ML20081L271

Final Draft, Estimating Economic Costs of Radiation- Induced Health Effects

ML20081L271
Person / Time
Site:	Limerick
Issue date:	11/30/1983
From:	Currie J, Hood L, Nieves L Battelle Memorial Institute, PACIFIC NORTHWEST NATION
To:	NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
Shared Package
ML20081L269	List: ML20081L271
References
PNL-4664, NUDOCS 8311160021
Download: ML20081L271 (175)

Text

_

d

.PNL-4664 FINAL DRAFT ESTIMATING THE ECONOMIC COSTS OF ,

RADIATION-INDUCED HEALTH EFFECTS Leslie A. Nieves J. William Currie Lance J. Hood Thomas M. Tierney, Jr.

, November 1983 Prepared for U.S. Nuclear Regulatory Comission Office of Nuclear Regulatory Research Pacific Northwest Laboratory Richland, Washington 99352 Operated by Battelle Memorial Institute For the U.S. Department of Energy 8311160021 831104 "

PDR ADOCK 05000352 h PDR F

l

SUMMARY

The research effort' covered by this report was performed by Pacific Northwest. Laboratory (PNL) for the Division of Risk Analysis, and the Division of Health, Siting and Environment, both within the Office of Nuclear Regulatory Research of the NRC. The purpose of this effort is to improve the quantitative information'available for use in evaluating actions that alter health risks due to population exposure to ionizing radiation. To pr,- ect the potential future costs of changes in health effect risks, PNL construued a flexible computer model, HECOM, which utilizes the output of an accident consequences model (CRAC2) to calculate the discounted sum of the economic costs associated with ionizing radiation exposure. Application of HECOM to value-impact and environ-mental impact analyses should greatly increase the quality of the information available for regulatory decision making.

Three major types of health effects .present risks for any population sus-taining a sign.ificant radiation exposure: acute radiation injuries (and fatalities), latent cancers.and impairments due to genetic effects. The liter-ature pertaining to both incidence and treatment of these health effects was reviewed by PNL and provided the basis for developing economic cost estimates.

The economic costs of health effects estimated ' by HECOM represent both the value of resources consumed in diagnosing, treating, and caring for the patient and the value of goods not produced because of illness or premature death due to the health effect.- Additional costs to society, such as pain and suffering, are not included in the PNL economic cost measures since they do not

' divert resources from other uses, are difficult to quantify and do not have a market value.

l l

r iii

y ACKriOWLEDGt1ENTS Because.of the multidisciplinary nature of -the effortlinvolved in develop-ing' estimates of;healthLeffect costs,~the authors have depended on the guidance landicriticism 'of findividuals from several fields.- We wish. to thank economists:

LJack LTawil:and Mac-Callaway for their critical comments on the conceptual basis

.in' economics for health effect cost ~ estimation. Guidance in regard to.the

-nature'and incidence of ~ genetic defects was provided by an epidemiologist; Lowell Sever.' . Ethel Gilbert, a statistician, assisted-by reviewing.our model-Ling off the incidence of cancers'. ' Sid Marks and Bill Bair:also provided criti-

.cism'from their. perspectives in health effects and environmental studies. 'In addition, we'wish to thank John-Burnham, Scot? Heaberlin, and Mark Mullen.for -

the advice!and supportf given to'our effort. We also wish to'thank the many Lindividuals at NRC who provided-support for this effort. Thanks especially to

. Clark Prichard, Don Cleary and -Brian Richter.

Leslie A. Nieves J. William Currie Lance J. Hood Thomas M. Tierney, Jr.

P e

v

(.  :. ~-:

CONTENTS

SUMMARY

............................................................... iii ACKNOWLEDGMENTS....................................................... v 1.0- INTRODUCTION..................................................... 1.1 1.1 THE NEED FOR HEALTH EFFECT COST ESTIMATES................... 1.1 1.2 OVERVIEW 0F SOCIETAL COSTS OF RADIATION-INDUCED HEALTH EFFECTS......................................*............... 1.2 1.2.1- Relationship of Health Effects to Costs. . . . .. .. . . ... . '1.2 1.2.2 Descri pti on of Heal th Ef fects . . . . . . . . . . . . . . . . . . . . . . . . 1.3 1.2.3 Composition of Costs................................. 1.7 1.3 SCOPE OF THE STUDY.......................................... 1.8 1.4 REPORT STRUCTURE............................................ 1.9

1.0 REFERENCES

....................................................... 1.11

2.0 CONCLUSION

S...................................................... 2.1 2.1 ACCOMPLISHMENTS............................................. 2.1

. 2.2 BOUNDING ESTIMATES.0F HEALTH EFFECT C0STS................... 2.2 2.3 RELATIVE MAGNITUDE OF HEALTH EFFECT C0STS................... 2.2

2.0 REFERENCES

....................................................... 2.6 3.0 REVIEW 0F HEALTH EFFECT PROJECTIONS.............................. 3.1 3.'1 ' RAD IATION INJURY INCIDENCE AND TREATMENT. . . . . . . . . . . . . . . . . . . . 3.1 3.1.1 Prodromal Symptoms................................... 3.1 3.1.2 Bone Marrow Syndrome................................. 3.2 3.1.3 Gastroi ntesti nal Syndrome. . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 3.1.4 Pu l mo n a ry Imp a i rme n t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4 3.1.5 I n - Ut e r o . I n j u ry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4 3.1.6 Other Radi ati on Inju ri es . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5 vii

.~ . . _ . . _ . . . _ . , _ _ . _-_ .- _,

3.1.7 The CRAC2 Projections of Radiation Injuries.......... 3.5 3.2 . CANCER INCIDENCE AND TREATMENT.............................. 3.8 3.2.1 In ci dence As sumpti on s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.9 3.2.2 Treatment Assumptions................................ 3.10 3.3 NATURE AND INCIDENCE OF GENETIC EFFECTS..................... 3.11 3.3.1 Kinds of Genetic Da' mage Associated with. Radiation.... 3.11 3.3.2 Estimation Methods................................... 3.12 3.3.3 Th e Ri s k Est i mat e s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.13 3.3.4 Clinical Manifestations of Genetic Disorders......... 3.14 3.3.5 The Reactor Safety Study and CRAC Model . . . . . . . . . . . . . . 3.15 3.3.6 Summary.............................................. 3.15

3.0 REFERENCES

....................................................... 3.17 14.0 VALUING CHANGES IN HEALTH RISKS.................................. 4.1 4.1 THE " HUMAN CAPITAL" APPR0ACH................................ 4.2 4.2 THE " INDIVIDUAL PREFERENCE" APPR0ACH........................' 4.4 4.3 , CONCLUSION.................................................. 4.7

4.0 REFERENCES

....................................................... 4.9 5.0 ESTIMATION OF THE DIRECT COSTS OF HEALTH EFFECTS................. 5.1

'5.1 DIRECT COSTS OF - RAD I ATION I NJUR IES. . . . . . . . . . . . . . . . . . . . . . . . . . 5.1 5.1.1 Prodromal Symptoms................................... 5.1 5.1.2 Bone Marrow Syndrome................................. 5.2 5.1.3 Gastroi ntesti na l Syndrome. . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3-5.1.4 Pu l mo n a ry Imp a i rme n t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4 5.1.5 I n - Ut e r o I n j u ry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4 5.2 DIRECT COSTS OF CANCERS..................................... 5.4 i

viii

5.2.1 Cancer Cost Data..................................... 5.5 5.2.2 Cost ~ Estimati on Methodol o gy. . . . . . . . . . . . . . . . . . . . . . . . . . 5.6 5.3 DIRECT COSTS.0F GENETIC EFFECTS............................. 5.11 5.3.1 Geneti c Ef fects Cos t Dat a. . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.12 5.3.2 . Cost Esti mati on Methodol o gy . . . . . . . . . . . . . . . . . . . . . . . . . . 5.13

5.0 REFERENCES

....................................................... 5.14 6.0 ESTIMATION'0F INDIRECT COSTS OF HEALTH EFFECTS................... 6.1 6.1 INDIRECT COSTS OF M0RBIDITY. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2 6.2 INDIRECT COST OF M0RTALITY.................................. . 6.3 7.0 HECOM STRUCTURE AND DEVELO PMENT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.1

.7.1 MODELING APPR0ACH........................................... 7.1

~

7.1.1 Fl exi b i l i ty- o f HEC 0M . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.1 7.1.2 Treatment of Costs Over Ti me. . . . . . . . . . . . . . . . . . . . . . . . 7.2 7.2 OVERVIEW 0F HECOM STRUCTURE................................. 7.4 7.2.1 Major HECOM Processes................................ 7.4 7.2.2 Cal cul ati on of Cancer Di rect Costs. . . . . . . . . . . . . . . . . . . 7.6 7.2.3 Calculation of the Direct Costs of Radi ati on In ju ri e s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.6 7.2.4 Calculation of. the Direct Costs of Genetic Effects... 7.8

'7.2.5 Calculation of Indirect Costs of Fatalities.......... 7.8 7.2.6 Calculation of Indi rect Costs of 111 ness. ... . . . . . .. .. 7.9 7.2.7 Projection of Fatalities............................. 7.11 7.2.8 Projecti on of Genetic Ef fects . . . . . . . . . . . . . . . . . . . . . . . . 7.12 7.2.9 Projecti on of Labor Value Over Time. . . . . . . . . . . . . . . . . . 7.12 7.2.10 Projection of Cohort Survi val Probabilities. . . .. . . . . . 7.13 7.3- MODIFICATION OF CRAC2 OUTPUT FOR USE AS HECOM INPUT......... 7.13 ix

l 1

1 1

7.3.1 Acute Effects...................................'..... 7.13 )

7 . 3 . 2 Ca n c e r s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.16 '

7.3.3 Genetic Effects...................................... 7.16 7.4 HECOM SENSITIVITY ANALYSIS.................................. 7.16 7.4.1 Sensi ti vi ty to the Di scount Rate. . . . . . . . . . . . . . . . . . . . . 7.17 7.4.2 Sensitivity to Labor Productivity Growth Rates....... 7.17 7.4.3 Sensi ti vi ty to Treatment Cost Growth. . . . . . . . . . . . . . . . . 7.18 7.4.4 Sensi ti vi ty to Earni ngs Level s . . . . . . . . . . . . . . . . . . . . . . . 7.18 7.4.5 Sensitivity to Treatment Costs....................... 7.19 7.4.6 Sensi ti vi ty to Week s of Il l nes s . . . . . . . . . . . . . . . . . . . . . . 7.19 7.4.7 Sensitivity to Labor Force Participation Rates....... 7.19 7.4.8 Comparison of Median and Interval Data Results....... 7.20 7.0 REFERENCE........................................................ 7.21 8.0 HEALTH EFFECT COSTS FOR A HYP0THETICAL REACTOR ACCIDENT.......... 8.1 8.1 HEALTH EFFECT ESTIMATES..................................... 8.1 8.2 COSTS ESTIMATES............................................. 8.1 APPENDIX A: HEALTH EFFECT COST M0 DEL................................. A.1 APPENDIX B: PRELIMINARY HECOM COMPUTER C0DE.......................... 8.1 X

- - . ~ , , . . .

. . FIGURES 1.1 Diagram of Radiation-Induced Health Effects and Resultant Social Costs.................................................... 1.4

'7.1- Overview of Health Effects Cost Model Processes.. . . .. .... . . . ... . 7.4 7.2 .HECOM Calculation of the Direct Costs of Cancers b

. a n d Ca n c e r Typ e . . . . . . . . . . . . . . . . . . . . . . . . . . .............

. . . . . . . ., . .y Se x7.7

-7.3 HECOM Calculation of the Direct Costs of Radiation In by Se x and In j u ry Typ e . . . . . . . . . . . . . . . . . . . . . . . . ...........

. . . . . . .ju ri es 7.7 7.4 - HECOM Calculation of the Direct ' Costs of Genetic Effects........ 7.8 7.5- HECOM Calculation of Indirect Costs of Premature Mortality, by Age, Sex, and Cause of Death................................. 7.9 7.6 HECOM Calculation of Indirect Costs of Illness, by Age, Sex, and Cause of 0eath.............................................. 7.10 77 HECOM Projection of Fatalities, by Age, Sex, and Cause of Death 7.11 7.8 HECOM Projecti on of Geneti c Ef fects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.12 7.9 HECOM Projection of Labor Value, by Age and Sex. . ...... . . .. . .. . . 7.13 7.10 HECOM Projection of Cohort Survival Probabilities b a n d Se x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .............

. . . . . . , . . y Ag 7.14 e

i.

l i

I o

l l

f xi l.

l l

1 1

-l TABLES 2.1 HECOM Present-Value Estimates of Radiation Injury, Cancer a n d Ge n e t i c E f fe ct Co s t s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 3.1 Clinical Progression of Acute Radiation Syndrome.. . . .. .. .. . . .. .. 3.3 3.2 Dose Values and Associated Mortality Rates Used in CRAC2........ 3.6 3.3 Dose Values and Associated Morbidity Rates Used in CRAC2........ 3.7 3.4 Summary of Early Inju ry-Related Information. . . . . . . . . . . . . . . . . . . . . 3.8 3.5 Estimated Increase in Genetic Disorders per Million Liveborn, from an Average Population Exposure of One Rad. . . . . . . . . . . . . . . . . . 3.13 4.1- Extent to Which Selected Methods Measure the Various Co mp o n e nt s o f Va l u e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.7 5.1 Radi ati on Inju ry Cos t Estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4 5.2 Corresponding Cancer Categories in CRAC2 and the Third Nat i o n a l Ca n c e r Su rv ey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.7 5.3 Direct Costs of Cancer Care for First Two Years of.Tr by Ca n c e r Typ e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. .. .. ... .. .. e. .a. .t men5.9 t

- 5.4 Calculation of Cancer Incidence Based on CRAC2 Fa t a l i t y Es t i ma t e s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.10

5. 5_ Oi rect Costs of ' Cancer Care by Cancer Type. . . . .. . . . . . . . . . . . . . . . . 5.10 7.1 Sensitivity of HECOM Estimates to the Discount Rate. .. . . . . . . . . . . 7.17

! 7.2 Sensitivity of HECOM Estimates to the Rate Labor of Productivity Growth............................................. 7.18 l 7.3 Sensitivity of HECOM Estimates to the Rate of Treatment Cost Growth..................................................... 7.18 7.4 Sensitivity of HECOM Estimates to Earnings Levels............... 7.19 7.5 ' Sensitivity of HECOM Estimates to Treatment Costs. . . . . . . . . . . . . . . 7.19 7.6 Sensitivity of HECOM Estimates to Weeks of Illness... . . .. ... . . . . 7.19

! 7.7 Sensitivity of HECOM Estimates to Labor Force Participation Rates............................................. 7.20 i

xii c

l

-,,w wm--*yw---,

F~.'

- 7.8 ' Comparative Results-of Median and Interval Data Cases...........- 7.20 8.1. Project Numbers.of' Health. Effects for One Reactor Accident Scenario Used as Input to the Sample HECOM Calculation.......... 8.2 8.2 Projected Health Effect-Costs for One Reactor

- Ac c i d e n t Sc e n a r i o . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3 O

1

(

f xiii

1.0. INTRODUCTION This report was prepared by the Pacific' Northwest Laboratory (PNL) for the

Division of Risk Analysis Jnd the ' Division of Health, Siting and Environment,

?both within the Office of Nuclear Regulatory Research"(RES) of the Nuclear i- Regulatory Commission (NRC). The purpose of this effort is to improve 'the-quantitative information used in evaluating actions that alter health risks.

~

To fulfill this purpose, _ PNL 1) evaluated the conceptual .and informational

_ basis for measuring the. total cost to society of radiation-induced health effects, 2): estimated economic costs for the major types.of' potential E

radiation-induced health effects,' and 3) developed a flexible computer code for calculating costs that could result over time due to a.s' ingle nuclear inci-

-dent. ' As a' result of 'this effort, quantitative estimates of the economic costs -

of health ~effect_ risks will be available for inclusion in environmental impact

-statements-for. nuclear facility siting and for evaluation of safety-related

. actions. This section'of the report covers the need for health effect cost estimates, the nature of the. health effects and the composition of resulting costs, the scope of the PNL effort .and an outline of the report structure.-

1.1 THE NEED FOR HEALTH EFFECT COST ESTIMATES Estimates'of the health effects that may result from radiation-exposure

' are used by NRC in many types-of analyses. Unlike other types of potential accident consequences, such as'offsite property damage,'a dollar value has not generally been' ascribed to potential health effects. This is in part due to the relative lack of economic models and data for the costing of health effects. A number of recent efforts have substantially improved the, economic b data in this area and this present work offers an economic model.

Th.e lack of economic treatment of health effects has also been due to the

. argument that it is -inappropriate, or even imoral, to place an explicit value

on human life and health. This study does not attempt to estimate the value of human-life or health; it estimates the economic losses to society that could L . occur due to radiation-induced illness and injury. Although the argument may t be.made that property damages and human health effects are qualitatively dif -

L ferent, the measurable economic costs of health effects are better included in i risk-related decision making than excluded. Although available information is i- incomplete, having it is preferable to having no information as to the relative magnitude of health effect costs.

n The cost estimates resulting from this study have applications in several types of analyses carried out by NRC. They may be used in developing health effect impact assessments for the nuclear fuel cycle, in total or in part.

They'are needed to evaluate safety goals, especially the benefits of avoiding health risks. In addition, there are applications in nuclear facility licens-ing procedures and National Environmental Policy Act (NEPA) related t-- assessments. ,

i L

l 1.1

. -. -. . ---. -_. -- . -~ -- -. =.. ..

i 1.2.0VERVIEW~0FSOCIETAL(a[COSTSOFRADIATION-INDUCE 0HEALTHEF:

s i The value of avoiding radiation exposure, whether for the general popula-tion or for workers, is determined by the total cost to society that is likely to result ' rom the effects of exposure. All health effects result .in costs to society because of the resources consumed in treating the illness and because of the . lost productivity"of the affected . individuals. These. primary economic

, costs are referred to as direct.and indirect costs. Direct costs include all f costs for treatment,. travel to obtain treatment, patient care, equipment and supplies, while indirect costs are the losses due to the reduced productivity of the patierit or his family. Such productivity losses may occur because the patient is too ill to work, the family is caring for the patient, the patient's functioning is permanently impaired or the patient dies.at a younger age than would have been likely without the radiation-induced health effect.

In addition to the pr1(sary costs of health effects, there are secondary costs that are nonmonetary in nature. ~ These costs include the ulue of pain and suffering; the cost of family members' stress-induced illness precipitated by the illness or death of the patient; the cost of depression or psychological stress due to actual or anticipated illness. While recent attempts have been made to measure some of these effects,' no rigorous estimates of secondary costs

, 'are available, eitner in absolute terms or relative to primary costs.

The relationship between the occurrence of health effects and the occur-rence of economic costs is discussed in Section 1.2.1. In Section 1.2.2 which

' follows, the types of health effects that may be induced by radiation are described briefly. Some of the difficulties in accounting for societal costs are discussed and the measurement approach taken by PNL is explained in Section 1.2.3.

1.2.1 Relationship of Health Effects to Costs Three major types .of health. impairments may result from accidental radia-tion exposure: acute radiation injury, cancer and genetic damage affecting future generations. Each of.these may result in premature mortality, as well as morbidity (illness) and physical impairments. Most types of acute radiation

+

injuries would become apparent within a few weeks of exposure and the resulting fatalities would generally occur wfthin six months. With a few important ex-l.

captions such as leukemia, cancers would not be apparent until ten to fifteen years after radiation exposure and incidence might be spread over the remaining lifetime of the affected population. The genetic effects of concern would occur in the offspring of the exposed population and then diminish in frequency over subsequent generations. As-a result of the delayed impact of genetic effects of radiation exposure, the costs of the health effects would be spread over a substantial period of time. While secondary nonmonetary costs would be

. associated with the health effects, they are not estimated and are not included

-in this discussion.

(a) Societal cost includes all monetary and nonmonetary costs, while the PNL health effect cost estimates include the subset of costs which are monetary, or economic, in nature.

1.2 l:

_. _ _ _ _ _ _ _ _ _ = _ . . . . . _ . _ _ . . _

Although the details are complex, the basic process by which health effects result in economic costs is shown in its simplest form in Figure 1.1.

The starting point is a population that has been exposed (1) to a source of radiation at some point in time. Depending on the dose received, and the.

period of exposure, individuals may develop acute radiation injuries (2) of varying severity. If symptoms develop, society incurs direct costs for the treatment of the illness and indirect costs due to the decreased productivity (3) of the stricken individual. Those individuals for whom treatment is inef-fective die (4) resulting in additional i edirect costs (5) to society from the premature loss of their productive capar';y.

Those who survive the radiation injuries, as well as those who were unin-jured, may develop cancer (G) at some time after the latency period. Both direct costs for treatment and indirect costs due to lost work (7) accrue to society as a result of the cancers. For those who succumb to cancer (8), there are additional indirect costs (9) of productivity loss due to their premature mortality. .

The portion of an exposed population that is unaffected by, or survives, radiation injuries would face the risk of bearing offspring with dominant or recessive genetic damage (10). Health impairment due.to these genetic effects could result in direct costs for medical treatment and indirect costs due to reduced productivity (11) of the affected individuals and the families who care for them. The health effects and their economic costs may continue for many generations.

1.2.2' Oescription of Health Effects Three major types of health effects are of concern for any population sustaining a significant radiation exposure: acute radiation injuries, cancers and genetic effects. Brief descriptions of the illnesses incluoed in each of these categories are provided below. Further detail related to the incidence and treatment of these effects can be found in Chapter 3.0.

l Acute Radiation Injuries The occurrence of acute radiation injuries among an exposed population is I dett mitied by the total dose, the dose rate at which the dose is received, and by the quality of the radiation.

A wide variety of biological effects may result from exposure to radia-tion. The possibilities vary in intensity from negligible or undetectable to those that are more severe: temporary discomfort, permanent impairment, and life-thraatening effects. Characteristics of the major types of radiation injuries are given below. For external sources of x-rays, gamma rays, and beta particles, the dose units " rad", and " rem" are equivalent.

r 1.3

EXPOSED g POPULATION NO ATION INJURY 2

?

, ,YES DIRECT AND \\

INDIRECT COST OF RADIATION

!NJURY 1 RYES INDIRECT SURVIVE NO / COSTS OF 5

? EARLY FATALITIES /

YES 1r ,

ir NO DEVELOP CHILD 10 NO (NO COST CANCER 6 WITH GENETIC NO COST)

? DEFECT

, ,YES ,7 YES DIRECT AND DIRECTAND INDIRECT 7 INDIRECT COST 11 COST OF OF GENETIC '

CANCER DEFECTS 1r SURV INDIRECT COSTS NO _

- OF EARLY 9 FATALITIES

, ,YES (NO COST)

FIGURE 1.1. Diagram of Radiation-Induced Health Ef fects and Resultant Social Costs 1.4

e' Prodromal Symptoms - These flu-like symptoms may result from a com-bination of the effects of tissue damage and anxiety about the ulti-mate effects of the individual's radiation exposure (Blakely 1968,

p. 35; Dalrymple 1973, p.192). Symptoms begin within a few hours of exposure and generally subside in a few days. Affected individuals may experience nausea, loss of appetite, headache, diarrhea and weak-ness. Occasionally, individuals receiving a dose as iow as 50 rads may be affected and at doses above 200 rads virtually everyone would exhibit these symptoms (Blakely 1968, p. 35).

e Bone Marrow Syndrome - This process is initiated by whole body expo-sures of 200 rads or more. There is damage to the bone marrow, spleen and lymph nodes which in turn results in impairment of the body's blood forming and immune functions (NRC 1975, Appendix VI,

p. F-1). The illness is characterized by infections, hemorrhage and anemia, which may be fatal alone or in combination. Approximately 50 percent of exposed individuals may be expected to die within two months of exposure at doses greater than about 450 rads (NRC 1975, Appendix VI, F-3).

e Gastrointestinal Syndrome - At.whole-body doses above 600 to 1000 rads, cellular damage may result in gastrointestinal symptoms.

. Symptoms include vomiting and diarrhea with severe fluid loss, fail-ure of food absorption and hemorrhage. Intestinal ulceration may occur, accompanied by bacterial invasion (Blakely 1968, p. 41).

Affected persons may be expected to die within 10 to 14 days or to survive to exhibit the bone marrow impairment described above.

• Pulmonary Syndrome - Doses of about 750 rads or more (Cooper, et al.

1982, p. 4-6) can result in impaired pulmonary function. The.e may be pulmonary infections, and shortness of breath may in turn affect heart function. Generally, injuries from lung exposure induce pneu-monitis, followed by pulmonary fibrosis (NRC 1975, Appendix VI,

p. F-3).

e Hypothyroidism - This is an impairment of thyroid function which can be, induced by radiation exposure. Oral medication is effective and inexpensive (NRC 1975, Appendix VI, p. 9-13).

e Sterility - Radiation-induced sterility may be either temporary or permanent. For males temporary effects occur at a lower dose than for females but a higher dose is required for permanent effects.

Permanent sterility, in males or females, is unlikely below doses that are life threatening if whole body exposure is involved (NRC 1975, Appendix VI, p. 9-15).

e Cataracts - Doses of 200 to 500 rads to the lens of the eye may result in formation of cataracts after a latency period that varies with both dose and dose rate (NRC 1975, Appendix VI, p. 9-18).

1.5

e Skin a'nd Hair Damage - Loss of hair occurs two to three weeks after external doses in excess of 300 rads. This is likely to be temporary unless the dose exceeds 600 rads (NRC 1975, Appendix VI, p. F-13).

The skin may also be affected'by doses in this range, resulting in radiation dermatitis. This condition has levels of severity compar-able to first, second and third degree thermal burns and in the most severe cases (due to doses of over 2000 rads) can result in permanent skin ulceration (Prasad 1974, p. 240-248). Survivable whole-body acute doses are unlikely to cause more severe injuries than hair loss and skin reddening.

e Prenatal Injury - The radiosensitivity of embryos is very nigh, resulting in deaths from doses as low as ten rads. Most such deaths would be unnoticed due to the early stage of the pregnancy. In later stages of development the fatality rate decreases but the probability of abnormalities increases. These generally take the form of growth impairment and mental retardation, especially microcephaly. As in the case of prenatal mortality, cases have been documented after exposures .of about ten rads (NRC 1975, Appendix VI, p. F-17-20).

Cancers Cancers induced by radiation exposure are indistinguishable from ether cancers. As a result, the cause of any particular cancer is rarely, if ever, identi fiable. Radiation-induced cancers may only be apparent as an increased statistical rate of cancer formation in an affected population. The " excess" cancer may then be attributed to the radiation exposure of the population.

Susceptibility to cancer varies among organs and tissues, so that the rates differ at which excess cancers appear in various sites. Cancer induction

- is influenced by sex, age when irradiated, and type of radiation, among other factors (BEIR 1980, p. 84-5). The cancers that are most susceptible to radia-tion induction are leukemia and cancers of the breast, bone, lung and gastro-intestinal tract. Both benign and malignant thyroid nodules may also be induced. While it is possible for radiation-induced cancers to occur in other organs and tissues, the types mentioned above are the most likely and are the focus of concern in the Calculation of. Reactor Accident Consequences Model (CRAC) as well as in this study.

Genetic Effects l  : Genetic effects, in the form of abnormalities and diseases, may affect many generations of the offspring of persons exposed to radiation, though at a decreasing rate over time. Radiation may increase the mutation rate, but does l not affect the nature of the mutation or the associated health effects. Thus, the health effects that occur are of the same type that occur spontaneously.

Of the possible types of mutation, autosomal dominant disorders are most likely to increase in direct proportion with radiation exposure. These disorders may cause chondrodystrophy, osteogenesis imperfecta, neurofibromatosis, eye

1.6 l

. .- . _- - .- - . . - ~ _ . - -. - - - - - _ . - - ~ .

^ - anomalies, polydactylism and polycystic renal disease. Other types of health effects due to autosomal dominant mutations occur much less frequently (NRC 1975, Appendix VI, Appendix I).

1.2.3 Composition of Costs ,

m L The value of avoiding radiation-associated illness'can be measured con-l ceptually in two different ways; by estimating the value that the public-places on decreasing risks to health and safety, or by measuring the costs l' associated with higher levels of risk. A review of the relative merits of the two approaches is included in Section 4.0. The PNL health effects model fo-cuses on costs because they are more directly measurable and because they ac-count for a substantial part of the public's evaluation of risk.

i i There are two ways to estimate the cost of illness, from either a pre-l valence or an incidence perspective (Hartunian, Smart and Thompson 1981).

g Essentially, a prevalence approach asks, "How much is an illness, e.g.,

. cervical cancer, costing U.S. society in 1983?" It 5%s the costs _in a given

year of all cases of an illness regardless of the ca.tse. In contrast, an incidence approach would focus on the question, "If a specified event occurred

'in 1983, what would be the resulting cost of induced cases of cervical cancer?"- This approach permits evaluation of the benefits.of changing the rate of development of new cases of disease. PNL employs the incidence approach in estimating the costs of radiation-induced health effects.

The economic costs of illness represent both the value of resources con-sumed in diagnosing, treating, and adapting lifestyles to the illness, and the j value of goods that do not get produced because _of morbidity or premature mor-

tality;from the illness. . It has been the convention in health economics i studios to label the constmed resources the direct costs of illness, the for-l gone production ~the indirect costs. Both the direct and indirect costs are L . measured in dollars. In addition to the economic costs, there are associated with . illness and death a variety of social effects that constitute intangible l- costs (Abt 1975). These elements of social costs, such as pain and suffering,

! are not included in economic cost measures, since they do not divert resources I from other productive uses, are difficult to quantify and do not have a market value. However, it is-clear that they are an appropriate matter of concern to the public in considering illness risks.

Direct costs are measurable in terms of monetary outlays both for health care and for other goods and services made necessary by the illness. Thus, direct costs include the costs of health care services on both an inpatient and i

outpatient basis for diagnosis and treatment. In addition, a full accounting l of direct costs would include expenditures for such things as treatment-related travel and modification of housing (a wheelchair ramp, for example) and for population screening for illness.- Unfortunately, the literature includes little information on these nonmedical direct costs; they are not included in PNL's cost model. Direct costs of health care may represent a stream of out-lays over a period of years. In tais study, future streams of direct costs are measured in terms of their preseat value in the year of radiation exposure.

i

( 1.7 1

l

__w.,_n__u.w._..,-.__-.-...-._...._,_-.-

Indirect costs involve no monetary outlays, .but reflect instead the value '

of lost productivity due to illness. Productivity losses may occur because the

- patient is too -ill to work, the family is caring for the patient (and they are

.; - therefore unable to work), the patient's functioning is permanently impaired, or the patient dies at a younger age than would have been likely without the radiation-induced illness. In any of those cases society forgoes the goods and services that would have been produced had the patient (or' family) been able to work. Valuing these productivity losses is similar to valuing capital invest-ments in terms of future. output and .is therefore generally known as a " human capital" approach to measuring indirect costs.

PNL employs the' human capital approach in valuing indirect costs. and includes among those costs the total present value of production forgone be-cause of radiation-related morbidity and mortality. A variant of the human

. capital- approach would include 'among indirect costs only the forgone net pro-

duction; that is, the value of the person's productiOb less his or her future consumption (for example, Weisbrod 1961). However, net production measures only

the value that the rest of the public places on someone's life and ignores

-the value that person derives from his or her own personal consumption. The total production approach comes closer, therefore to a full measure of the ,

indirect costs in terms of human capital.. ( A fuller discussion of alternative

, approaches to measuring indirect costs is provided in Section 6.0)

Assuming workers are paid the .value of their marginal product, the value

. of lost production is equal to the value of forgone future earnings. Following an incidence approach, as is employed for direct costs, indirect costs are

,  ; measured in terms of present value in the year of exposure..

. . In estimating the costs of health effects, we' assume that in the event of

population exposure, the change in demand for health care services would not be o sufficient to affect the . price structure. A similar assumption is made in l regard to indirect costs, that the ' numbers of fatalities involved would be insufficient to affect wage rates or prices. Thus, only small, or marginal changes within our economic system are considered in estimation of health ef-fect costs.

1.3 SCOPE OF THE STUDY Radiation-induced health-effects may result in both economic costs and

- nonmonetary impacts on society. PNL cost estimates are limited to the eco-l nomic costs

1) the. direct costs of health care provision and 2) the indirect l costs of productivity losses resulting from illness or premature mortality.

Other measures of health effect ' impacts, such as the value of pain and suffer-ing, are beyond the scope of this effort.

i:

. The PNL cost estimates represent the present value of probable future costs.that are .likely to be associated with each of the major types ( f radia-tion-induced health effects. In the case of acute radiation injuries. PNL 1.8

,c,._ .-,-.,,,w.i ,. w-, e--+. , , , ,o y,-_,-,~,,,,y-,-.m. -,,,.,.m...-,,---.,-. ,-e,.,, , - , , - ,e.,., .w,--.c.,,,,.,,, .-.-,..,,,n-,v. -

estimates the costs of bone marrow syndrome, gastrointestl ngl. syndrome, pulmon-ary impairment, prodromal symptoms and panatal injuries. . For cancers, the

PNL cost estimates cover the-same.categcries projected by the.CRAC2 model

<

1eukemia, lung, breast, bone, gastrointestinal tract, thyroid and all others.

In addition, direct 'and- indirect costs are considered for radiation-induced genetic effects occurring in future generations.

The cost. estimation methodology is designed to be compatible with the health effect output of the CRAC2 model, but also to accommodate health effect

. projections from other sources as well. PNL has developed a Health Effect Costs .Model' (HECOM) for implementation of this methodology. The model is modu-lar in structure and is designed for flexibility and ease in modification and ,

updating. It is expected that HECOM'can be readily adapted to future changes in CRAC2 and related models for projecting health effects.

a 1.4 REPORT STRUCTURE i

This report presents the conceptual and informational 3dse f;'om which PNL~ l has developed health.effect cost estimates. It dest-ibes in detail the meth-odology employed in estimating each component of these costs. In addition, it provides a description and documentation of the model (HECOM) developed to cal-culate the present value.of possible future health effect costs. Conclusions and recommendations of the effort are. presented in Chapter 2. This includes a

. discussion of the limitations.of the cost estimates, the relative importance of the major. cost components and recommendations for-further research.

In Chapter 3 we review the ' major. health effect studies and models which provide the basis for cost estimation. Assumptions as to health effect inci-

.dence and timing that affect cost estimates are discussed, as are the uncer-tainties involved in the health effect projection.

Though the estimation of health effects costs is difficult, the diffi-

-culties stem from incomplete medical and economic data and information, rather than inadequacy of the conceptual basis .for such cost estimates. Chapter 4 presents the conceptualL basis and discusses the two major approaches to mea-

-surement of health effect costs: the individual preference approach and the human capital approach.. Because of its greater tractability, PNL employs the human capital approach in developing cost estimates.

The methodology used in this cost estimation is detailed in Chapters 5 and 6 In Chapter 5 the direct costs of radiation-induced morbidity are dis-cussed. - Costs for radiation injuries are developed in Section 5.1, costs for cancers cin Section 5.2,'and costs for genetic effects in Section 5.3. These sections present information as to likely treatments and the associated costs, and describe the methods used to calculate each cost component. Similar (a) .0ther types of radiation injuries, such as cataracts, are not included

, because they are dominated by the effects of actue whole-body exposure.

1.9

..-.:-. ,, . a- -,_:.-.--_-.--.-_..-.-.._.-.-..

. i:

information is presented in Chapter 6 for the indirect costs of morbidity. The-same cost' estimation metholodogy applies to each type of health effect.

An'~o verview of PNL's Health Effects Cost Model (HECOM) is provided in Chapter.7 'The _ general approach employed to develop a flexible health effect

. costs model,is presented in Section 7.1. HECOM will accept input data from various sources, will' allow simulation of alternative health effect incidence assumptions and can . easily be modified or updated. The model -structure is described in- Section 7.2 and use of CRAC2 data as inputs to HECOM~is discussed in Section' 7.3. The sensitivity of health effect cost estimates to various data and model parameters is explored in Section 7.4.

Documentation of the model appears in Appendix A, along with summaries of the data.used in the base case. The computer code is listed in Appendix B.

M q.

2 -

.I.

fc s.. .

.t' j_

- vp ~

Q(l' b

w Y4 7$ 1.10 7.

.f f , - . , . - - . - ....J... . _ , , _ , ,. . - ;-f .:_..__._ , _ - - . . . ._ .

1.0 REFERENCES

Abt, C. 1975. "The Social Costs of Cancer." Social Indicators Research, Vol.

2, pp. 175-190.

Blakely, J. 1968. The Care of Radiation Casualties. Charles C. Thomas, Springfield, Illinios.

Comittee on the Biological Effects of Ionizing Radiation. 1972 and 1980. The Effects on Populations of Exposure to Low Levels of Ionizing Radiation.

National Academy of Sciences, Wasnington D.C. .

Cooper, D. W. et al . 1982. Reactor Safety Study Radiological Health Effects Model: Critical Review. Sandia National Laboratories, Albuquerque, New liexico.

Dalrymple, G. V. et al., eds. 1973. Medical Radiation Biology. W. B.

Saunders Co., Philadelphia, Pennsylvania.

Hartunian, N. S., C. N. Smart, and M. S. Thompson. 1981. The Incidence and Economic Costs of Major Health Imoairments. Lexington Books, Lex 1ngton ,

Massachusetts.

Prasad, K. N. 1974 Humbn Radiation Biology. Harper and Row, Hagerstown, Maryland.

U.S. Nuclear Regulatory Comission. 1975. Reactor Safety Study. Appendix VI. WASH-1400, National Technical Infonnation Service, Springfield, Virginia.

Voilleque, P. G., and'R. A. Paslick, 1982. " Societal Cost of Radiation Exposure." Health Physics. Vol. 43, No. 3, pp. 405-409.

Weisbrod, B. A. 1961. The Economics of Public Health. University of Pennsylvania Press, Philadelpnia, Pennsylvania.

i 1.11

f l

2.0 CONCLUSION

S

~ Preliminary conclusions from the cost estimation effort are presented-below. -_ An overview af the accomplishments in this first attempt to rigorously estimate health effect costs is presented in Section 2.1. The scope and focus c of the study are indicated and some of the limitations are explained. Section 2.2 describes the . level of uncertainty inherent in; the HECOM c'ost estimates, apart from the uncertainty in the estimated numbers of health effects that are

-input. to HECOM. - The' estimated ranges of costs for each type of health effect are then presented in Section 2.3. This is followed by suggestions for further research in regard to refinement of cost estimates, improvement of health effect incidence estimates and application of HECOM to risk analyses.

2.1 ACCOMPLISHMENTS

. To improve the . quantitative . information used in evaluating actions that alter ' health risks, this study and the health effect cost model (HECOM) provide estimates of the economic costs of the principal types of radiation-induced health effect L Tho study presents the conceptual basis for measuring direct and indirect economic costs and it describes in some detail. likely medical treatment of radiation-related health impairments. PNL's cost model, HECOM,

. calculates the present dollar value of resources that would be consumed in treating radiation-induced health effects and the resources that would not be produced because of exposure-related morbidity and early mortality.

HECOM is a flexible computer code that combines health effect incidence and timing with streams of treatment costs and lost productivity values to-approximate the sum of direct and indirect costs of potential acute radiation injuries and fatalities, cancers and genetic effects. The flexibility of HECOM allows analysis of costs while varying key parameters. The model can accept changes in incidence estimates, in treatment costs, in the discount rate and in real growth rates. Because of its flexibility, it will be adaptable over time as information improves regarding risks, treatment regimens and costs.

, Use of HECOM estimates requires a clear understanding of the model's focus. .Two general points are important in this regard: first, the model includes only the major forms of potential radiation-induced health impairments and second, the model centers on health effect costs and not on society's val-uation of risk to life and health.- HECOM calculates costs for acute radiation injuries and fatalities, cancers, and genetic disorders. However, it leaves uncounted other potential effects that may be nonetheless important considera-

, tions to the public, such as psychological stress and sterility. For the major types of health effects, HECOM calculates the associated monetary costs. Thus, the HECOM cost estimates do not measure the total value of life or health but only the value of resources that would be used or not produced because of ill health or early mortality.

The economic cost figures obtained from HECOM are useful as rigorous and documentable cost estimates for health effects potentially associated with 2.1

, . . .- -..::. - -. - . u - - - - . - -. - -- - - - - - -- - -

population exposure to ionizing radiation. They constitute heretofore unavail-able information that is appropriate for use in value-impact analyses and environmental impact statements for nuclear facility siting. While there is room for refinement of the health . cost estimates, they provide an indication of the relative magnitude of health effect costs for use in regulatory decision making.

, 2.2 BOUNDING ESTIMATES OF HEALTH EFFECT COSTS There is considerable uncertainty in the health effect incidence estimates that are currently available for input to HECOM for cost calculation. In addi-tion, there is uncertainty regarding.the distribution of cancers and genetic effects over time. For cancers the choice of an absolute versus a relative risk model has a major effect on cost estimates. We are. currently using an absolute risk model to distribute cancer incidence over time in HECOM. In regard to genetic effects, there is uncertainty as to the frequency of defects of various degrees of severity. We have made a number of assumptions to develop cost estimates, however, but available information regarding genetic defect severity is inadequate for estimating the level of uncertainty.in our severity estimates.

There is considerably less uncertainty regarding the direct ano indirect cost estimates we have developed for radiation injuries, cancers and genetic effects. Using the HECOM base case parameters of a four percent discount rate and one percent growth rates for medical . costs and labor productivity, the level of uncertainty in total- costs due to the uncertainties in the direct and indirect cost components is about 25 percent.

2.3 RELATIVE MAGNITUDE OF HEALTH EFFECT COSTS Since a probabilistic methodology was used in developing HECOM, the resulting health effect cost estimates do not represent the costs for any par-ticular individual. Rather, the HECOM cost 'estir:ates are representative of costs for a population with a specified age' and sex distribution, for whom both health effect risks and resulting costs vary with age, sex and other factors.

For instance, cost estimates for cancers and genetic effects are based on prob-ability distributions of incidence and associated costs over long time periods. These cost estimates should not be confused with the average cost of a cancer or genetic effect occuring at any specific future time; they are sta-tistical constructs.that weight the probability and magnitude of costs in each year of the period modelled by HECOM and discount this stream to a base year.

It is this characteristic of the HECOM estimates that makes them most suitable

.for use in evaluating changes in health effect risks.

Results of the HECOM base case are shown in Table 2.1 where direct, indi-rect and total costs are listed for acute radiation fatalities and injuries, cancers and genetic effects. For total costs, .a. *25 percent range of uncer-tainty is shown, based on a sensitivity analysis of HECOM cost estimates. The

.present-value cost estimates in Table 2.1 are for one case of each type of 2.2

TAR.E 2.1. EO(N Present-Value Estimates of Radiation Injury, Cancer and -

Genetic Effect (bsts (1981 $)

Direct Cbst (000 $) Indinect Oost (000 $) Total 0)st (000 $)

Base t25 Percent Radiation Injuries Prodrmal 1.0 0.1 1.1 0.8 - 1.4 Ebne Mirrow 56.0 72.5 128.5 96.4 - 160.6 Lung 3.6 72.5 76.1 57.1 - 95.1 Gastrointestinal 28.0 72.5 75.4 - 125.6 Prelatal 100ja) 100.0 181.2(a) 281. 210.9 - 351.5 Cincers Leuksnia 11.0 120.4 131.4 98.6 - 164.3 lmg 7.6 , 18.9 26.5 19.9 - 33.1 Gastrointestinal 5.8 18.7 24.5 18.4 - 30.6 Reast 3.1 20.4 23.5 17.6 - 29.4 Bone 21.2 96.5 117.7 88.3 - 147.1 All others 4.0 20.2 24.2 18.2 - 30.3 Thyroid 1.8 0.4 2.2 1.7 - 2.8 Genetic Effects 34.3 17.2 51.5 38.6 - 64.4 (a) Because of the HEON aggregation proced;res, this fig 2re includes sane indirect costs of cancers affecting individuals irradiated in-utero, health effect probabilistically distributed over an exposed population and over time. Because the costs cover such a wide range due to the underlying varia-tion in health effect severity (such as the difference between prodromal symp-toms and prenatal injuries), an average would not be representative of the cost distributions.

For radiation injuries the total costs range from those for prodromal injuries ($0.8K to $1.4K), through those for lung injuries and for manifesta-tions of acute radiation syndrome, to the costs of prenatal injuries that are over $200K per injury. Since these injuries are qualitatively different in nature, as well as in costs, they are best considered as five separate cate-gories of effects rather than as a single category, radiation injuries.

Cancer costs cover a slightly narrower range, from those for nonfatal thyroid nodules and thyroid cancers ($1.7K to $2.8K) to those for leukemia

($98.6K to $164.3K). The indirect costs of leukemia and bone cancer are sub-stantially higher than those of other cancers, mainly due to the potential brevity of the latency period.

The cost estimate for a genetic effect has a range from $38.6K to

$64.4K. This cost estimate may be interpreted as the value of avoiding the i

2.3

-- ._. _ _ _ _ _ - . ~ . - , _ _ _ _ ._ - - . _ -. ,_. -

risk of one individual's health impairment due to a genetic effect that would occur'within the subsequent ten generations. Serious and minor effects are weighted in estimating the genetic effects costs so the estimate appli.es to the broad category of genetic effects.

The individual health effect cost. estimates given above may be applied to numbers of soc.ii!c types of cancers or injurias (e.g. leukemia, prodromal symptoms) tJ evaluate total health effect costs for an daffected population.

Ways in which the above cost estimates could be improved are discussed below.

Regarding the estimation of cancer risks, there is reason to believe that recent data from the Japanese A-bomb survivors may lead to increased use of relative risk models to model cancer risk (Cooper et al.1982, Section 5).

Currently, HECOM employs an absolute risk model to distribute cancer fatalities over time; this is consistent with the CRAC2 methodology. HECOM is designed to accomodate a relative risk model option, that has not yet been implemented.

We recommend that this option be developed.

Concerning radiation injuries, there is uncertainty regarding the sensi-tivity of both mortality rates and costs to variations in the level of medical care provided. The question arises partly from the Reactor Safety Study's (NRC 1975) suggestion that the lethality of radiation exposure can be avoided to an extent by sufficiently intensive levels of medical care. Currently HECOM applies the cost of relatively intensive care in a well-equipped medical center to all bone marrow and gastrointestinal injuries. However, it does not treat the costs or the mortality implications of either minimal or heroic treat-ment. Emergency plaaning efforts would benefit from examination of the cost effects that would stem from the difference in mortality rates associated with various types of medical care?

An effort to assess the costs of the principal diseases associated with mutation would entail first the identification of those diseases and second the gathering of relevant cost data. To distribute genetic diseases according to severity would be a simpler task that could employ, perhaps, a panel of experts.

Changes in the estimation of particular health effect costs as discussed in the preceding paragraphs would add increased precision to HECOM. Regardless of whether those changes are made, an importot next step is the application of the model to examples of hypothetical reactor accidents. The current output from the model shows the richness of information that can be obtained. Appli-cation of the model may be expected, in addition, to lend a new empirical basis to the enduring policy question concerning the potential costs associated with irradiation.

Aside from improvements to and application of the current model, benefi-cial advances could be made in the valuation of risk by further conceptual and empirical work toward the development of a contingent market study of the pub-lic's risk valuation. The ideal approach to estimation of the value of a change in risk is to measure individuals' willingness to excnange income for that risk change. A carefully designed contingent market survey can provide 2.4 m -- .- . _ _ - , _

.:t y

information about individuals' preferences toward nuclear risk;~ from a rigorous

. theoretical perspective, such information about individual valuation is most

- appropriate in measuring the benefits of risk. reduction.

9 l'

e D

4 2.5 J

, , - - , - , . . . , y - ,, . - . - - ,--.,,,,--~...,,_,,r-,e , -

,w,,.., + . . - -., ,7--n. , - - , + - . - , -r,~4',<

If 9

2.0 REFERENCES

Cooper, D. W. et al. 1982. Reactor Safety Study Radiological Health Effects Model: Critical Review. Sandia National Laboratories, Albuquerque, New Mexico.

U.S. Nuclear Regulatory Comission. 1975. Reactor Safety Study. Appendix-VI. WASH 1400, National Technical Information Service, Springfield, Vi rginia.

e rl 2.6

l l

-1

_ 3.0. REVIEW OF HEALTH EFFECT PROJECTIONS In this chapter we review the information about the incidence of health effects that.provides the bases for cost estimates. This includes experimental and epidemiological studies of dose and effect relationships, information on

- the clinical symptoms associated with each type of illness',. and the treatments likely to be requirad for each. Radiation injuries are discussed in Section 3.1, cancers in Section 3.2 and genetic effects in Section 3.3.

3.1 RADIATION INJURY INCIDENCE AND TREATMENT 10epending_on dose levels ~and on individual sensitivitie's, exposuretto significant amounts of radiation may result almost immediately in acute symp-toms that could range from nausea to death. Treatment required for recovery may. range from a few days of-bed rest at home to heroic intervention in a well-equipped regional medical center. It is convenient to consider the range of possible acute effects by grouping radiation injuries into three categories:

1) prodromal symptoms, which.last only a few days; 2) bone marrow syndrome, gastrnintestinal syndrome, and pulmonary impairment, which are all potentially life-threatening; and 3)'in-utero effects, which cause severe and permanent impairment _ to the irradiated fetus. In this section we provide a review of how each category of injury relates to radiation dosage and how the clinical signs of the injury are likely to progress. We also suggest parallels with more conunon diseases in. order to estimate the . levels of treatment that may be involved for each injury category.

3.1.1 Prodromal' Symptoms Prodromal symptoms may include nausea, loss of appetite, headache,' diar-rhea, and weakness. The higher the radiation dose.and the shorter the time

- ~

over which. exposure oc:urs, the sooner these symptoms occur and the longer they persist (Blakely 1968, p. 35 and NRC 1975, p. F-13). Blakely (1968, p. 35) reports that prodromal symptoms may occur occasionally after a dose as low as 50 rads, but are more likely at 100 rads and'are seen in all cases at- 200 rads and'above.

Prodromal symptoms may be treated like a case of the flu, and are not serious in themselves, except perhaps for the very young, the old, and those with recent illness or injury (Dalrymple 1973, p.191). The. appearance of prodromal symptoms,-however, serves to identify persons who.may have received sufficient exposure to result .in more serious radiation injuries, such as bone marrow syndrome. Because closely monitoring prodromal-symptoms is the only way to detect the existence of serious injury, we assume that people would be treated as though seriously injured until evidence develops to the contrary.

- Such treatment could involve two or three days of hospitalization, with the administration of fluids and medications and the performance of numerous labo-ratory tests.

3.1 4

w .~,,,.,,-+w-, -

-,,+~-.,-.-r,,,.,,,.,w- , - .n , , , , - - , - , , <w-.w.-,-,-m-a-.,-,-+,-,-..,,.-,-.w-rr-, - , . ,w,w-ew-,..rw.--- = , r - - + --, -

. Following the prodromal symptoms there is a latent period before the mani-festation of more serious injury. The duration of the latency per.iod varies

' inversely with the dose rate. Table 3.1 provides a summary .of the ' progression of acute radiation .symptons for various whole body dose levels. In the less serious cases this latency period lasts from 1-3 weeks,-during which time the

_ individual may experience weakness and . fatigue and should have both mental and

. physical rest'to minimi_ze the severity of the hemorrhage and infection that may follow (Blakely 1968, p. 50). This is a time when preparations can be made at.

. regional medical centers for, the treatment of severe cases and a time when patients can be transported to centers with adequate facilities. The cost L estimates developed in this study assume that facilities are available

-locally. LIf unusual efforts were required to deliver medical care, the costs

.could-be substantially higher.

3.1.2 Bone Marrow Syndrome

. Failure of the bone marrow system would be the primary cause, of serious illness or death as a result of radiation exposure in a reactor accident.

Blakely (1968, p. 37) places the' lower threshold for bone marrow syndrome at about 200 rads, with milder manifestations resulting from doses between 200 L'd 400 rads and severe symptoms at doses between 400 and 600 rads.

~

The Reactor Safety. Study (NRC 1975, Appendix VI pp. F F-3) presents

-dose-response curves for bone marrow damage depending on the extent of medical intervention. That . study predicts 50 percent of the people exposed to 340 rads would: die .within 60 days if.they were given only minimal treatment. With sup-portive medical treatment, the estimate is that 510 rads would be a lethal dose within 60 days to 50 percent'of those exposed. Supportive treatment is.

described later in this section.' With heroic treatment the-report asserts that i the 50 percent lethal dosage may be as high as 1050 rads for whole-body expo-i sure. Heroic treatment would involve bone marrow transplantation. We consider

, transplants to be an unlikely form of treatment because of the difficulties of l

finding a compatible donor for most patients, a problem that may be accentuated in the aftermath of a . reactor accident. In addition, at least one researcher p ( Andrews 1980) advises that marrow transplant may not be helpful.

L l Bone marrow syndrome is characterized by impairment of the blood forming h system,,with the degree of impairment depending on the dose. The clinical manifestations include severe susceptibility to infection, hemorrhage, and anemia. Treatment is centered around keeping the patient free from complica-tions until bone marrow function is regained. Supportive treatment involves sterile isolation, controlling infection by employing special air filtration

. systems and sterilizing everything that comes into the room ( Andrews 1980, p.

306;- Blakely 1968, p. 61; NRC 1975, Appendix VI p. 9-3). Administration of antibiotics is prescribed (Saenger 1982; NRC 1975), as well as continual moni-toring with. laboratory tests ( Andrews 1980; Saenger 1982) and use of blood transfusions.

For purposes of outlining the probable course of treatment and its costs, we suggest there are relevant similarities between the characteristics of bone

' marrow syndrome and those of burn trauma. Both are potentially lethal threats, 3.2 I

..,,_ _ n._ . . u ._ ._.~_.___==__._n-.__..__..._

TABLE 3.1. Clinical Progression of Acute Radiation Syndrome (a)

' Approximate-Dose Levels (whole-body rads) Clinical Progression 50-200 prodromal + recovery symptoms.

200-400 prodromal + latency. + mild bone + most symptoms 2-3 weeks marrow crisis recover 2-3 weeks 400-600 prodromal + latency + severe bone + about 50 symptoms I week marrow crisis percent 4-6 weeks recover (D) 600-1000 - prodrcmal + latency + gastrointes- + probable symptoms few days tinal . injury death 1-2 weeks 1000's - death within hours lfrom cerebrovascular crisis E (a) Blakely (1968) describes a similar pattern of disease progression, except that he predicts near-100 percent mortality at 600 rads, with cerebrovas-cular crisis occurring at around 1400- rads. NRC (1975) considers dose ranges from 350 to 550 rads as critical for the bone marrow. That study distinguishes between whole body doses and locus-specific doses to gastro-intestinal tract and lungs. NRC 1975 (Appendix VI p. I-7) suggests that in the absence of bone marrow complications mortality fran gastro-intestinal injury.alone would not occur below 1000 rads.

(b) 'At dose levels of about 450-500 rads 50 percent of the exposed population are expected to die within 60 days even with supportive treatment. At 600 rads the death rate may be close to 100 percent without heroic interven-tion. - NRC (1975) suggests that 50 percent could survive' whole body doses as high as 1050 rads with heroic treatment (i.e., with a bone marrow transplant) .

with infection as the immediate concern. In addition, a possibility of severe

, hemorrhage is present in either condition. Because of the clinical similari-ties, we assume that.the services involved in the provision of " supportive treatment" are similar to those given a nonsurgical burn trauma patient. The costs of.such services are estimated in Section 5.1.2.

3.1.3 Gastrointestinal Syndrome At whole-body doses over approximately 600 rads the symptoms of gastro-

- intestinal syndrome are likely to precede those of bone marrow damage (Saenger

' 1982;' Blakely 1968). The onset of symptoms-comes after a shorter latency 3.3

-ry v , - , - - -.,,-eiw-2,. .~ w e. - - + , ,,m.-w-<,,---w,- y ---w-www.m , ,,w s-,w,r-.+w-+-w~w. .,,vrwry--,+,s , w- -,w---

period than at lower doses (a few days to a week). The symptoms include vomit-

ing and diarrhea of a. severity that is qualitatively different from that exper-ienced in the prodromal phase. Death is probable within a week or two of expo-sure (Blakely 1968, p. 41). For local irradiation of the gastrointestinal tract without a high whole-body dose, the lethal dose may be closer to 3500 rads (NRC 1975).

For either local or whole-body irradiation, treatment involves the replacement of fluids and electrolytes. Such treatment may keep the patient alive long enough for healing of the intestinal lining (Blakely 1968, p. 41).

However, recovery will result in the patient facing severe bone marrow syndrome a short time later. Because of this threat of bone marrow syndrome in patients who survive the gastrointestinal problems, we assume that gastrointestinal patients would be treated from the start in the same isolation prescribed for bone marrow patients.

3.1.4~ pulmonary Impairment Pulmonary impairment can be expected in approximately five percent of cases after inhalation doses of 3000 rads and in 100 percent after inhalation of 6000 rads (NRC 1975, Appendix VI p. F-6). Depending on the source of the radioactivity,100 percent mortality can be expected from lung doses of 15,000 to 30,000 rads. Although it is possible to receive that high an inhalation dose with relatively low whole body doses, at any given distance from the reactor the probability of death from lung dose would always be substantially lower than that from the associated bone marrow dose (NRC 1975, Appendix VI p.

9-5).

Symptoms of pulmonary injury include pneumonitis and pulmonary fibrosis.

In the absence of bone marrow syndrome, we assume these symptoms could be treated in an average hospital room.

3.1.5 In-Utero Injury A category qualitatively different from other radiation injuries is in-utero or prenatal effects. Injuries and deaths would be due mainly to irradia-tion during the second trimester of pregnancy, with spontaneous abortion likely for embryos in earlier gestation. The nervous system is particularly sensitive

- to injury and effects such as growth impairment, microcephaly and mental retar-dation have been observed at doses as low as 10 to 20 rads (NRC 1975, Appendix VI p. F-18) . Microcephaly, which is generally associated with severe retarda-l tion, occurred in about 50 percent of fetuses exposed to 150 rads as a result of atomic bomb exposures (p. F-36). Using information about the age structure of the potentially exposed population and dose rates, the nurnber of in-utero injuries can be estimated, though it is not by CRAC2.

Long-term institutionalization may be required for individuals irradiated in utero. The care provided may be similar to that given to individuals who are severely affected by Down's Syndrome or spina bifida. For lack of informa-tion specific to in-utero radiation injuries, we rely on the probable similari-ties with those two other prenatal-onset diseases with long-term impairment to guide our cost estimates.

3.4

___ -_ _ _ _ r _ _ _

n_ .. _ _ _ _ _ _ - ____ _ __ _ . . _ _

d 3.1.6 ~ 'Other' Radiation Injuries

.There are' other possible forms of injury from irradiation that are of less.

concern than those outlined above, either.because they cause relatively minor problems or because they become serious only at doses high enough to preclude probable survival: .

• . Hypothyroidism - This Lis an impairment of thyroid function that can

, .be induced by raajation exposure. Oral medication is effective and-inexpensive (NRC 1975, Appendix VI, p. 9-13).

e Sterility - Radiation-induced sterility may be either temporary or permanent. Males may have temporary effects at lower doses than females but require higher doses for permanent ' effects. Permanent sterility, in males or females, is unlikely below doses that are life-threatening if whole body exposure is involved (NRC 1975, Appen-dix VI, p. 9-15).

e Cataracts - Ooses of 200 to 500 rads to the lens of the eye may result in formation of cataracts after a latency period that varies .

with both dose and dose rate (NRC11975, Appendix VI, p. 9-18).

e Skin .and Hair-Damage - Loss of hair occurs two to three weeks after external doses in excess of 300 rads. This is likely to be temporary unless the dose . exceeds.600 rads (NRC 1975, Appendix VI p. F-13).

LThe skin may also be affected by doses in this range, resulting in radiation ~ dermatiti s. This condition has levels of severity compar-able to first, second and third degree thermal burns and in the most severe cases (due to doses of over 2000 rads) can result in permanent s skin ulceration (Prasad 1974, p. 240-248). Survivable whole body doses are unlikely'to cause more severe injuries than hair loss and

skin reddening.

p 4

3.1.7 The CRAC2 Projections of Radiation Injuries The CRAC2 output includes estimates of early fatalities and injuries-i.e., those occurring within one year. of accidental radiation exposure. fin actuality, most of these effects would occur within the first three months.)

For exposures of less than 1000 rads, which includes most hypothetical accident scenarios, the primary cause of early fatalities would be dose to the bone marrow. In some cases, however, pulmonary exposure could also be instrumental in inducing mortality. To estimate fatalities the CRAC2 computer code calcu-

.lates-population exposures and then applies a probabilistic fatality rate to

-the estimated exposure level of each segment of the population. The dose and associated mortality rates used in these calculations are shown in Table 3.2.

The methodology used is documented in the Reactor Safety Study (NRC 1975, Appendix VI) and in the CRAC2 user's manual (Sandia 1981). Mortality rates for dose levels between those listed are developed within the model by linear interpolation. Early fatalities, as estimated by the CRAC2 model, are the r,um 3.5 5.

, CA_ , .; - , - ,, ,f % ;- ,, . . , . . , ,,. . J . Jr . 1, . .i.-,,__,,,.,-., .

-,__,-.-~_,..-,m. _ _ ,

~

TABLE 3.2. Dose Values and. Associated Mortality Rates Used in CRAC2 (Sandia 1981)

Mortality Organ Dose (rem) Rate

- Bone Marrow 320 0 400 .03 510 .5 615 1.00 Small Intestine 2000 0 Lining- ' 5000 1.00 Lung 5000 0-14,800 .24 22,400 .73 24,000 1.00 of probhble fatalities for the entire exposed population; double counting of fatalities due to multiple fatal organ doses is avoided in the model.

CRAC2 use of the mortality rates shown in Table 3.2 is based on the assumption that .alliof the injured would receive a level of medical treatment designated as " supportive" by the Reactor Safety Study (Appendix VI, p. 9-3).

Unfortunately, an estimate of the total number of people who would require this treatment is.not available from the CRAC2 output. While the fatalities are

. counted,: the survivors.of bone marrow exposure are not explicitly included in the category of "early injuries" and their number cannot be derived from the number of fatalities. It would be advantageous to indicate the population

' receiving doses within 100 rem intervals, so that cost estimates could be linked to the severity of the injuries.

Injuries evident in the immediate post-accident period are calculated by the CRAC2 model from the information in Table 3.3. As in calculating fatali-ties, the injury rate is applied to the population projected to have received each dose level and the resulting estimates are summed. The threshold for injuries-is approximately 50 rads. Injury rates at intermediate dose levels are derived by. linear interpolation within the model. At the levels of pos-sible doses to offsite ' population developed in most accident scenarios, it is whole-body dose that is primarily responsible for injuries.

People receiving whole-body doses above 50 rads may experience prodromal symptoms such as nausea, vomiting, anorexia and diarrhea within a few hours of exposure and continuing for a. day or two. While CRAC2 calculates the number of people likely to experience actual prodromal symptoms, it does not provide any indication of the number likely to require medical care. As noted by Dalrymple (1973, p.'192), people in the vicinity of an accident may experience circula-tory system or gastrointestinal system symptoms that are due to anxiety rather 3.6 o-- -

e - g-t

• g s -vw-H7+g W-98t' g -+wipv,-yg e y'w--v-g- gW-mbF pp- y- yP gm*wr- p wywi-y-w a- ev- * -p- gw*,s -wwwv-

i TABLE 3.3. Dose Values and Associated Morbidity Rates Used in CRAC2 (Sandia 1981)

Morbidity Organ Oose (rem) Rate Whole Body 55 0 150 .3 280 .8 370 1.00 Lung 3000 0 3000.1 .05' 6000 1.00 Small Intestine 1000 0 Lining 1000.1 .05 2500 1.00 than radiation exposure. Thus, both injured and uninjured individuals may initially experience identical symptoms. In the event of an accident where the occurrence of significant population exposures is suspected, a major population screening and treatment effort would be required. The number of people who would require treatment for prodromal symptoms and screening for more severe injuries would be at least as large as the number of early injuries calculated by the CRAC2 code. There is a high probability that the actual number would be substantially larger.

The present form of CRAC2 output for early injuries is ill-suited to pro -

jection of direct costs. Only an aggregate measure of early injuries is avail-able, on,e that includes transient, prodromal symptoms on the same basis as

. life-threatening pulmonary and gastrointestinal effects and that omits bone marrow injuries. . Major types of potential injuries and their status in the CRAC2 calculation are shown in Table 3.4. If those effects that are included in the CRAC2 calculations were available by organ (e.g., lower intestine lin-ing), the estimates could be used directly in calculating costs. No technical reason for the exclusion of bone marrow syndrome from the estimate of early injuries has been identified. CRAC2 modifications required to calculate num-bers of bone marrow injuries are discussed in.Section 7.3.

There is an additional category of health effects that is omitted from the CRAC2 calculations but which may have substantial impacts. That is in-utero fatalities and injuries. An analysis of the numbers of fatalities potentially involved indicated that " embryonic and fetal deaths would be fewer than 10 to 5 percent, respectively, of the early fatalities..." .(NRC 1975, p. 9-11). The rationale given for excluding them from reported early fatalities is that the embryonic (first trimester) deaths would not be noticed and the fetal (second and third trimesters) deaths fall within the range of uncertainty of the CRAC estimates.

3.7

. - - - w < , .,.w,- -,,m - . - . _ . , - , , - , , - - - , - , .,,,,,,-,,-,v---,,v ..-e,-,.n--e-,-- w m - ~ . - - - * - . ,m-- -,

TABLE 3.4 Summary of Early Injury-Related Information Major Included Duration of Injuries in CRAC2 Acute Categories Estimate Symptoms Prodromal symptoms yes 2 days Bone marrow syndrome no 4 to 8 weeks Gastrointestinal syndrome yes (a)

Lung effects yes 1 year (b)

'In-utero injuries no lifetime (a) Patients who die generally do so within 10 to 14 days. No estimate of the recovery period was noted in the literature but it is likely to be several months.

(b) No information on treatment or likely length,or recovery period was found.

While projection of fetal injuries would not have much effect on the total number of early injuries calculated by CRAC2, it is important in the calcula-tion of accident costs since these injuries are the most costly type of health effect. (See the discussion of the in-utero injury treatment costs in Sec-tion 5.1.5). Sufficient information to project in-utero injuries is available from the Reactor Safety Study (NRC 1975) and other sources.

3.2 CANCER INCIDENCE AND TREATMENT There is wide consensus among scientists that an association exists between ionizing radiation and cancer. In fact, scientists may know more about the carcinogenic effects of ionizing radiation than about those of any other environmental agent (Land 1980). Nevertheless, there is considerable uncer-tainty regarding dose-effect relationships, to the extent, as Land (1980, p.

1197) reports, that scientists contributing to BEIR 80 differed by as much as a factor of 100 in their assessment of the risk from exposures to a single rad of ionizing radiation. Because there are basic disagreements about central fea-tures of the techniques used to estimate dose-effect relationships, and because scientific knowledge is rapidly changing concerning the risks from radiation, there are several issues to be raised pertinent to the CRAC2 estimates of cancer effects. In this section we do not attempt to provide resolution of those issues, but rather to explain how reasonable estimates may vary from those used as inputs in this study.

In regard to estimates of incidence, there are reasons to suggest that the CRAC2 estimates may be too high, and other reasons why they may be too low. In addition to questions of dose-effect relationships, changes in treatment may also have an important influence on the cost estimates provided by this study. Questions are raised relevant both to incidence and to treatment.

i 3.8

,i

- 3.2.1. Incidence Assumptions There are at least two general issues of current concern in regard to cancer. incidence estimates,' each relevant to the Reactor Safety Study (NRC 1975) and CRAC2 projections.: First, dosimetry. data and incidence esti-mates for the Japanese atomic bomb casualties-have come into question. Se'cond , -t uncertainty about the. shape of the dose-response curve may have an important  !

impact-on the estimates of responses to low-dose radiation.

4

-The issue regarding tihe accuracy of dosimetry data- for the Japanese A-bomb

, casualties is central to dose-effect estimates because BEIR 72 and BEIR 80 (and- ,

therefore .the . Reactor Safety Study and CRAC2) base their projections of inci-

- dance on the Japanese data. Each of those incidence projections employs dosi- i metry estimates computed in 1965 and labeled " temporary" (the "T65" dose). .

Further study now suggests that the neutron component.of the Hiroshima bomb may i have'been -lower than previously' calculated, with some corresponding increase in the gamma component (Loewe 1981). The net result may be that some risk esti-mates will be double ( (Beebe 1981).

l In addition to changes in dose estimates, other new information on the Japanese casualties suggests that cancer incidence and related mortalities may be higher than previously estimated (Wakabayashi et al.1983). Consideration of the new estimates reinforces a conclusion 'that earlier incidence estimates

Dased on the Japanese data may be significantly too low.

Unlike the new Japanese data that suggest current dose-effect estimates ,

are too. low, the dual problems of inadequate sampie size and uncertainty regarding the. shape of the dose-response function result in an ambiguous con-clusion that current estimates could be either too high or too low. A problem

arises in estimating the effects of low-level radiation because such an esti-mate requires a study with very large sample size. Land (1980, p. 1197)

. describes the problem with an example: "If the excess risk is proportional to

- dose, and if a sample of 1000 persons is necessary to determine the effect of a 100-rad exposure, a sample of 100,000 may be needed. for a 10-rad exposure, and about 10 million for 1 rad." The Japanese Life Span Study sample includes data ,

, on 110,000 people, some from. Hiroshima and some from Nagasaki, with exposure to

a very different mix of radiation types in the two cities. While the sample may be adequate for projection of high-dose effects, it is unlikely that the Japanese data can provide estimates of risk in the low-dose region except with i

. the assumption of a specific dose-response function (Beebe 1981). Since the sample is too small and too diverse to derive estimates at low doses, the exper.ience at high doses must be extrapolated to obtain low-dose estimates.

The critical question is, on what basis should the extrapolation be made: i s the dose-response function linear or of some curvilinear form?

- Extrapolation assuming a linear dose-response curve may overestimate low-

. dose responses, if the true function actually curves up more steeply at high doses. - Conversely, adjustments that imply a curvilinear (positive second deri-

' vative) dose response curve may cause an underestimate of the response if the function is linear.

3.9

'f v --,,e--.,-p v s -9,-3,. ' , .-em,e<. .-- , en,. , - , _ , .w--w.,,-e,--e ' w we,w-+-+w.,-e,-

- +,m, w w w. -vvmm+----,,, -e-w+ .+.--eyvew,~ -we-w,-,w--- +

,. . _ __ .._. 4 .-. _ _ . -. _ _

- BEIR 72 assumed a linear dose-response function for all types of cancer.

BEIR-80 subsequently asserted a curvilinear (linear-quadratic) dose-response function for all cancers,'against a dissent from the Committee Chair (BEIR 80, -

pp.~227-253) who argued for. a linear form. Beebe (1981, pp. 780-781) supports the use of the ' linear form for its ease of ' application and its interpretability as an upper boundary. . Land (1980, p.1202) observes that the linear model appears to overestimate: leukemic. effects.of low-dose radiation, although it fits reasonably well the evidence of breast cancers associated with low doses.

Basing its " upper. bound" estimate on the lin. ear extrapolations of BEIR.72,

' the Reactor Safety Study offers a " central estimate" for all cancers other than

- breast cancer to' account for "the ameliorating effects of dose protraction and the lesser . effectiveness of very small ' acute doses." The central estimate is r

- not a representation of a curvilinear dose-response function but in modifying the linear function it has a'similar'effect. Cooper et al. (1982, p. 5-4) cite more recent studies that suggest that fractionation by dose protraction may make low doses even more effective at low dose rates. Cooper et al. conclude that such studies would argue against dose reduction factors (such as used. in the Reactor Safety Study to. adjust from the BEIR 72 functions to the central estimates) . In fact, they observe that those studies support dose factors that

- would result in higher dose-effectiveness at low, protracted doses. (Cooper

et al. 1982, p. 5-4)

~

CRAC2-(Sandia 1981) employs the central estimate from the Reactor Safety Study. As discussed'above, there are some reasons to suspect that projection

- is too high, others to consider if too low. It is likely that the. central estimate adopted in CRAC2 lies within the band of uncertainty. The data and models that provide the basis for CRAC2 estimates are currently. being reviewed -

. (Cooper et al.=1982). Completion of this review is expected during 1983.

3.2.2 Treatment Assumptions Due to lack of more recent information, this study of health effect costs

- assisnes cancer treatment effectiveness to be the same today as it was in the early 1970s. . First, the estimates of cancer mortality input to the PNL model via CRAC2 are derived ' rom BEIR 72. Therefore, the estimation of fatality L costs is based on- fatality rates that do not consider any medical progress

- since 1972. Second, the direct costs of cancer treatment included in this study are based on information obtained through the Third National Cancer Sur-

. vey completed in 1974. - That information includes the recollections of survi-I vors.and of nonsurvivors' kin regarding treatment received in the early 1970s.

l In its effects on cost estimation, the assumption of unchanging treatment t

modes yields an ambiguous result. On the one hand, to the extent that medical advances have lowered cancer mortality rates since 1972, the projection of early mortalities should be adjusted downward. The indirect costs would be expected to be lower as a result of such an adjustment. On the other hand,

[; - medical advances have been obtained only with increases in real costs. Cancer treatment is more intensive than in the early seventies, and consequently the direct costs may be higher in real terms. We are unable to discern whether

[ these changes have led to increased or decreased economic costs.

3.10 e

w +9-- g e- g erc+ y 9 ,, y e- -*+-e- r y 9o9 ,--

w-*Pr " etg y - m- r-9, ge wi#tvvev v '+www '

er m ' ere*'- *w-*"&-2m-e'+ ' ' * " ~~---**w-=**-*"" * ' ' "" * - ' ' ' " ' - -

• +

" " - -v:.r

. 3.3 NATURE AND INCIDENCE OF GENETIC EFFECTS In the-1950s, as government and public attention ' focused on the possible

risks of ~ radiation, genetic _ risks were the -predominant concern (Denniston 1982). Over ' time, attention has shifted to radiation-induced cancers. This

- shift may be due partly to the perception that cancers pose a more immediate

- threat, andf also partly to the fact that science has displayed a greater facil-

. ity in quantifying cancer risks than in estimating genetic effects. -In this l section, we_ review selected relevant literature for a discussion of the diffi-culties11n: predicting radiation . induced genetic disease.

There are several enduring impediments to the estimation of the~ genetic-

' effects of _ increased radiation levels. First, evidence is weak regarding the

- linkage between radiation and genetic damage in humans. Since radiation causes

. identifiable mutations in other mammals, geneticists generally agree that radi-

~

ation can cause' harmful mutations .in humans. However, there remain difficult i questions concerning what kinds of genetic disorders may. be caused by. radiation

. and how the dose-response relationship may be quantified. Even if the effects in terms of genetic material changes are identified and quantified,- there

-remains an imposing problem of predicting the nature and severity of clinical manifestations (observable diseases). of each type of genetic damage.

3.3.1 ' Kinds of Genetic Damage Associated with Radiation

+

Among the categories of genetic damage, autosomal dominant disorders have special importance in . radiation genetics: the relationship between the muta-tion rate and birth defect frequency is relatively direct and radiation-induced increase in the mutation rate .is _ expressed most strongly in early generations (Carter.1977) . The collective incidence of autosomal dominant disorders is roughly one percent of persons born (Stevenson 1959; Carter 1977; Oftedal and -

. Searle~ 1980).1 Trimble and _ Doughty (1974) estimate.the incidence at only 0.1 percent,but-they ignore late-onset diseases.

Another category of genetic disorders that would almost directly reflect a

- radiation-induced increase in the mutation rate are X-linked disorders. These

- mutations involve genes located on the X chromosome and are expressed almost

~ exclusively in males. These disorders behave as dominants in males. Estimates of their numbers, are typically included with the dominants in a single cate-gory in estimates of radiation-induced genetic effects.- As with the dominants, these disorders appear most frequently in the early generations after a one-time increase in the mutation rate. The current incidence of X-linked disorder is approximately 0.8 per 1000 liveborn males. -(Stevenson 1959; Trimble and Doughty 1974; Carter 1977).

Unlike ainant and X-linked disorder that require the presence of only one mutant gene for their expression, autosomal recessive disorders appear only

. when two mutant genes are inherited, one from each of the parents. There is a very low probability of a newly induced recessive mutation pairing up with a F previously existing mutant allele in a way that will express a deleterious condition in the first or early generations (0ftedal and Searle 1980).

Instead, the interval between the induction of a recessive mutation in a gene 3.11

= - = : =. . . . = _- -.-:---a..-.-..---..-_...-

and the birth of an affected individual may be. centuries or even millenia ( Ash, Vennert, and Carter 1977). For that reason the category of recessive genetic disorders is usually considered to be negligibly affected by increased radia-tion (UNSCEAR 1977; BEIR 1980). In contrast, Edwards (1979) holds that the recessive disorders are particularly severe and that over a very long time they represent the " main hazard in man" (p. 467). Estimates of the current inci-dence of recessive disorders vary from 1.1 per 1000 liveborn (Trimble and Doughty 1974) through 2.1' per 1000'(Stevenson 1959) to 2.5 per 1000 (Carter 1977).

In addition to the disorders discussed thus far, all of which have unambi-guously genetic causes, there is a 1arge category of multifactorial disorders (also called irregularly inherited disorders.) These may stem partly from dominant mutations and partly from environmental causes. In order to predict their increased incidence, it is necessary to estimate their " mutation compo-nent," the proportinn of their frequency that depends on the mutation rate.

Each multifactorial disorder has its own mutation component and very little is known about these components (Denniston 1982). UNSCEAR (1977) estimates the mutatior component to be 5 percent; the BEIR-(1980) estimate is 50 percent.

Estimates of current-incidence range from 4 per 100 liveborns (Stevenson 1959) to 9 per 100 (Trimble and Doughty 1974).

In addition to the genetic mutations discussed above, ra'diation exposure may cause a bread class of chromosome anomalies. This class includes three types of disorders: numerical aberrations, rearrangements, and deletions (Den-niston 1982). The deletions may have effects indistinguishable from those of single gene mutation and thay are included among those disorders. The numeri-cal aberrations contribute heavily to'very early prenatal mortality, acc sunting for approximately 50 percent of spontaneous abortions, often so early that pregnancy is undetected (Denniston 1982, p. 331). They also result in genetic diseases such as Down's syndrome, Turner's syndrome, and Klinefelter's syndrome (Denniston p. 331). As a class, chromosome anomalies lead to impairment in approximately 0.6 percent of liveborns,'according to Denniston (p. 331).

3.3.2 ' Estimation Methods

(

There are two principal ways to estimate the effect of increased radiation dosages,in terms of the incidence of genetic and chromosomal disorders. Both l involve extrapolation to humans from experience with irradiated mice and other l mamai s.

The. doubling dose method is based on the equation (Dennison 1982) l Induced burden per rad = so " " x mutational component d n d The spontaneous burden is estimated from human population studies such as Stevenson (1959), Trimble and Doughty (1974) and Carter (1977), as reported above for each type of genetic disorder. The mutational component is the part of the existing burden expected to increase in proportion to the mutation rate. It is 100 percent for autosomal dominant disorders and open to question l 3.12

for others. The doubling dose itself is calculated f' rom nonhuman data, gener-

-ally from the mouse. The increase in dominant disorders in humans is estimated from the' induced mutation rates of recessive genes in tihe mouse. Each of the variables in the doubling dose equation is dependent on interpretation of evi-dence that permits widely divergent estimates. _

The direct method of dose-response estimation rel.ies on skeletal structure anomalies in the ' offspring of irradiated mice. This method requires extrapola-

.. tion of skeletal effect rates to other body systems ar'd then projection of the experimental findings in mice to effects in humans. The method also calls for adjustment by various " correction factors" to compensate for high dose rates and for fractionation and to estimate a total population inc dence from experi- 4

'nentation with males alone (Denniston 1982).

3.3.3 The Risk Estimates -

There are three major studies of primary relevance to the estimation of radiation-induced genetic disorders, the two. reports from the National Academy of Sciences Committees, BEIR I in 1972'and BEIR III in 1980, and one from a United Nations Committee, UNSCEAR, in 1977. Table 3.5 shows these committees' estimates for an average population exposure of 1 rad. Estimates are given both for the first generation following exposure and for equilibrium, which is the level at which, after several generations, the incidence rate would level off and be sustained if there were no further changes in exposure (i.e., a new steady state.)

TABLE 3.5. Estimated Increase in Genetic Disorders per tiillion Liveborn, from an Average Population Exposure of One Rad -

Ournant BEIR 72 UNSCEAR 77 BEIR 80 bisease Type Incidence 1st(a) Ea(a) 1st En 1st Ea Dominant and X-linked 10,000 10-100 50-500 20 100 5-65 40-200 R!cessives 2,500 slight very slow slight slow very few slow increase increase incntase thbalanced 4,000 12 15 40 40 <10 increase only Rearrangments slightly Aneuploids 1 1 -- --

0 0 Irrt9:larly Inherited 90,000 1-100(b) 10-1,000(b)

- Disorders J 45 - 20-900 Totals 106,500 25-215 75-1,500 65 185 - 60-1,100 (a) First gener3 tion; equilibriun.

(b) Used a curren incidence of 40,000.

Soun:e: Adapted frun Dennistan 1982.

3.13

, ..~ .. - . . . . .-

7 . . _ . -

1Both UNSCEAR =1977 and BEIR.1980 employed the doubling dose method for estimation of single gene effects'in equilibrium,..and the direct method for-

,  : first' generation estimates. UNSCEAR used the doubling dose method for esti-

, - mating. chromosomal ' rearrangements, while BEIR.1980 used human and marmoset data for direct: estimation (That is the reason for the divergence in the connittees' estimates regarding rearrangements.) BEIR 1972 employed the direct method for estimating first generation incidence of induced chromosomal: aberration, but

. used 'a doubling dose method throughout for gene mutation.

There are additional reasonsLwhy the estimates vary. ' We concentrate here' on' the differences between UNSCEAR 1977 and BEIR.1980 as reflections of the

!  ; current sta'e of the art.- The UNSCEAR Committee -accepted a dou'oling dose of

- 100 rad;~ B'.iR 1980 considered it ~ to be in the range of 50-250 rem (Selby

' 1979). BEIR 1972 had placed it .in the range of 20 to 200 rem using the direct

method to forecast the effects in the first generation. The committees used different estimates of both the mouse-human relationship and the skeleton-whole

- body relationship. The UNSCEAR Committee accepted an estimate t".at about one-half- of the dominant mutations found in mice would cause se'rious disorders:if

^

.found in humans; the BEIR Committee felt the true range to be from one-quarter to three-quarters of the mouse disorders. The UNSCEAR Committee multiplied the

' skeletal. disorders by.five.to estimate the whole body effects; .BEIR preferred a range from five to 15.

Scientific interpretation causes estimates of increased genetic disorders

- to vary even though, as'Denniston (1982) observes, tne UNSCEAR and BEIR Com-mi.ttees-have overlapping memberships and they used the same data (p. 332) . In order to estimate genetic effects in terms of their clinical manifestations instead of as'_ genetic disorders, a. further interpretive step must be taken.

3.3.4 Clinical-Manifestations of Genetic Disorders Stevenson (1959), Trimble and Doughty (1974) and Carter (1977) all provide lists of clinical diseases classified according to category of genetic dis-order. Those lists'are usually employed in the calculation of genetic effects: they provide the estimates of current incidence to which the doubling J dose _is ' applied. In this study we 6re interested in the clinical manifesta-tions' as final' outcomes, as the observable, impact-producing health effects related to radiation-induced genetic damage. ~ It is the effects of inherited

' disorders . such as blindness, muscular dystrophy, chorea, and kidney disease, that' produce costs for society.

To project the impact of genetic disease both the types of diseases that

may occur and their. relative frequency of occurrence must be known.. In forma-tion about the nature ~of genetic-related disease' has been expanding capidly.

~

For example, in 1966 Mckusick catalogued 169 diseases categorized as actosomal

~

dominant disorders; by 1978 a total of 736 were listed (with another 753 not

- yet fully confirmed) (Mckusick 1978). Similar growth in knowledge has occurred

- for the other types of genetic disorders as well.

Estimates of the relative frequency of various genetic diseases varies

' depending on the disease classifications used as the basis for enumeration and on the population studied. It is apparent that different populations have

~

3.14 t . . - , - m., , . .--wy.h ,. ..e,-vy.,,.--~,,~%, -e e r,ew.,-,-w.-.we v-e  %, , , ,+,,,.,-y-..,., ,,w. -is,m,--,e,,ew,w.,---,-e--+-- - , . , - ..--e,,,,

widely varying rates for some genetic diseases. Of the major population studies, Trimble and Doughty's (1974) for British Columbia probably most closely represents the U.S. population. This study could be used as the basis for identifying the relative -frequency of genetic diseases with different levels .of impact for society. At the present time, information is' unavailable as to the frequency of severe genetic diseases relative to those that create little or no cost.

3.3.5 The Reactor Safety Study and CRAC Model In the Reactor Safety Study (NRC 1975) information from BEIR 1972 is modi-

- fied to some extent, to estimate the genetic effects of a reactor accident.

Instead of documenting each step, we include here a brief discussion of the differences and similarities between the Reactor Safety Study's assertions and those of BEIR 1972 and BEIR 1980:

e 'In order to make the estimates of genetic effects comparable to the estimates of pther health effects, the Reactor Safety Study makes several computational changes from BEIR 1972: effects are calculated per million in the population, not per million liveborn; effects are calculated per rem instead of per 5-rem dose.

• BEIR 1972 employed a doubling dose in the range 20-200 rem, BEIR 1980 in the range of 50-250 rem. The Reactor Safety Study uses a point estimate of 100 rem for the doubling dose.

• The Reactor Safety Study uses the BEIR 1972 range for mutation com-ponent of multifactorial disorders: 5'to 50 percent.

In general, the Reactor Safety Study indicates there are reasons to consider-the estimates from BEIR 72 to be too high. And, as shown in Table 3.5, BEIR 1980 supports that assessment, lowering very slightly the estimates of the previous BEIR Committee.

Genetic effects have been estimated by the CRAC model, though they are

~

neither included'in the CRAC2 versio" nor documented in the user's manual (Sandia 1981).- While the discussion of ' genetic effects in the Reactor Safety Study indicates an approach to projection based on BEIR 1972, the CRAC model actugy uses a simple calculation of 260 genetic effects per million person-rem. This procedure is currently being ' revised as part of a larger NRC risk modelling effort.

3.3.6 Summary The level of uncertainty inherent in genetic-related disease projections is very high due to the major information gaps in each stage of the projection 4

(a) Conversation with Roger Blond, Division of Risk Analysis, Office of Research, NRC, April 5, 1983.

3.15

process. Given the state-of-the-art and the recent rapid expansion of informa-tion regarding genetic disease, PNL currently uses the following assumptions to provide a basis for genetic effect cost estimates:

1. Genetic effects are expressed within ten generations.

2. Half of all effects are due to autosomal dominant and ha are due to multifactorial genetic disorders and chromosomal _ damage.

3. Autosomal dominant disorders are eliminated from the population-at a rate of 20 percent per generation and multifactorial (and chro-mosomal) disorders at a rate of 10 percent (NRC 1975, Appendix VI,

p. 9-30).

4. Genetic diseases are equally distributed between those most disabling and those that have little or no impact.ges that are Because advances in the state-of-the-art are expected, PNL's cost model (HECOM) has been designed for ease of modification of these assumptions regarding gene-tic effects incidence.

(a) This is based on the midpoint of the range of uncertainty regarding incidence of multifactorial disorders (NRC 1975, Appendix VI, p. I-11).

(b) This assumption is made in the absence of an empirical information base.

3.16

3.0. REFERENCES Andrews, G. A.- 1980. "The Medical Management of Accidental Total-body Irradi-ation. In The Medical Basis for Radiation Accidant Preparedness. K. F.

HDbner and S. A. Fry (eds.) Elsevier/Nortn-Holland, New York, New York pp.

297-311.

Ash, P., J. Vennart, and C. Carter. 1977. "The Incidence of Hereditary Disease in Man." Journal. of Medical Genetics Vol. 14 pp. 305-306.

Beebe, G. W. 1981. "The Atomic Bomb Survivors and the Problem of Low-dose Radiation Effects." American Journal of Epidemiology. Vol. 114, 6.6, pp. 761-783.

Blakely, J. 1968. The Care of Radiation Casualties. Charles C. Thomas, Springfield, Illinois.

Carter, C. O. 1977. "Monogenic Di sorders." Journal of Medical Genetics.

Vol. 14, pp. 316-320.

Committee on the Biological Effects of Ionizing Radiation. 1972 and 1980. The Effects en Pooulations of Exposure to Low Levels of Ionizing Radiatf on.

National Academy of Sciences, Washington, D.C.

Cooper, D. W. et al. 1982. Reactor Safety Study Radiological Health Effects Model: Critical Review. Sandla National Laboratories, Albuquerque, New Mexico.

Dal rymple, G. V. et al ., eds. 1973. Medical Radiation Biology. W. B.

Saunders Co., Philadelphia, Pennsylvania.

Denniston, C. 1982. " Low Level Radiation and Genetic Risk Estimation in Man." Annual Review of Genetics. Vol. 16, pp. 329-355.

Edwards, J. H. 1979. "The Cost of Mutation." In Genetic Damage in Man Caused by Environmental Agents. K. Berg, ed., Academic Press, New York, New York, San Francisco, London. pp. 465-453.

Land, C. E. 1980. " Estimating Cancer Risks from Low Doses of Ionizing Radia-tion." Science. Vol. 209, pp. 1197-1203.

Loewe, W. E. 1981. " Revised Dose Estimates at Hiroshima and Nagasaki."

Health Physics. Vol. 41, No. 14, pp. 663-666.

Mckusick, V. A'. 1978. Mendelian Inheritance in Man (5th edition). Johns Hopkins Univ. Press, Baltimore, Md.

Oftedal, P. and A. G. Searle. 1980. "An Overall Genetic Risk Assessment for Radiolo91 cal Protection Purposes." Journal of Medical Genetics. Vol. 17, pp. 15-20.

3.17

Prasad, K. N. 1974 Human Radiation Biology. Harper and Row, Hagerstown, Maryland.

' Saenger, E. L. 1982. " Radiation Accident Preparedness" A course manual from the University of Cincinnati, Cincinnati, Ohio.

Sandia National Laboratories. 1981. Calculations of Reactor Accident Conse-

~

quences, Version 2, Computer Code Users Guide. Oraft SAND 81-1994 NUREG/CR-2326. Albuquerque, New Mexico.

-Selby, P. B. 1979. " Genetic Risks from Radiation: Recent Assessments by the BEIR and UNSCEAR Committees and Suggestions as to How Future Research Can Improve Such Estimates." In. Proceedings of the First International Confer-ence on Health Effects of Energy Production, Eds. N. E. Gentner and P. Unrau, pp. 115-124. Atomic Energy of Canada Limited, Chalk River, Ontario, Canada.

Stevenson, A. C. 1959. "The load of Hereditary Defects in Human Popula-tions." Radiation Research. pp. 306-325.

Trimble, B. K., and J. H. Doughty. 1974 "The Anount of Hereditary Disease in Human Populations." Annals of Human Genetics.- Vol. 38, pp. 199-223.

United Nations Scientific Committee on the Effects of Atomic Radiation.

1977. Sources and Effects of Ionizing Radiation. Report to the General Assembly, United Nations, New York, New York.

U.S. Nuclear Regulatory Commission. 1975. Reactor Safety Study. Appendix VI. WASH-1400, National Technical In formation Service, Springfield, Virginia.

Wakabayaski , T., et al . 1983. " Studies of the Mortality of A-Bomb Survivors, Report 7." Radiation Research Vol. 93, pp. 112-146.

l r

i. -

3.18 l

4.0 VALUING CHANGES ~IN HEALTH RISKS j Among the risks of exposure to acute radiation doses are increased illness and .a lowered life expectancy. That.is, compared with statistical norms, an

. exposed population faces a risk of more morbidity and of excess (i.e., earlier)

. mortality. People are generally averse to risk: a ' decrease in risk is consi-dered a good, and to be sought; an increase in risk is a bad, and to be avoided. Concentrating, for simplicity, on the issue of excess mortality, this

-section provides a discussion of the difficult problem of evaluating (in dollar terms) the cost of an increase in risk.

It is useful to begin the discussion by emphasizing that the effort here is to evaluate an incremental change in risk, not to put a value on human life. - Two general approaches have been followed to measure the cost of .

increased risk: measuring individual preferences and measuring the risk to the value of human capital. A descriptioq of each of these general approaches follows, along with an analysis of how comprehensive each is in terms of cap-turing each of the components of the cost of risk.

There are at least five reasons why someone would prefer a lower societal risk of mortality to a higher one.

The first three stem from valuing life per le: e

1. If lower societal risk means he himself is at lower risk, he prefers that state 'of lower risk. Call the value of his preference in regard to his own. life 1 .

2. If lower. societal risk means his loved ones are at lower risk, he prefers that state. Call the value of his preference in regard to loved ones 1 3, Even if neither'he nor his loved ones benefit, he prefers a lower risk for other (anonymous) people purely out of beneficence. Call that d.

Aside fr'om beneficence or valuing life per se;

4. He would value a lower risk to anonymous others because it means a lower risk to his claim on their net production. Call that 1 5 He prefers lower risk because he values the resources that would otherwise be consumed in treating illness or in trying to avoid death. Call that j .

These five components of the value society places on changes in risk levels are employed in the following discussion of risk valuation methods.

l They are used to illustrate the extent to which each method captures the major aspects of society's valuation of changes in risk.

l l

t l

4.1 g p .--.p.c.- , ---. H w - wpe

A 4

4.1 THE " HUMAN CAPITAL" APPROACH Society is willing to forgo current consumption and to invest in produc-tive plant and equipment to an extent that depends on the value of the result-ing output. That is, we value physical capital in terms or the goods and ser-vices produced with it. Similarly, we may value human capital in terms of the value of goods and services produced by labor. So a risk of losing productive years of labor, through increased morbidity and lower life expectancy, is also a risk of losing the value of the goods and services produced by that labor.

Assuming that the value of the marginal product of labor is equal to the wages paid for that labor, lost wages (including the equivalent value of self-employ-ment)'are a measure of the value of health risk.

Employing the human capital approach in practice, the cost of health risk

- is computed by multiplying 'a measure of the value of human capital by tne change in the probabilistic risk of death. For example, consider an individual who expects to earn a discounted total of $100,000 over his remaining life-

. time. That expectation depends to an extent on his life expectancy: he has some discrete probability of dying in each year. The level of his expected future earnings reflects both future wage levels and the probability of death in each subsequent year. Now suppose a reactor accident imposes on that indi-vidual an increased probability of death every year in the future; now his risk-weighted expected future earnings are, say, only $90,000. Then the cost of the risk to that individual.is estimated to be $10,000 discounted to present

' value in the year of the accident.

In computing the cost of risk this approach considers both the increased level of risk and the value of the human capital at risk. This section discus-ses the several ways in which the value of human capital can be measured. Each of the principal variants to human capital valuation is discussed briefly in the following paragraphs.

One conanonly used measure of human capital is the share of each person's

- net production at risk of being lost to society, given risks of increased mor-bidity or early mortality. The value of a person's net production is the value of his or her total production (as measured by total earnings) less the value of what he or she consumes. It is a measure of the value of goods and services a person "gives" to society, over and above what he or she " takes away" through personal consumption. Weisbrod (1961) proposed this as the appropriate measure

of human capital at risk.

This " net production" measure, however, evaluates only one component of the total value at risk; it corresponds only to the value v4,of the components listed at the beginning of this section. It ignores completely the value the individual places on the risk to his or her own life. And even from society's viewpoint, it takes no account of the beneficence that makes us prefer a lower l . risk to the lives of those whose net product is negative (that is, who consume L more than they produce). ~

L l

4.2 L

l l

l r

y -o ,_- v , , , - - , . - ---,,w-- ,

9-m  %  ;,  %

- 4 ) :-

~q 4 '

L, jf .

Y{.

y-c.

y

~

K .

Y

. QA more-comprehensive a'pproach to human capital valuation measures the' ivalue of total production,' including personal- consumption. lThis is the fapproach.taken in cost 70f-'illnessXstudies over the past twenty years by Rice -

'andfnerfdssociates -(Rice 1966; Cooper and Rice 1976; Hodgson and Rice 1982). -~

These3 studies compute the .value offhuman capital 3f rom the total average earned

-income for a~ person in ancage,and isex cohort at risk.: The total value of a Llossifrom early mortality is measured over the period between; the -age at death

• and < the, year of normal :11fe expectancy, and is Lequali to the present value of

~

- the" stream of lost. earnings (i.e.,slost ' production) . For Rice's purpose of:

" ' estimating the annual: cost,of illness- it is" appropriate to discount this stream to .its;present value in'the year. of; death; to apply such costs to a decision

. that'affects risk-(e.g., the risk ofla reactor accident), it.is appropriate to'

'discountjto present valueiin the year;in which resources would be committed. .

,The latter; approach is followed.by Hartunian, Smart and Thompson (1981) and by PNL in thijs. study.

- ^

s t

. Furthk refinements are often made to both the total and net production .

= measures.nespecially to account in:different ways for the human capital of the

' nonwage-earning population.;Since available data on earnings exclude values ,

~

for nonmarket) production,1the value ofih'ousehold services, for example,;must be-

' imputed if. the. value of. women',s (and some. men's) production is not:to be sig-nificantly ' understated. J(This is also true for other types-of nonwage earning

.- .laborjbut datatto-carry it out are lacking.) Imputed values may be' based on-the market valine'of domestic services -(Brody 1975) or. on the opportunity cost- '

! . principle, accounting for wages that could'be earned:in the marketplace as an

, ~a lternative.useiof the homemaker's productive time (Prest and Turvey 1965). A problem with the;1atter approach is that it is difficult to determine likely

. wages that'could be earned in the marketplace if a large number of homemakers,,

i not currently in7the' labor. mark't,' e suddenly entered it. Besides, the wage that could be earned in:the market is, by observation . insufficient to reward the

' household for- giv'ing up the homemaker's services (Gronau 1973). In spite of the problems with the opportuniti cost approach, we employ 'a modification of it

~

in this study for. practical ease of calculation. 'We compute the mean earnings

-of non-institutional 1 zed, wage-earning individuals in each age and sex cohort and apply that.figureL to all individuals in the cohort.

' 'When refined to include an imputed value.for household labor by those who are not'otherwise emoloyed, the measurement of human capital in terms of cotal

production captures both net production ~(jv4 and also some ' portion of 1, the value an : individual places on a risk to his or her own life. This assumes that the dollar'value of consumption is a rough measure of the . satisfaction a person

- will' receive out- of life.' Thus, an approximation of f is provided by the

, value of the-person's future consumption.

!: The value of ' personal consumption is usually considered an underestimate

.of d. 'In Jan -argument requiring some' theoretical rigor, Schulze et al. (1979),

have shown that the principle of " risk aversion" is one reason why the value'of consumption understatesf. min addition, consumers are often willing to pay more for a' good than they actually.end up paying in the market; therefore, they -

get more satisfaction than is ' represented by the price they pay. Thus, expen-ditures on future

consumption probably understate 1 i

i 4.3 w - * - ' e- w s m-n---- -- -we n ev - N, *- --*m~+~+w vs '~-*n no -*-e-w,-- ->n~^~~~ -***-ww~e"**w'~~"~ ~ " * ~ ~ ~ ~ ~ " " ~ ~ ' ~ " "

The value a person places on his.or her own life (vi) is an elusive mea -

sure. It is not constant over various risk levels; it Ta' ries among individuals; and .for one individual it varies with circumstances and over time. Therefore, it is unclear just how much of vi is measured by total pro-duciton. Nevertheless, given the practical consiTe~ rations of obtaining an

. estimate, in this study we add the value of direct treatment costs to the value of- total production to develop an estimate of the total value at risk that includes v4,, _v_5, 5 and some measure of H.

/

A number of problems with the human capital approach have been observed, both in terms of particular methodological troubles and more generally in termt of theoretical shortcomings. Particular methodological problems include the tendency of the approach to value risk to life based on earnings; those who have low earnings tend to be assigned low values (Mushkin and Dunlop 1979,

.p. 6). Mushkin and Dunlop list other problems involved in human capital valua-tion: changing trends in workforce participation rates at different ages and for males and females, changes in productivity growth rates, and changing earn-ing patterns over a working life (1979, p. 6).

Aside from the methodological problems, significant challenges have been raised against using the human capital approach in risk valuation, on the grounds of incompatibility with economic theory. Neoclassical economists are uniformly in agreement that a measure of human capital simply has no place in 4

cost-benefit analysis. (See for example Mishan [1971].) Instead of using human capital, the benefits of a particular project should be measured in terms of individual preferences, according to economic theory.

In summary, for reasons both of problematic details in the valuation of

, hunan capital, and because of that approach's theoretical shortcomings, many economists have urged that risk to longevity be measured in terms of the value of individual preferences. (For general descriptions of the theoretical sup-port for measuring individual preferences and for comparison of this approach with human capital valuation, see Schelling 1968; Mishan 1971; Acton 1973; Zeckhauser 1975; Jones-Lee 1976; Rhoads 1978; Clarke 1979; Dorfman 1979; and Weinstein, Shepard and Pliskin 1980.)

4.2 THE " INDIVIDUAL PREFERENCE" APPROACH When the total costs are accounted for, the introduction of a particular project (e.g., a project that lowers risk from a reactor accident) will make

.some members of the public better off on balance, some worse off on balance, and others will be indifferent to the project. For example. an investment in safety equipment may decrease public risk but require increased worker expo-sures. If in the aggregate the . total of individual preferences regarding the project is positive, there is a potential for improving overall public welfare by going ahead with the project. In the individual preference approach the value of that potential improvement is interpreted as the excess of benefits over costs arising from the introduction of the project. The value of the 4.4

~

improvement is measured by observing directly the behavior of the public.

Methods to observe individual preferences are discussed in detail below.

Economic theory suggests that the value of a change in an -individual's perceived well-being can be measured 'by the amount of money the individual would' be willing to accept (WTA) or willing to pay (WTP) to remain indifferent

'to the change. The benefit of a risk-reducing project can best' be measured, in theory, by how much the consnunity, in aggregate, would be willing to pay for a

, . decrease -in the level of risk, Er would be willing to accept to face an increase _.in the existing' level of risk.

Selection of the appropriate measure (WTA or'WTP) depends upon the assign-ment of rights within the affected society. If consumers have a right to a lower-risk state, their willingness-to-accept-payment to face a higner risk is the relevant measure. If consumers do not start with the right to a lower

risk, then .we should measure individuals' willingness-to-pay to obtain a lower risk.- In practice, the distinction between WTP and WTA is often blurred, with the availability of information a more important criterion for the choice of either measure than the distribution of rights.

Among the attempts to evaluate individual preferences, three approaches stand out: measurement of WTP by questioning. consumers directly ( Acton 1973; Jones-Lee 1976), measurement of WTA by wage differentials paid to workers in risky occupations (Thaler and Rosen 1976), and measurement of WTP by public budgets for . life-saving programs (Cohen 1980). We ignore the last here because the factors in a program's success in the battle over budgets do not appear to be directly related to society's valuation of the risks ' averted by that pro-gram.

Acton (1973) describes the use of a questionnaire to elicit willingness-to-pay responses directly from the public. While concerned more with general patterns of responses and with the applicability of the technique than with nunerical estimates, Acton concludes that the questionnaire method yields results that are' reasonably consistent internally. He finds that when con-fronted with a hypothetical situation involving risks to themselves~, people are i generally willing to pay-more for larger reductions in risk than for smaller I

ones (p. 105). He notes also, however, that this relationship is non-linear,

[ varying directly with the absolute level of risk faced by a respondent.

i -Because people face and perceive different . levels of risk, the nonlinearity of

_ responses means a single " willingness-to-pay" measure cannot be expected from l such studies (p. 108).' Acton reports that his respondents were willing to pay

[ an average of $43 to reduce annual mortality risk by one death per 1000 people, j and 556 to reduce risk by one death in 500 people (p.109). (Both figures are l- in 1971'$). These values are for risk to a group of which the respondents were i members.

l , It is important to note here that Acton and other investigators of indi-vidual preferences measure individuals' valuations of risk directly. These risk valuations are often discussed in the context of "the value of a life."

In that use, it is necessary to perform a calculation from the risk value to j obtain what Freeman (1979) calls the value of a " statistical life." For l

I- 4.5

~. _ _ - . _ - - --

9 s

1 example, if the average individual willingness-to-pay,for a program that reduced the mortality rate of a given group from seven deaths per 100,000 to six deaths ~per 100,000 'were 55, the "value of statistical life" would be

$500,000. (Freeman 1979, p.- 168) . Thus Acton's -results are commonly. presented in terms o_f a "value of life" ranging from $28,000 to $43,000, depending on the risk change evaluated. For most policy purposes, however, it is the value of-

~

risk that is relevant, not the secondary calculation of value of life.

In an approach generally similar to Acton's, Jones-Lee (1976) also uses a 7

question _naire to estimate willingness-to-pay. Posing a hypothetical situation tin which the respondents themselves are at risk, Jones-Lee finds that, effec-tive.over relatively short periods'of time, the average. reported value of a decrease in risk of one death per 500,000 people is about 6 pounds sterling (19751) (or,about 510.at 1975 exchange rates).

Thaler and Rosen (1976) seek a measure of willingness-to-accept (WTA) in an alternative to the questionnaire. approach. They reason that the wage dif-ferentials paid to individuals-in high-risk industries constitute a measure of those individuals' valuation of risk. Controlling for a variety of nonrisk-

"related characteristics of laborers, Thaler and Rosen present four equations that yield risk valuation estimates in a range from $136 to .$260 (in 1967 $)

for reducing risk from one death per 1000 people to zero.-

Just as. the . human capital approach can be faulted for. ignoring certain

1. components of the cost of risk, so can the empirical studies. undertaken to measure individual preferences. .The risk values reported by Acton correspond

~

only to vi, the value an individual places on risk to his or her own life.

Acton atTemptis measurement of v2 an individual's valuation of risk to loved ones, but does not quantify thTr,esponses in dollar terms.

~

. JJones-Lee (1976) suggests that v4 the risk of losing a share of net pro-duction, and _v5,, the risk =of having To~, share in treatment costs, should be added to vi for a full valuation of the cost of risk. He acknowledges that he has' not aEounted forsv2 the value put on a loved one's life, and he ignores

altogether what we have labeled,v,3,, 3 the preference for lower risk stemming purely from beneficence.

- Kneese and Schulze (1977) employ Thaler and Rosen's high estimate in a rough approximation of the costs of cancer associated with selected environ-mental hazards. However, they reason that even that high estimate is "probably seriously; biased downward." They argue first that workers in risky jobs are less risk ; averse than the general population, and therefore accept risk at a lower wage differential. Second, they suggest- that people may be more willing to take risks voluntarily than to have risks imposed externally. To the extent that risks from environmental carcinogens are-accepted involuntarily, people may demand more compensation for that acceptance. Finally, they argue that

- job-associated death risks may not entail the particularly unpleasant pain and suffering of cancers, for which people would seek higher compensation (Xneese and Schulze 1977, p. 331).

4.6 4

, ~- -+ , s' ,-v- we, wen -

~$v-w ,v,v-+ Awe-mv- w-,

Neither wage _ differentials, as used by Thaler and Rosen for a measure-of WTA~ nor other similar marketplace valuations are capable of including values Lother than-vl, an individual.'s concern for risk to his or her own _ life. Thaler and Rosen c Ecentrate only.on _v_1,.

Table 4.1 provides a summary of the value' components measured by each of the< approaches' assessed in this section. As can be seen in Table 4.1, none of the approaches quantifies adequately all of the ' components of risk value.

, TABLE 4.1. Extent to Which Selected Methods Measure the Various Components of Value Methods vi v2- v3 v4 v5

, _ Human Capital Partial 0 0 Full 0 Human Capital (gus Direct Costs Partial 0 0 Full Full Acton Full (D) 0 0 0 Jones-Lee Full (b) 0 (c) (c)

Thaler-Rosen Partial (d) 0 0 0 0 (a) Approach taken in this study.

(b) This component is considered, but not quantified.

(c) Addition of this component is recommended, but- the study does not attempt. it.

-(d) The critique of Kneese _and Schulze (1977) indicates

' several.. reasons why the wage differential measure may understate 1 Depending on the age, sex, and kinship relationship of the person (s) being considered, Needleman (1976) suggests adding to v1 a value ranging between 25 and 100 percent of vi to account for v2. If any'Talue were added for v3 in -

that scheme it wou1Tbe less than 25 pe~rcent of g.

That still leaves the question of whether the other components could be appropriately added together. Perhaps, as Jones-Lee suggests, one may add WTP or WTA to other component values of risk costs. However, that approach is neither practicable nor desirable in the present PNL effort.

4.3 CONCLUSION

The_ human capital approach is not ideal; it measures only a portion of the probable "true" value of risk ' reduction. And it measures that portion in a way inconsistent with certain principles of economic theory.

4.7

- . . - - . . - . w . - - . . - :_ - - -- - ..=.-. -..-. - - . - . . -

However,' the individual preference approach; while~ firmly rooted in eco-nomic theory, is difficult and costly to implement. Mishan (1971) suggests-

'that a " contingent market" study (i.e., measurement.through surveys) is a pro-per. vehicle for measuring WTP or WTA.i Cronin (1982) shows, however, that such studies must be rigorously designed in order to avoid several kinds of respon-dent-bias. ' While such an approach may be implemented in the future, no broadly

~ based studies-are presently available.

The valuation of individual preferences through WTP or WTA depends to.a significant degree on how the risk _ valuation question is asked, on the per-ceived risk levels, and on the pain and suffering expected. (See Currie and Kidd (1980) for a demonstration of how WTP and WTA values may vary depending on how the question is asked.) It is not appropriate,-therefore, simply to trans-fer a WTP or WTA estimate from one study to another. Instead, it would be

-necessary to perform a special survey to explore individual preferences regard- '

ing the risks of radiation-associated morbidity and mortality. And it would

- still be useful to pursue both the human capital valuation and the direct cost valuation for risk-weighted measures of vj. and v5, respectively, to provide a baseline.

To-gain an. understanding of the magnitude of the value of risk reduction with minimum investment, we have adopted the human capital approach in this study. A contingent market survey would offer greater potential for a full valuation of. health effect risks but it could be implemented only after sub-stantial investment in survey design and testing.

4.8

- , ,e e e < , - - , - - -- ,,--a* .,- ,, --.:-,+. -.n-,---.

4.0 REFERENCES

Acton, J. 1973. Eyaluating Public Programs to Save Lives: The Case of Heart Attack. Rand Corp. Santa Monica, California.

Brody, S. 1975. Economic Value of a Housewife. DHEW, SSA 75-11701. U.S.

Government Printing Office. Washington, D.C.

Clarke, E. H. 1979. " Social Valuation of Life-and Health-Saving Activities by the Demand-Revealing Process." In Health: What Is It Worth? Measures of Health Benefits. S. Mushkin and D. Dunlop (eds.) Pergamon Press, New York, New York.

Cohen, B. 1980. " Society's Valuation of Life Saving in Radiation Protection and Other Contexts." Health Physics Vol . 38 No.1, pp. 38-51.

Cooper, B. Se and D. P. Rice. - 1976. "The Economic Cost of Illness Revisited."

Social Security Bulletin. Social Security Administration, Washington D.C. pp

~

21-36.

Cronin, F. J. 1982. Valuing Nonmarket Goods Through Contingent Markets.

PNL-4255. Pacific Northwest Laboratory, Richlanc, Washington.

Currie, J. W. and J. Kidd. 1980. "A Documentation of Bidding Games Used In Measuring Social Value." FNL-2798. Excerpt from NUREG/CR-0989, PNL-2952,

, Vol. II, Appendix C. Pacific Northwest Laboratory, Richland, Washington.

Dorfman, N. 1979. "The Social Value of Saving a Life" In Health: What Is It Worth? Measures of Health Benefits. S. Mushkin and D. Dunlop (eds.) Perga-mon Press, New York, New York.

Freeman III, A. M. 1979. The Benefits of Environmental Improvement: Theory and Practice. Resources for the Future, Inc. Wasnington, D.C.

i Gronau, R. 1973. "The Measurement of Output of the Nonmarket Sector: The Evaluation of Housewives' Time." In The Measurement of Economic and Social l Performance. M. Moss , ed. Volume 38 in The National Bureau of Economic Research Studies in Income and Wealth. Columbia University Press, New York, New York.

Hartunian, N. S., C. N. Smart, and M. S. Thompson. 1981. The Incidence and

! Economic' Cost of Major Health Impairments. Lexington Books, Lexington, Massachusetts.

Hodgson, T. A. and D. P. Rice. 1982. " Economic Impact of Cancer in the United States." In Cancer Epidemiology and Prevention. D. Schottenfeld and J. F.

Fraument (eds.) W. B. Saunders Co., Philadelphia, Pa. pp. 208-228.

1 Jones-Lee, M. W. 1976. The Value of Life: An Economic Analysis. University of Chicago Press, Chicago.

4.9

' w

.,wa- -- , - , - n -c - - - ~ ~ - . - n.- ,y -.wr

i Kneese, A. V. and W. D. Schulze. 1977. " Environment, Health, and Economics -

The ' Case _ of Cancer." American Economic Review. Vol. 67 No. 1, pp. 326-333.

Mishan, E. J. 1971. Evaluation of Life and Limb: A Theoretical Aporoach; Journal' of Political Economy. - Vol . 79.No. 4, pp. 687-705.

Mushkin, S. and D. Dunlop, eds. 1979. Health: What Is It Worth? Measures of Health Benefits. Pergamon Press, New York, New York.

Needleman, L. 1976. " Valuing Other People's Lives." Manchester School of Economic and Social Studies. Vol. 44 No. 4, pp. 309-342. Cited in Freeman 1979.

Prest, A. R. and R. Turvey, 1965. " Cost Benefit Analysis, A Survey." The Economic Journal (December). pp. 680-735.

Rhoads, S. 1978. "How Much Should We Spend to Save a Life?" The' Publ ic Interest Vol. 51, pp. 74-92..

Rico, D. 1966. Estimating t_he Cost of Illness. U.S. Public Health Service o Publication No. 947-6. U.S. Government Printing Office, Washington, D.C.

Schelling, T. 1968. "The Li fe You Save May Be Your Own" in S. Chase (ed.)

Problems in Public Exoenditure Analysis. Brookings, Washington D.C.

Schulze W. et al. 1979. " Economics and Epidemiology: Application to Cancer." in S. Mushkin and D. Dunlop eds. Heal th: What is It Worth? Mea-suces of Health Benefits. Pergamon Press, New York.

Thaler, R. and - S. Rosen. 1976. "The Value of Saving a Life: Evidence from the Labor Market" in N. Terleckyj (ed.) Household Production and Consumption. Columbia University Press, New York, pp. 265-298.

Weinstein, M., D. Shepard, and J. Pliskin. 1980. "The Economic Value of Changing Mortality Probabilities: A Decision-Theoretic Approach. Quarterl y Journal of Economics. Vol. 94 No. 2, pp. 373-396.

Weisbrod, B. A. 1961. The Economics of Public Health. University of Penn-sylvania Press. Philadelpnia, Pa.

Willig, R. 1976. " Consumer Surplus Without Apology." American Economic Review Vol. 66, pp.587-597.

Zeckhauser, R. 1975. " Procedures for Valuing Lives." Public Policy. Vol. 3 No. 24, pp. 419-464.

4.10

~

m..

' 5.0' ESTIMATION'0F THE DIRECT COSTS OF HEALTH EFFECTS If one' measures ethe values of life and livelihood by the human capital approach, an additional accounting of.the _ direct costs of treating an illness

_is necessary to measure tha total benefit achievable by risk reduction. Con-

. ceptually, in al consumer's re:ponse that- he is willing to pay $X for some risk-reducing program,'there is implied both a.value of life and limb and'an assess- i ment of- the actual monetary outlays he will face if the risk is not reduced.

, Since the consumerfis unlikely' to know the total value.of the monetary outlays,

- the questioner'should be-expected to provide an estimate. - Thus even in a will -

ingness-to-pay approach, an estimate of-?ctual outlays (direct costs) is necessary.

Direct costs 'of radiation-induced health effects include all of the costs of hospitalization, physicians' care, drugs, nursing, special equipment, trans-

- portation required for medical treatment, medical supplies,= etc. Regardless of

whether these costs are paid by individuals, private insurance, or government programs, or represent-bad debts that are paid indirectly by other users of medical . services, they involve costs to society for medical treatment and should be counted.. The rest of this section describes the bases for developing 4

direct cost estimates for /adiation injuries, cancers, and genetic effects.

P 5.1 DIRECT COSTS OF RADIATION INJURIES

~

Depending on dose levels and on-individual sensitivities, exposure to significant amounts of radiation may_ result almost immediately. in acute symp-

, toms 'that could _ range from nausea :to death. Treatment required for recovery Lmay range from a few days of bed rest at home to heroic intervention in a well-equipped regional medical: center. Cases of acute radiation syndrome have occurred too infrequently to result in the development of information regarding treatment practices and costs. However, specialists in radiation medicine have reached relatively close agreement about the clinical manifestations of radia-h tion illness. We-estimate the costs of treatment from information on the cost of treating patients with similar clinical' problems. For this analysis, radia-

~

tion injuries are grouped into three categories: 1) prodromal symptoms, which last only a few days; 2) bone marrow syndrome, gastrointestinal syndrome, and pulmonary impairment, which are all potentially life-threatening; and 3) in-utero effects, which cause severe and permanent impairment to the irradiated fetus.

5.1.1 Prodromal-Symptoms c

' ' Prodromal symptoms, consisting of nausea, vomiting, and diarrhea may occur

'within a few hours of.whole-body exposures over about 50 rads and may continue for-a~ few' days. - Andrews (1980) suggests that individuals displaying prodromal

, symptoms should be kept at home, partly to avoid the infectious environment of

' in hospital and partly to avoid undue apprehension. However, because closely

-monitoring prodromal symptoms is the only way to detect the existence of serious injury, we assume that people would be treated as though seriously i

i 3 5.1 4

.n. ,,, ',,-,e--,- , %,::-6,-.-, ,..n.n.,,w,,, e,,-_,,_,,,nn.,_n.,,n--,,_,.,,_. ,,--,n - -~, ,- _ ,_ n-~n-,.,.-a

-injured 'until evidence develops-to the contrary. -Such treatment could involve two or three days of hospitalization, with the administration of fluids and medications 'and the performance of numerous laboratory tests. In 1981 the

average total hospital charge for . inpatient services was approximately $300 per

-day _(Health Care Financing Administration, June 1982). If physicians' fees average some one-third of_ hospital charges, as they do for cancer patients (Scotto and Chiazze-1976), then' they will total another $100 for each day of care. We assume a 2.5-day stay in the hospital, resulting in an estimate of about $1,000 per case of _ prodromalf symptoms. While provision of such high-quality care may be unlikely in the event of a major accident, lack of-it would probably increase fatality rates and, hence, societal losses. Unless the injured.are quickly identified and isolated to prevent infection, fatalities may occur even 'among those exposed to as 11ttle as 150 to 175 rads (NRC 1975, Appendix VI, F-1).

5.1.2. Bone Marrow Syndrome Bone marrow syndrome is characterized by implirment of the blood forming system;Edepending on the extent of damage, the clinical manifestations include

- . severe susceptibility to infection, hemorrhage and anenia. For purposes of outlining the probable course of treatment and its costs, we suggest there are relevant similarities between the characteristics of bone marrow-syndrome and those of burn trauma. In both cases the most immediate concern is the threat of infection. -In addition, patients suffering from either face a threat of severe-hem' orrhage.

To control infection, burn patients are placed in reverse sterile isola-

~

tion, usually employing special air filtration systems and sterilizing every-thing that comes into the room. Because of- all these special precautions, a regionp{) burn care center charges $1255 per day for " room and board"

~ alone. That is the cost for nonsurgical- burn patients; those requiring surgery receive additional precautionary measures, and pay up to $2,000 per day for a room in sterile isolation. Patients with radiation-induced bone marrow syndrome would require somewhat similar precautions to avoid infection

( Andrews 1980, p. 306; Blakely ~1968, p. 61). Therefore we apply a cost for 4

hospital room of about $1250 per day for about 3 weeks for those patients with bone marrow syndrome.

~

In addition to hospital room charges, a typical nonsurgical burn patient may pay $200 per day for medicapjyns, $180 per day for laboratory tests, and

$50 for each blood transfusion. Saenger (1982) suggests both prophylactic and systemic antibiotic therapy should be used to fight infection in the bone marrow syndrome patient. He advises the use of antibiotic and antifungal agents such as neomycin, oxacillin, and nystatin. That aggressive approach to (a) Communication with staff at Harborview (Seattle) Medical Center's burn care unit March 1983.

5.2

,a, , - -4 .a v

medication is probably not very different from that followed for a burn

~ patient, so we include the full $200 per day for medications in the total cost of treating bone marrow syndrome.

Similarly the' continual _ monitoring of b.lood counts along with laboratory cultures results in high laboratory costs for a bone marrow syndrome patient

-(see Andrews 1980 and Saenger 1982). . The daily costs could easily reach levels

.similar to those of a burn patient. So we add $180 per day for laboratory tests.-

Each -bone' marrow patient can expect a number of transfusions both to replace white blood cells in moderate forms'of bone marrow failure and to

' replace whole blood and platelets in case of hemorrhage in severe cases. We L add another $20 per day to account for cost of a transfusion approximately every second day.

Based on these estimates, total daily cost of hospital services for bone marrow syndrome may run approximately $1650. Because of the relatively high cost of the hospital services component of this care, physicians' charges may not amount to the full 33 percent we have applied to other services based on

-the-experience with cancer care. If physicians' fees amount to about one-fifth of hospital costs in this case, they may total some $350 per day, resulting in

.a total cost close to $2000-per day.

Depending on' the severity of injury, patients may be hospitalized for from two to six weeks. Costs could range, therefore, from $28,000 to $84,000 for bone marrow syndrome. This does not include the cost -of a bone marrow trans-plant, which is often recommended for patients with severe bone marrow 4

syndrome, especially for those who have received a probabl-e fatal dose (Blakely 1968; Dalrymple 1973; NRC 1975; Sa The cost of a bone marrow transplant is approximately $70,000.gerWe 1982). have not included bone marrow transplant as a likely form of treatment because of the difficulties of finding a compatible donor for most patients, a problem that may be more difficult in the aftermath of a reactor accident. In addition, at least one

' researcher (Andrews 1980) advises that marrow transplant may not be helpful.

Although bone marrow syndrome is not the most severe manifestation of acute

' radiation injury, it-is probably the most costly, since other severe forms are almost certain to end in_ ~ death before large amounts of medical resources can be used.

7 5.1.3 Gastrointestinal Syndrome Symptoms of gastrointestinal syndrome include severe diarrhea and vomit-ing. Patients are likely to die within two weeks of the onset of these symp-toms. There is some chance that treatment involving replacement of fluids and r

electrolytes may assist the patient to recover from the associated symptoms.

However,,a radiation dose high enough to cause gastrointestinal injury is also (a) Connunication with staff at the Fred Hutchinson Cancer Research Center, Seattle, March 1983.

5.3 c.

', .2 - _ _ _ , _. ac _ ,_-_u..- ., -- _ _ ___ _ . _ . .~ .- _ _ . . . . _ _ _ _ _ _ .

~

probably_ high enough to ' damage the bone marrow; a patient surviving the~ former-will almost surely suffer the latter. - For that reason, we consider it plaus-ible that patients with gastrointestinal injury will be treated from the: Start

_w ith infection-preventing measures similar to.the treatment given bone marrow

. _ patients. However, since'they are likely to die within two weeks, we apply-to these patients a treatment cost for only two weeks: . $28,000.

' 5.1.4 Pulmonary Impairment Symptoms of pulmonary injury include pneunonitis and pulmonary fibrosis.

. We assume that:(in the absence of bone marrow' syndrome) these symptoms could.be treated .in an-average hospital room at-the average 1981 charge of $300 per day. At 33 percent of: hospital' charges, physicians' fees may add another $100

- per day. - Thus, pulmonary impainnent may cost some $400 per day for all hos-pital and medical services.. In 1977 the average length of stay in acute-care hospitals 'was 8.0 days for pneumonia, and 9.8 days for emphysema (National Center for. Health Statistics 1982). . Lacking simila'r statistics for radiation-induced pulmonary complications, we average the data for those similar diseases and assume a nine-day length of stay. Diat leads to a total cost for pulmonary

. impairment of approximately $3600.

- 5.1.5 In-Utero-Injury-

~'

Cost estimates for direct care of individuals with congenital defects,

similar. in effect to the. retardation and nervous systen. anomalies induced by in-utero radiation injury, are applied to all in-utero injuries. Two studies provide estimates of the-present value of streams of costs that can be incurred in the care of Down's Syndrome (Conley and Milunsky 1975) and spina bifida (Layde, Allmen and Oakley 1979). The_ studies': cost estimates are $116,000 and

$86,500, respectively, in 1981 dollars.1 We are currently using a rough average

'of. those estimates, $100,000, as the_ cost of an in-utero injury.

In summary, the resulting cost estimates are used in the HECOM Model base case-for different-manifestations of radiation injury:

, TABLE 5.1.- Radiation Injury Cost Estimates (1981 $)

Prodromal 1,000 Bone marrow syndrome 56,000 (a mean value)

Gastrointestinal' injury 28,000 Pulmonary injury 3,600 In-utero injury 100,000

^

5.2 DIRECT COSTS OF CANCERS 1 Tim different perspectives have been employed in the past in measuring the direct. costs of treatment for selected diseases including cancers: prevalence

.and incidence. The prevalence approach asks, conceptually, "What is (for

- example) cancer of the cervix costing the nation this year in terms of direct 5.4 o e- < --me e- w www es,- e,m,~,,, ww- w v-n e a:.-- -w.=r--~=~-------ew.ma -+-~~e--=~- - - - - = ~ - ~ ~ - " - *

$ }

a outlays for treatment? - It is this approach that has been followed by Rice and her-associates (Rice 1966; Cooper and Rice 1976; Hodgson and Rice 1982). The-

- prevalence approach-is well-suited to an aggregate or ' op down" accounting of.

illness costs, in which-total national expenditures for selected health ser-Vices are allocated to the various illness categories. Direct costs thus com-

puted are of little'use, however, in evaluations of actions that affect the '

risk -of illness.

The incidence approach asks, conceptually, "Given a certain event--a reac-tor accident, for example--what will be the total cost of treating the assoc 1-

- ated health effects?" The incidence approach requires a " bottom-up" measure-ment'of treatment costs based on scenarios of expected treatment. -

In' practice, the treatment regimens .used in the two principal studies of-costs-of cancer incidence (Cromwell et al.1976 and Hartunian, Smart.and Thomp-e son 1981) are based on infonnation regarding treatment as reported in the Third National Cancer Survey. The PNL H6COM direct cost estimates for various types of cancer mirror the basic approach taken by both Cromwell and Hartunian:

given the treatment regimens reported in the Third National Cancer Survey 2

(TNCS), compute current costs by inflating TNCS costs to current dollars (with

. a few adjustments).

Cromwell' and Hartunian provide the only incidence-based measures of direct cancer costs across a range of cancer types presently available. A nunber of other studies have undertaken " bottom-up" measurements of costs for particular

< types of cancer, for. example, Scitovsky and McCall (1976), Kodlin (1972), and Schneider and Twiggs 1972).- Unfortunately, those studies concentrate typically on patients with specific cancers, and are unrepresentative of treatment regimens and costs for a broad range of cancer types.

5.2.1 Cancer Cost Data Because the 'TNCS is the primary source of information, both on services

- rendered.and on costs, it is useful to review the strengths and limitations of r the TNCS data. As part of the. TNCS, a. sample of approximately 8500 cancer L patients, newly diagnosed in the years 1969-71, were interviewed in depth with l a Patient Interview Booklet (PIB). (That study represented slightly less than 10 ' percent of the full TNCS sample). The PIB elicited details both on the services received by each patient and on the payments for those services. In addition to the PIB, information on hospital charges was extracted from patient

. records for 6332 of the TNCS patients. Scotto and Chiazze (1976) report hos-pital charges-as contained in the hospital records sample of the TNCS. Crom-g - well et al. (1976) 'uses payments from patients to hospitals and to other health

! providers, as reported on the PIB. As Cromwell shows (pp. 66-68), the differ-l ence between the two data sources is small in terms of average hospital cost l

per cancer case. Among the various types of cancer, however, Cromwell shows

). that there are significant differences between the two data sources (differing l

l' by.as much as 50 percent). ~ Cromwell concludes that the self-reported data from

the PIB may be an unreliable source of hospital costs by cancer type.

l 5.5

. .-_ - -_ m m._. :_________________ - . _ . _ _ _ _ _ _ _ _ _

Nevertheless because the PIB is also the source for other treatment costs, Abt (1975) records hospital costs as on the PIB. In comparison, Hartunian, Smart and Thompson (1981) use data from Scotto and Chiazze (1976) to measure

. hospital costs. Then they use the ratio of hospital costs to other service costs, as reported on the PIB, in order to estimate the cost of all nonhospital services.

In addition to the details of particular data collection instruments, there are other limitations to the cost data from the TNCS. Cromwell (1976, pp. 56-73) identifies biases in that the high cost Northeastern states are not.

represented, nonresponse occurred more heavily among those with the more aggressive cancers, and interviewees exhibited selective memory.

A final structural limitation of the TNCS data particularly worth mention-ing is that the PIB data cover a time interval between the onset of symptoms and the date of the interview. This time interval varied widely (Cromwell .

1976, p. 72) and the wide range of time spans makes it difficul.t to interpret the cost data. Ideally, direct costs would include monetary outlays for the entire course of the illness, discounted to present value 'in the year of decision making. Lacking such data, direct costs should be measured over a standard time frame, such as considered (for hospital costs, but not for the costs of other services) by Scotto and Chiazze. Their data include hospital costs over the first two years after diagnosis.

5.2.2 Cost Estimation Methodology The direct costs of cancer include all' of the costs of hospitalization, physicians' care, drugs, nursing, special equipment, transportation, radiation treatments, chemotherapy, etc. Disaggregate data from the TNCS are used to create the following cost categories:

Hospital / inpatient - includes physicians' and nurses' services, laboratory, diagnostic, radiotherapy and surgical charges as well as hospital bed charges, supplies, and special services.

Outpatient / doctor - office, home and clinic, outpatient visits and surgical and other physician inpatient costs.

Nursing home - includes daily room charges, nursing costs, and supplies.

Private nurse - costs of in-hospital private nursing, billed separately.

In-home nursing - includes nursing and supply costs.

Drugs - includes everything from prescription drugs used in chemotherapy to over-the-counter medications.

4 Rehabilitation - includes physical therapy, special equipment, and prosthetics.

5.6

_ .c

i

~

These direct cost components are then used to construct a direct cost

^

estimate for each CRAC2 cancer category. Since the TNCS data are categorized partially by cancer type and partially by cancer site, these. categories are combined to correspond with the CRAC2 output as shown in Table 5.2. For the CRAC2 categories of " gastrointestinal tract" and "other," the TNCS data are aggregated using the proportional incidence of the major types or sites-of cancers as weights for the costs. Those cancers constituting less than five percent of the total incidence are not included. Thus, based on the distribu-tioc h Table 5.2, the cost estimate for gastrointestinal cancers is a weighted average of the costs for the sevea major. types of cancers falling within that category.

TABLE 5.2. Corresponding Cancer Categories in CRAC2 and the Third National Cancer Survey CRAC2 Category Third SurveyNational CategoryCpnfer a

Percent (D)

Leukemia Leukemias --

Lung Lung --

Breast Breast --

Bone Bone --

" Colon 34%

Bladder 14%

Rectum 11%

-Gastrointestinal < Pancreas 11%

Stomach 11%

Oralcaviti 12%

, Kidney 7%

' Larynx 44%

Cervix 10%

Other <

Uterine corpus 11%

Prostate 26%

, Lymphomas 10%

(a) The table excludes TNCS categories that constitute fewer than i

5 percent of the corresponding CRAC2 cases.

(b) TNCS category (by site) shown as a percent of the corresponding, broader CRAC2 category.

An estimate of each cost component, such as " hospital / inpatient," is then calculated for each of the CRAC2 cancer categories, using the proportional weighting for TNCS categories described in Table 5.2 for gastrointestinal and "other" cancers. These cost estimates, representing first and second year 5.7

treatment costs (Cromwell 1976, p. 70) for the eight categories of direct

~

costs, are shown in Table 5.3. While treatment of some patients may extend over several years, the brevity of median survival periods makes application _of two years' costs- to all cases a reasonable approximation of total costs. Shown along with each: cost estimate is the percentage of patients surveyed who incur-red this type of cost. These percentages are applied to each cost category to calculate the weighted total cost shown in the last column for each type of cancer. In calculating benign thyroid nodule costs the base case assumes that 75 percent of the benign nodules are diagnose ithout surgery and that only outpatient costs are incurred _in these cases, a The weighted total of cancer care costs is converted to 1981 dollars using the hospital room and medical care cost ' components of the Consumer Price Index (CPI). Once the' direct cost estimates are calculated in this form, they are

'used with the CRAC2 health effects projections to calculate the total direct cost of cancer care over time. CRAC2 health effect estimates, which except for thyroid are for fatalities only, are converted to incidence estimates by appli-cation of the ratios shown in Column 2 of Table 5.4 Since the thyroid health effect estimate produced by CRAC2 reflects incidence (NRC 1975, Appendix VI,

p. 9-27), it needs only to be partitioned between benign and malignant cases.

The resulting estimate of thyroid cancer (and benign thyroid nodule) incidence can then be allocated across age groups and time to calculate total direct cost due to exposure.

-Since the cancers would not occur immediately after radiation exposure but would generally have n.inimum latency periods of from two to 15 years, the dir-ect costs must be discounted to a present value estimate. This is accomplished by discounting the costs that' are projected to occur over the remaining lifetime of. the exposed population. First, cancer incidence is allocated to age groups in proportion to.the size of each age group in the exposed population and the relative risks for people in each age and sex category.

Members of each age group are then assigned a probability of developing cancer in each year after the minimum latency period until they reach the maximum age considered.

The preliminary cost estimates shown in 1970 dollars in Table 5.3 have been inflated to 1981 dollars using the appropriate components of the CPI.

Table 5.5 presents the resulting PNL estimates for each CRAC2 cancer cate-gory.- These are the costs presently being used in the base case of PNL's Health Effect Cost Model (HECOM). They can be converted from 1981 dollars to

[ any other year's dollars using the medical care cost component of the CPI (see Table A.13).

l (a) Comunication with Oncology Department staff, University of Washington.

5.8 l

t i

TMIE 5.3. Direct Costs of Cancer Care for First Tw) Years of Treatment by Cancer Type (1970 $)

I Ibhabili-Ibspital/

CRN2 Categories Inpatient %

h .&tpatient/ Mrsin9 Private In-lbne .

Wi@ted Doctor % lhne % m rse % m rsing % Drugs % tation  % Other % Total leukania 3,914 100 881 100 1,421 13 ,317 7 677 10 198 68 186 18 106 70 5,312 1.ung 3,905 100~ 1,376 100 1,369 12 608 6 572 17 134 75 122 36 67 81 5,814 Ikeast 1 ,745 100 996 100 3 /72 8 458 10 - 464 5 104 71 50 100 47 69 3 228 Done 7,908 100 2,041 100 - --

200 19 181 90 2,395 100 132 100 12,677 Castro-irtestinal 3,140 100 1,138 100 1,p67 5 694 6 990 14 120 73 115 15 81 30 4,822 1 Other 2,366 100 1,132 100 1,144 9 266 6 941 4 138 100 81 14 72 53 3,842 1hymid-ben 19n 1,516 25 1,108 100 -- -- -- -- -- --

1,487 i

Thyroid-malimant 1,516 100 1,108 100 1,250 5 234 21 187_ 14 81 100 153 24 34 100 2,914

• Source: O

, 11, J., et al.1976. The Measurenent of the Cost of Cancer Care. Mt Ibport No.76-152; et Associates, Inc, and Ibston thiversity Gincer Ibsearch Gnter, Cait> ridge, Miss. Ibspital and &tpatient/ Doctor costs are fran Table 3.4, p. 60; all other costs are frun Table 3.3, pp. 58-59. .

(a) Ibspital costs are increased by 20 pertent to reflect mcollected charges. Acon 11n9 to 20tto and (hiazze (1976) an investi9ation of selected survey cases showed that 20 percent of actual hospital charges were not reflected 6 the Thini ibtional Oncer Survey data since they are not paid by the patient, private instrance, itdicare, or Hslicaid.

(b) Dercentages represent the proportion of patients with a 91ven type cancer, dio receive each type of service. For each type of service, patient totals are adjusted for missin9 data, as stryysted by Quaell et al.

i i

\

TABLE 5.4. Calculation of Caacer Incidence -Based on CRAC2 Fatality Estimates Incidence Health CRAC 2 to Fatality Effect Health Effects Estimates X Ratio = Incidence Fatalities Leuk'emia X 1.00 Lung i Yi X

2 1.00 Y 2

Breast X 3 2.00 Y3 Bone- X4 1.25 Y4 Gastrointestinal X 5 1.20 Y5 Other X 6 2.00 Y 6

Incidence .

Thyroid-benign (a) y Thyroid-mal ignant(a) X X7 7

0.6((

0.4 D) b) Yg Source: U.S. - Nuclear Regulatory Connission. 1975, pp. G18-G23.

(a) CRAC2 provides a single incidence estimate for all thyroid effects.

(b) Proportion. of nodules that are benign or are malignant.

TABLE 5.5. . Direct Costs of Cancer Care by Cancer Type (1981$)(a)

CRAC2 Catego*1es - Weighted Total Cost Leukemia 16,300 Lung 17,400 Breast 9,400 Bone 37,600 Gastrointestinal 14,000 Other 11,400 Thyroid-benign 7,700 l Thyroid-malignant 8,400 Source: Table 5.3 and Consumer Price Index inflators from the US l Bureau of Labor Statistics. (Monthly)

(a) To convert from 1970 to 1981 dollars, hospital costs are inflated using the hospital room cost component of the f

CPI; all others ar1 inflated by the l all medical care cost component.

1 l

l 5.10

5.3 DIRECT COSTS OF GENETIC EFFECTS .

Estimating the costs of radiation-induced genetic disesse is a task made difficult by both conceptual ~ issues and limited information. In this section we examine some relevant conceptual issues, describe an approach to estimating costs, and apply limited data within that approach to construct a preliminary cost estimate.

In Section 4.0 'we suggested several reasons why individuals wGuld prefer lower health risks: because they value life itself, for themselves (v_l), for _

their loved ones (g), and for anonymous others (g); because they prefer not

and because they prefer not to losethe to bear theresource

' net production of others' costs of treating otherslabor s (g) illnessIn(1).

estimating the costs of radiation injuries and radiation-related cancers, we have proposed that since the sum of direct and indirect costs accounts for most of vl, v4, and i that sum is a reasonable approximation of total costs. -

With respect to genetic disease, the rationale for use of direct and indirect measures is similar, albeit more difficult to see. If genetic disease affects only future generations and not this one, does an estimate of future direct costs measure _v,5,5 for this present generation? And does an estimate of the loss of future earnings measure either v1 or v4 to this generation? That is, we (this generation) are not the ones aTrisk7 rom genetic disease and we need not bear the cost of those health effects at all; why then, should we value either resources consumed by future generations (v5), or net production (earnings) forgone (v4)? And if we are not at risk, wny include a measure of

, forgone consumption (vi) as a measure of loss from genetic mortality?

The answer lies to some extant'in the fact that generations overlap; this generation will actively share in v4 and v5 for the next generation and to a lesser, but still positive extent 7n thaTof the second generation hence. In addition, the satisfaction (" utility") of this generation is usually considered to. depend not only on one's own opportunities but on the income and consumption opportunities of future generations. Thus, the welfare of future generations affects this generation directly, to the extent they will soon co-exist with us, and. indirectly to the extent that our levels of satisfaction depend partly on theirs.

Employing direct and indirect costs as a. measure of this generation's valuation of future health effects goes even further than mere concern for the future. It treats future generations in an egalitarian way, valuing their L . health effects as though they were our own. That is, if vi is a measure of how

~

much an individual values his own life (because it is a treasure of his future consumption), then it is an appropriate component of the valuation of health risk only if the individual is among those at risk. Therefom, for this gener-ation to consider direct and indirect costs, vi + v4 + v5 as the valua-tion of genetic effects means that this generation i.e., Waluates tnose health effects on the same basis as if we were the ones at risk.

In practice, these future costs are discounted to present value, just as costs incurreo later in this generation would be. Discounting results in a 5.11

u

~

measure.of the funds thatLwould need to .be placed in an annuity -at the time of a reactor. accident in order to pay the costs occurring at some future date.

(Choice of the ' discount . rate- for intergenerational valuations is a methodologi-

-c  ; cal issue in.'itself. which we ~ address _briefly in this section.) .

- .Given that' future . direct costs;are an appropriate measure .of this genera-

- tion's valuation of genetic . effects, there remain a number of problems in esti- -

mating those. direct costs. The remainder of this section presents an ' approach -

to estimating the direct costs of genetic disease. The. associated indirect

costs are examined L in Section 6.0.-

'5.3.1 ' Genetic Effects Cost Data

. Information on the' costs of treating' disabilities'.and diseases .that are l genetic in origin is very limited. In this section- we rely on two studies for >

specific diseases to estimate the magnitude of the direct costs of genetic

effects.-

l' '

Hall et al. (1978) present data on the hospital treatment at one urban medical center of children with genetic disease. That study reports an average cost' per hospital -admission of ;approximately $1100 -(1981 dollars) for children with diseases unambiguously attributable to genetic causes. Those children had been admitted to the-hospital an average of 5.3 times each at the time records

..w ere reviewed for Hall's study. If that were the total number of admissions per child the total hospital. cost per child would average approximately $5830 (1981 dollars). Of course, there is no reason to surmise that the end of the

j. study coincided with the end of hospitalizations for the chfidren sampled, so

$5830 is doubtless an underestimate of the average total hospital ' costs.

1 Assuming that physicians' fees average approximately one-third of hospital I- costs, as is the experience with cancer patients (Scotto and Chiazze 1976), the average total costs for acute care of. childhood-onset genetic diseases may be as low as $7775 (1981-dollars), but are most likely higher because of multiple hospitalizations. We treat the total acute _ care costs as if they were incurred

-in-the first year of life.

, In addition to the costs of acute care, a portion of the genetically diseased population also will incur costs for long-term institutional care.

Conley and M11unsk viduals with Down'y (1975) s and examine Hunter's the costThose Syndromes. of. institutional' syndromes care for indi-are related to chromosome aberrations and would account for a very small percentage of the genetic diseases associated with radiation exposure (UNSCEAR 1977 and BEIR III 1980). .However, costs- for those two syndromes may be somewhat representative of the' costs for long-term institutionalization of other genetically impaired individuals.. Assuming the costs are representative, for an individual born in 1981 and; institutionalized for the'next 70 years the cost would be approxi-mately $14,000 annually in 1981 dollars (inflating Conley and Milunsky's 1972

, estimates by. the medical care component of the CPI). Conley and Milunsky

, .' report that.approximately 20 percent of the cost of institutionalization is comprised of normal personal consumption and should not be considered to be a l

5.12 "m---..-w%,, --.aw.~ .,,,-.+,,[. .-w.w-, ...+m-,,d,,m-. - ,, .. g.--.w h [ -.,,--,-.n.m--.7,w--,-mm,,--o..-mm--m--.v- --r,. ,._,n,y-,- ,

- . . - . ._- . . -.- . - . ~ . ~ - . _ . - - - - . . - - - -

O' ,

i

' le- '

result of disease. We. subtract that amount so that only tne incremental costs 4 -

of illness are consideredr$11,600 per year. These costs are distributed over 1the person's. lifetime. :

5.3.2 Cost Estimation Methodology

The genetic effects associated w'ith increases in . radiation exposure may

- range in severity.fran color blindness to mortal or debilitating diseases.

Obviously the costs vary _as well.

The two studies cited in the previous section result in rough estimates of  !

$8,000 for. acute care and $11,600 per year.for long-term care for individuals *

. affected by severe genetic diseases. The problem is to determine what propor-tion of genetic effect result'in costs of this magnitude.

Mckusick-(1978) lists 736 " diseases" that can be traced to autosomal ,

. dominant genetic defects. Along with X-linked defects, the autosomal dominant

-would be a major. category of genetic effects likely to result from radiation -

" exposure. Some of those " diseases" cause little or no symptomatic problems; others.are life-threatening or totally' debilitating. The genetic effect may oe s' obvious at birth in some: cases and disability onset may occur in adulthood in

others. . Unfortunately; we are unaware of any studies that provide the fre-l' quency of genetic-effects classified by severity. . Lacking any information as

.to the frequency of genetic effects.with no economic costs relative to those i' resulting in maximum cost, we assume. .for current working purposes,- that .the median point .in the range is representative. To implenent this assumption me treat. half of the genetic effects as resulting in maximum costs and half as resulting in no treatment or institutionalization cost.

The resulting estimated lifetime costs for treatment are discounted to p' resent value at the time of each affected individual's birth. We allocate 7~

those births over 10 generations after the hypothetical . reactor accident. The defects projected for the first generation are allotted to the first 30 years,

~

the second 30 years for the second generation, ar,d so forth. Those genetically

. impaired births projected for the first generation' are distr. 'uted evenly over l

~ the first 30 years post-accident. The number of' affected births projected for the second generation are distributed evenly over the years 30 through 59 and so forth.

L After applying these stimates of average lifetime costs to each birth dis-

, -tributed over the appropriate' generations, it is necessary to discount those j' costs to their present value in the year of the hypothetical accident. That '

L process yields an estimate of the funds that could be placed in an annuity at  ;

the time of a reactor accident to pay for future direct costs of genetic

-effects. There is an enduring question -in economic theory concerning the appropriate discount rate for analysis of intergenerational cost streams.

Because the diacount-rate must be treated as an important factor in any evalua-E  : tion of future costs, a sensitivity analysis including the application of dif-forent discount rates is presented in Section 7.4.

l 5.13 1

-,,r., , , ,-.

T. ' ~% 2,,-..m,

. . . J : ' ?" ;",. k w . . -,~,,--A, R ,-s i ,--, , nd, ,, h w . -n n m m e , , ,m _- _-, ww,--,w,-

I i

5.0 REFERENCES

Abt, C. 1975, "The Social Cost of Cancer." Social Indicators Research.

vol ?. pp.175-190.

Andrews, G. A.1980. "The Medical Management of Accidental Total-Body Irradiation." In the Medical Basis for Radiation Accident Preparedness. K.

F. Hubner and S. A. Fry (eds.) . Elsevier/ North Holland, New York, New York, pp. 297-311.

Blakely, J. 1968. The Care of Radiation Casualties. Charles C. Thomas, Springfield, Illinois.

Committee on the Biological Effects of Ionizing Radiation. 1972 and 1980. The Effects on Populations of Exposure to low levels of Ionizing Radiation.

National Academy of Sciences, Wasnington 0.C.

. Conley, R. and A. Milunsky. 1975. "The Economics of Prenatal Genetic Diagnosis," in The Prevention of Genetic Disease and Mental Retardation. A Milunsky, ed. W. B. Saunders Co., Pn11adelphia, hnnsylvania.

Cooper, B. and D. Rice. 1976. "The Economic Cost of Illness Revisited."

Social Security Bulletin. - Vol. 39, No. 2. Social Security Ad.ninistration, Baltimore, Maryland, pp. 21-36.

. Cromwell, J. et al. 1976. The Measurement of the Cost of Cancer Care, Task Two Report. Abt Associates, Inc. and Boston University Cancer Research Center, Boston,Itassachusetts.

Dal rymple, G. V. et al . , eds. 1973. Medical Radiation Biology. W. B.

Saunders Co., Philadelphia, Pennsylvania.

Hall, J. G. et al. 1978. "The Frequency and Financial Burden of Genetic Disease in a Pediatric Hospital." American Journal of Medical Genetics.

Vol. 1. pp. 417-436.

Hartunf an, N. S., C. N. Smart, and M. S. Thompson. 1981. The Incidence and Economic Costs of Major Health Impairments. Lexington Books, Lexington, Massachusetts.

Health Care Financing Administration, U.S. DHHS. 1982. Health Care Financing Trends. Published quarterly by the Department of He.alth and Human Services, Washington, D.C.

Hodgson, .T. A. and D. P. Rice. 1982. Economic Impact of Cancar in the United States." In Cancer Epidemiology and Prevention. D. Schottenfeld and J. F.

Fraumeni eds. W. B. Saunders Co., Philadelphia, Pennsylvania.

Kodlin, D. 1972. "A Note on the Cost-Benefit Problem in Screening for Breast Cancer." Methods of Information in Medicine, Vol. II, pp. 242-247.

5.14

Layde, P. M., S. D. Allmen, and G. P. Oakley, Jr. 1979. " Maternal Serum Alpha-Fetcprotein Screening: A Cost-Benefit Analysis." American Journal of Public Health,. Vol. 69, No. 6, pp. 566-573.

Mckusick, V. A. 1978. Mendelian Inheritance in Man. (Fifth edition). Johns Hopkins University PreTs, Baltimore, Maryland.

National Center for Health Statistics, U.S. DHHS. 1982. " Inpatient Utilization of Short-Stay Hospitals, by Diagnosis,1979. Vital and Health Statistice. Series 13, No. 69, OHHS Pub. No. (PHS) 83-1730. U.S. Government Printing Of fice, Washington, D.C.

, Rice, D. P. 1966. Estimating the Cost of Illness. Health Economics Series, No. 6. Public Health Service Pub. No. 947-6. U.S. Government Printing Office, Washington, D.C.

Saenger, E. L. 1982. " Radiation Accident Preparedness." A Course Manual from the Universit of Cincinnati.

Schneider, J. and L. B. Twiggs. 1972. "The Costs of Carcinoma of the Ce rvi x." Obstretics and Gynecology. Vol. 40, No. 6. pp. 851-859.

Scitovsky, A. and N. McCall. 1976. Changes in the Costs of Treatment of Selected Illnesses, 1951-1964-1971. DHEW Publ. No. (HRA) 77-3161.

Rockville, Maryland.

Scotto, J. and L. Chiazze. 1976. Third National Cancer Survey:

Hospitalization and Payments to Hospitals. National Institutes of Health, liashington, D.C.

United Nations Scientific Committee on the Effects of Atomic Radiation.

1977. Sources and Effects of Ionizing Radiation. Report to the General Assembly, United Nations, New York, New York.

U.S. Bureau of Labor Statistics. Monthly. Consumer Price Index Detailed Report. U.S. Government Printing Of fice, Washington, D.C.

U.S. Nuclear Regulatory Commission. 1975. Reactor Safety Study. Appendix VI.

WASH-1400, National Technical Information Service, Springfield, Virginia.

5.15

6.0 ESTIMATION OF INDIRECT COSTS OF HEALTH EFFECTS In addition to the direct costs of treating radiation-induced illnesses, there are potentially much larger indirect costs associated with those health effects. Indirect costs do not involve monetary outlays, but rather represent other losses incurred by society as a result of the health effects. In Section 4.0 we presented a conceptual discussion of how those other societal losses might be valued. Using a " human capital" approach, societal losses due to increases in illness and premature mortality are measured in terms of the value of lost production. That is, when an individual is too sick to work or when he or she dies earlier than might be expected, that person produces less. Because wages are a measure of the value of a person's marginal product, the value of the lost production is measured in terms of the value of lost earnings.

The value of earnings lost due to increas.ed morbidity or premature mortality provides an approximate measure of two components of the societal losses due to illness. Lost earnings mean lost consumption to the indivi-dual. (That corresponds to 1 in the taxonomy employed in Section 4.0). The rest of society incurs a loss as well, consisting of the value of what the individual would have produced over and above what he or she would consume.

(That is a measure of net production and it corresponds to f in Section 4.0).

The value of lost earnings should be considered an underestimate of the v2) and the loss society in general feels purely out of beneficenceFu(v,3). full indirect costs, -

using an individual's lost earnings as the measure of lost production ignores the lost production experienced in addition by family and friends who take time out to care for the stricken individual.

There is another way in which the use of earnings often underrepresents the full indirect costs: earnings data do not reflect the value of services performed in the home. In this study we avoid that shortcoming by two steps.

First, we consider the population incurring indirect costs to be all non-institutionalized individuals, not just persons in the labor force. Second, within each age and sex cohort, we apply the average earnings of employed individuals to all non-institutionalized persons in the cohort. That is, the production of a female homemaker, aged 35, is considered to be equal in value r to that of an employed woman of the same age. (The method treats all males equally as well, although it does not treat men and women equally.)

The following sections relate how lost earnings measures are applied to evaluate both morbidity and mortality related to radiation-induced health i

effects. In general, several causes of lost production are associated with health effects: inability to work during acute phases of radiation injury or cancer, reduction in capabilities as a result of the illness, inability to work due to mental or physical impairment as a result of prenatal injury or genetic defect, and permanent cessation of work due to early mortality. In this study, we explicitly calculate costs related to all those causes except those due to 1

6.1

illness-related reduction in capabilities; the average earnings data used implicitly reflect a low rate of handicaps among workers.

. 6.1 INDIRECT COSTS OF H0RRIDITY lost production during illness is estimated based on weeks of missed work for each type of illness. The value of that loss is measured by average earn-

.ings, for individuals of a particular age and sex, in each ~ post-exposure time period. The incidence of illness is assumed to fall across age and sex cohorts in proportion to age- and sex-related risks of radiation-induced illness and to each cohort's relative numbers in the exposed population. The estimate takes l into account the individual's age at the time of illness and also accounts for the fact that normal probabilities.of, death lead to an expectation that some exposed individuals would die of other causes before latent cancer can result in any lost production.

'For cancers, we apply an estimate of lost work ranging from about 6 weeks to more than 23 weeks depending on the cancer type. (See Section A.2.4.) .

Among the cases of radiation injury, prodromal symptoms are assumed to cause

.one lost week of work; all other types of radiation sickness are assumed to result in a loss of six months of work. Individuals disabled by growth impair-ments and mental retardation resulting from prenatal exposure to more than 200 rem are assumed to suffer a 100 prcent income loss, beginning at the age of 15 and continuing over the person s expected lifetime. hnong individuals afflicted with genetic defects, we currently assume 50 percent to suffer a 100 percent income loss similar to those injured in utero. The remaining 50 percent are currently assumed either to have no handicap as a result of genetic disease or are considerd to have been successfully treated before age 15.

The model considers the incidence of genetic effects through ten generations (300 years). The indirect costs of genetic effects are calculated l in a manner similar to those for premature mortality due to illness. A review E of the literature, unfortunately, does not disclose any estimate of the rate at-which productivity impairment results from genetic effects. We currently assume that one-half of the individuals experiencing genetic effects ~ will never be productively employed and that the remainder have no impairments. Applying this assumption, the expected earnings of each age cohort (given normal mortality probabilities) provide the basis for estimatt.ig the stream of

- potential indirect costs for a genetically damaged individual born in each year after population exposure. The rate at which such individuals are born is calculated as it'is for the direct costs, allocating first generation effects equally across the first 30 years post exposure, the second generation effects across the next 30 years, and so forth. The resulting indirect cost streams are then discounted and summed to-the present value at the time of population exposure.

For all types of health effects, the indirect cost of morbidity is esti-

-mated from the amount of work lost, valued by expected earnings. These costs are computed for the specific age and sex cohorts in the population and the 6.2

1

.1 time period in;which they would face health effect risks. .To apply those costs to the year of exposure,.the projected stream of future costs is discounted to present value as of.the year of a hypothetical reactor accident.

6.2 ' INDIRECT ~ COST OF MORTALITY

' The indirect. cost of mortality is valued by the earnings lost as a result

of' exposed individuals dying earlier than would be-expected in an unexposed ,

population. The basic: computation is most easily seen in-an example: For an ]

individual who dies at-the age of. 30, the indirect cost would be the discounted l sum of his or. her expected future earnings. It is assumed that in each poten- '

tial year of life after age 30 the individual would have produced (and there-fore earned) a value equal to the average for his or her age. and sex. The average earnings in each future year.are weighted by the probability that the individual would have survived to that age, had he or she not died at age 30 due to radiaton exposure.

Fatalities from acute radiation injuries are assumed to affect individuals

, of each age ~and. sex cohort in proportion to their relative numbers in-the total population. .For those who suffer fatalities from acute injury, the fatality is

~

assumed .to come.in the first year after exposure. Thus, expected losses begin in the ~ year.of the accident and extend out for many years, until all those exposed would have been dead of other causes. The total indirect cost is the sum of the discounted stream of-future-losses for each fatally exposed individual.-

Cancer-related mortality costs are calculated in a similar manner, except that cancer fatalities, and therefore the onset of losses, occur over a period of years. CRAC2 estimates of cancer fatalities are assigned to age and sex cohorts in proportion to their risks of radiation-induced cancers and relative numbers in the population. Each type of cancer has a specific minimum latency period (see Section A.2.2) between exposure and the onset of cancer symptoms.

After the latency period has passed, individuals are expected to show signs of cancer and to die from those cancers over a time period distributed over what would h ave been their normal lifetime. That is, not -all individuals will .show -

i. cancer symptoms in'the years immediately following the end of the 1atency period;.and even.after the onset of symptoms some people will not die for many years. .Thus, cancer fatalities are treated as having an equal probability of occurring in each year after the latency period and continuing for a normal

, life span.

. For example, CRAC2 may project that two persons in the 30 year-old age

, group will contract a fatal bone cancer. After a 10-year minimum latency period between exposure'and bone cancer symptoms, the probability of fatality

-is. treated as being proportional to the probability of survival in each remain-ing year of normal life expectancy. The resulting fatality rate due to bone i - cancer is constant over the remaining-lifetime of the 30 year-old cohort. Thi s

!' probabiity of death in each succeeding year is applied to the value of the earnings. loss that would occur if an individual from that age cohort died in that' year. The total indirect cost is.the sum of the discounted stream of probabilistically weighted future losses for each individual.

6.3

. - - ['. ,,_v ,, .--w.,...m , , . . ,,.m ,.v m.__%.,E.,,..,z., , . , , ,.,,.,.,w. ,,, reg , , . - ,_n .v., ., y. - %+.,-

1 7.0 HECOM STRUCTURE AND DEVELOPMENT l

This section provides a conceptual overview of the Health Effects Cost Model (HECOM) structure and processes. A more detailed, user-oriented discus-slon of the data base, subroutines-and processes is contained in Appendix A and the computer code is listed in Appendix B.

In Section 7.1, the general approach used in developing HECOM is dis-cussed. Aspects of the model's flexibility and the treatment of future cost streams in present-value, real terms are emphasized. An overview of the HECOM structure is then provided in Section 7.2. This is followed in Section 7.3 by a discussion of the steps required to modify CRAC2 output for use as input to HECOM. The sensitivity of HECOM to both input data and parameters has been )

examined and this analysis is presented in Section 7.4.

7.1 MODELING APPROACH The general approach employed in developing HECOM was dictated by the need ,

to develop a flexible model that could be easily updated or modified. The ways in which this flexibility have been achieved are discussed in Section 7.1.1 and the method used to discount future cost streams is explained in Section 7.1.2.

HECOM is a probabilistic model designed for analysis of changes in popula-tior, health risks. The cost estimates calculated by the model are dependent on population distribution by age and sex,- cohort arvival probabilities, excess health effect risk estimates by cohort, and probabilistic distributions of incidence over time. As a result of this approach, HECOM can project the societal impacts of health effects for which timing and population incidence are indeterminate. .

The cost estimates calculated by HECOM are expressed in real, or constant dollars, excluding strictly inflationary changes in costs. As a result of this approach, the cost estimates reflect comparable real resource costs regardless of the future year in which the costs may be incurred. All future costs are discounted to the base year of analysis so that the resulting HECOM estimates refle ture.gthe Detail present value of-the of costs that discounting mayemployed method actually is beprovided incurredinin"aection the fu-7.1.2.

7.1.1 Flexibility of HECOM HECOM has been designed to be as flexible as possible, subject to the limitations imposed by the computer. code used to develop the model. This (a) To analyze the consequences of axposures in years after the base year of analysis, costs and wages can be escalated to the level of the year of exposure before being input to HECOM. HECOM cost-estimate output can then be discounted back to the base year.

7.1

flexibility enables the model- to use input data ' , several different forms and to easily calculate cost. estimates for a variet) of population exposure scenar-ios. Flexibility has been achieved through the model's modular construction, through user-specified control parameters, input data files that can be easily modified, the use.of real costs and _ growth rates, and the ability of the model

.to aggregate and report costs in 'a variety of ways. _

The modular construction enables a user to avoid gathering and using input data .for_ calculations that are not of interest. For example, a user may wish

- to study the costs of treating radiation-induced cancers. The model's modular F construction enables him to skip the calculation of radiation injury.and gene-tic effect treatment costs, as well as the calculation of indirect costs. Only i those steps essential to calculating the direct cost'of cancers must be per- .l formed and onlye the data essential for performing these cal-culations is needed. l Execution of HECOM is' controlled by several parameters that define the e number of years of costs to process, the types of cancers,- radiation injuries

- and genetic effects to be included in cost calculations, and the number of. age

, categories and sexes _ defined in the input data. The value of each of these parameters can be specified by the user. The input data file can be easily modified to alter various economic (i.e., _ income and growth rates), demographic

-(1.e., cumulative' life' probabilities, labor force. participation rates and popu-lation fractions) and health effect data. This enables a user to easily run

, . different scenarios and-thereby ' develop a range of estimated health effect costs in addition to'a point estimate. -

. HECOM is designed to run with age and income data for. user-specified time

' intervals. The data may cover ten year age intervals, for instance, or the

- data.may consist of median values for the whole population. This allows HECOM to be run with available data at any level- of aggregation.

Costs calculated _ by HECOM are stored in the lowest level of aggregation

possible. This enables the model (with minor algorithm modifications) to .

aggregate costs in a variety of ways. For example, health effect costs could be aggregated and reported by' age cohort, type of illness, year of occurrence

! and sex, depending on the specific needs of the user.

c

^

T.1.2 : Treatment of Costs'Over Time t ~ The effects of radiation exposure are long-term, with both direct and indirect costs occurring over the lifetime of the affected population and suc-

.ceedi ng generati ons. To evaluate the merits of various measures that affect health effects risks, the. cost stream must be reduced to a single current dol-

- lar. estimate for the base year, the year in which action would be taken. This J' is accomplished by discounting the costs expected in each future year back to the base year. A present value estimate of both direct and indirect costs of

health effects ' projected in future years is calculated using the following basic approach:

4 7.2

-- a _ . ,~_ .- .. - _ - . - . - _ . , , . . _ .

z.

/

"t A EC(a-jtl)a s Present Value of Costs = {

a=j (1 + R/100,)n-1 where y 1 e3 a = ages of the affect.ed individual',Lfrom 0 to maximum (A) consideration '*

j = age at onset of morbidity or mortality

~

EC(a-j+1)a,s = expected cost in current dollars (given' direct or indirect cost levels, real escalation rates and esti-mated survival probabilities) for an individual of age  !

a and sex s, [for the (a-j+1) year after morbidity onset]

s = sex R = real discount rate (in integer form) n = year after population exposure.

The real oiscount rate used is an input parameter, thus facilitating sensitiv-ity analyses.

The time dimension of potential health effects also necessitates accommo-dation of changes in the levels of direct and indirect costs relative to the general rate of inflation, that is, changes in the real value of treatment costs and productivity losses. This is handled by~specifying the real escala-tion rates for treatment costs and productivity losses as input parameters.

, Expected costs of morbidity or mortality occurring in any given year are pro-jected as follows:

EC(n)a,s = C a,s P(a+1)a,s - (1+E)"~1 where EC(n)a,s = expected cost, or loss, in year (n) for an individual of age a and sex s Ca ,s = average cost, or loss, for an individual of age a and sex s P(a+1)a s = probability that an individual of a,e a and sex s would norm-ally survive to age a+1 E = real cost escalation rate.

7.3

i. -

7.2 -0VERVIEW OF HECOM STRUCTURE The algorithm developed to estimate the direct and indirect costs of health effects is described in this secticn. Figures that identify major com-putational processes and the types of data used to carry them out are provided to present a conceptual overview of the algorithm. Figure 7.1 shows the rela-tionships among the major algorithe processes. Each box represents a process and each line represents a flow of information. The remaining figures describe the individual processes shown in Figure 7.1 in more detail. Since the same processes appear in various figures, they are always shown in the same position l on the page.

7.2.1 Major HECOM Processes Health effect costs are calculated for five cost components: direct costs of cancer, radiation infuries and genetic effects, and indirect costs of ill- )

ness and fatalities. These cost calculations are represented by the five boxes in the middle row of Figure 7.1. Four intermediate processes are necessary to calculate these health effect costs: projection of genetic effect incidence, of cohort survival probabilities, of labor value over time and of fatalities ,

over time. These intermediate processes are represented in Figure 7.1 by the four boxes in the top row. The final step in the algorithm is to aggregate direct and indirect cost estimates into a form usable for analysis. This step is represented by the bottom box in Figure 7.1. The nu11ber of the figure which provides detail on each process is shown in parenthesis in each box.

m PROJECTION oF PROJECTION OF PROJECTION OF PROJECTION OF GENETIC COHORT SURVIVAL LABOR VALUE FATAUTIES INTERMEDIATE -

EFFECTS PROBA88uTIES oVER TIME OVER TIME PROCESSES (FIG. 7.8) (FIG. 7.10) (FIG. 7 91 (FIG. 7.7)

_ _J L_

4 5< rir ir ir2 , <r <r GENETIC EFFECTS RACtATION INJURY INDIRECT IND4 RECT CANCER DIRECT DIRECT COSTS OF COSTS OF DIRECT COST COSTS COSTS ILLNESS FATAUTIES COSTS CALCULATIONS (FIG. 7.41 (FlG. 7.3) (FIG. 7.8) (FIG 7.5) (FIG. 7.21 hi r D6 RECT AND

INDIRECT COST _

AGGREGATIONS FIGURE 7.1. Overview of Health Effects Cost Model Processes 7.4

1 The cost calculations;shown in the middle row of Figure 7.1 each represent a component of the tutal costs of health effects. Direct cancer costs are shown in the'righthand box of the middle row. These.are the costs of providing medical care to affected individuals at the point when the cancer develops and

~; is diagnosed.. While this cost component is referred to as cancer direct costs, it also includes the cost of treating benign thyroid nodules, since the incid-

- ence is available dirt.ctly from CRAC2 output. CRAC2 fatality projections. for  !

other cancers are converted to incidence by HECOM. Since the costs of treating cancer vary with age and sex, due to differing mortality probabilities, the model is designed to calculate direct costs by age, sex and cancer type. The j process is described in Section 7.2.2.

Radiation injury direct costs, shown in the second box from the left, ,

consist of the costs of providing medical care to persons with bone marrow, '

~

gastrointestinal, or pulmonary injuries or with prodromal symptoms. Costs of providing care to persons with growth and mental retardation due to prenatal exposure are also included. The calculational process is described in fection 7.2.3.

Direct costs of genetic effects are shown on the far left. The calcula-tion of these costs is explained in Section 7.2.4 While costs of caring for

- persons with genetic effects may stretch into the indefinite future, the costs are calculated as though all- future effects would occur within the first ten generations after population exposure. Direct costs for the portion of individuals _ assumed to be disabled by genetic effects include both acute medical care and institutional care costs.

' Indirect costs of fatalities are covered by the second box from the right. -While these indirect costs should include the value of all of an indiv-idual's productive activities, earnings data are presently being used in the _

HECOM base case, with only a partial correction for nonwage-earning labor. The indirect costs of fatalities depend on the sex and age of the deceased as well as other factors such as the rate of labor productivity increase over time.

The computational elements and general process for calculating indirect costs of fatalities are presented in Section 7.2.5.

[ The indirect costs of illness, shown in the center box, are similar to

. the indirect costs of fatalities except that generally they are of shorter duration. There is an exception in the case of prenatal injuries, which are assumed to prevent productive employment over the individual's lifetime.

Indirect = costs are calculated for the total incidence of cancers, rather than just the cancer fatalities projected by CRAC2. They also include losses during the period of. illness for those with radiation injuries. The calculation pro-cedure is explained in Section 7.2.6.  ;

The top row of Figure 7.1 shows the aa,ior processes that prepare the input data for use in the cost calculations. On the right-hand side, the projection of fatalities over time involves the calculation of cancer fatality probabil-ities in each subsequent year for each age and sex cohort depending on its

remaining expected lifetime. Based on these probabilities, the cancer fatality 7.5 i

4

~

---

..m - _.,e - . , _ . . . - . . . , , ,

[ ,,,,w-,--

,-..--, y _,,-y--m, , mm,,v-- i--e .~..----.e,,..-v -. -we- ,- - --,.#-.rm.,-.,.

I incidence from CRAC2 is distributed.over time. Acute fatalities are assigned to age and sex cohorts in proportion to their fraction of the population and are treated as occurring in the base year. Additional details of the procedure 1 are given in Section 7.2.7.

On the far left is a box. representing the projection of genetic effects i over. time. Procedures used to allocate genetic effects are explained in Sec-

. tion 7.2.8. Different types of genetic damage are treated as being eliminated from the population at different rates across generations. The genetic effects allocated to each succeeding generation, however, are treated as having an equal probability of occurrence.in each year of the 30-year generational period.

The projection of labor value over time is shown to right of center. A -

full description of the process is provided in Section 7.2.9. It is based on data for the median income of any specified number of median age categories.

When five-year age intervals are used, the cohorts are " aged" through succes-sive median age and income levels with labor value changing at some real rate over time.

To the left of center is the box representing the projection of cohort survival probabilities. This process is discussed in Section 7.2.10. Annual survival probabilities by sex and age are used to develop the cumulative sur-vival probability for each cohort as of the base year. These estimates are then applied to future labor value to calculate probable earnings in each year for each cohort.

7.2.2 Calculation of Cancer Direct costs Cancer direct costs are composed of the cost of treating cancers induced by radiation exposure. The information used to perform this calcu'ation is shown in Figure 7.2. To calculate total cancer direct costs by cancer type and

. sex, data from the intermediate process (which projects fatalities over time) are combined with data on the real treatment cost escalation rate, the discount rate, cancer treatment costs by cancer type, cancer incidence per fatality and duration of treatment.

The cost of treating each type of cancer in each subsequent year is deter-mined using base year treatment costs and the treatment cost escalation rate.

Incidence of cancer in each year after exposure is based on projected fatal-ities by cancer type and the ratio of cancer incidence to fatalities for each type of cancer. With this information, direct cancer costs are determined for each' year. These costs are then discounted back to the base year, using the discount rate. In the final process these data are aggregated to totals by sex and type of cancer.

7.2.3 Calculation of the Direct Costs of Radiation Injuries Direct costs of radiation injuries are composed of the costs of treatment for both the injured who survive and for fatalities. The flow of information involved in calculating radiation treatment costs is shown in Figure 7.3.

7.6

TREATMENT CANCER I

-- COST TREATMENT INCIDENCE TREATMENT f DISCOUNT DATA /

ESCALATION PATE ,

DURATION PER

, FATALITY COST f f

RATE PARAMETERS PROJECTION OF FATAUTIES INTERMEDIATE OVEA TIME PROCESS (FIG.7.7) ir CANCER COST ON DIRECT AND INDIRECT COST AGGREGATION (FIG.7.1)

FIGURE 7.2. HECOM Calculation of the Direct Costs of Cancers, by Sex and Cancer Type POeuuTiON TREATMENT RADIATION DATA /

FRACTIONS COST PER INJURY PARAMETERS INJURY INCIDENCE RADIATION INJURY DIRECT  : COST COSTS  : CALCULATION DIRECT ANO

_ INDIRECT COST AGGREGATION (FIG. 7.1)

FIGURE 7.3 HECOM Calculation Of the Direct Costs of Radiation Injuries, by Sex and Injury Type 7.7 e

+v a , - , - + - , --r,e,, ,-g.- ,., --,w,--- - - - - - - , , , - - , -v ,-- ,,-

These costs are based on data for the fraction of the population in each age cohort, treatment costs by injury type, and radiation injury incidence pro-jected by CRAC2. Direct radiation injury costs are assumed to occur only in the first year. Radiation injury incidence allocated according to each cohort's relative size in the population is combined with the treatment cost for each injury to estimate treatment costs by injury type. Finally, direct costs by sex and injury type are calculated.

7.2.4 Calculation of the Direct Costs of Genetic Effects Direct costs of genetic effects consist of the cost of treating persons born with severe genetic defects and institutionalizing them over their life-time. Inputs to the process include the number of persons requiring care in

.each year, the cost of treatment and institutionalization, the discount rate and the rate of treatment cost escalation.

Direct costs are calculated as the sum of lifetime expected institution-alization and treatment costs for each person born with a severe genetic defect. Expected institutionalization and treatment costs are based on cohort survival probabilities and the real costs of treatment and institutionalization in each year an individual is incapacitated. These costs are all discounted back to the base year using the discount rate. An overview of the process is provided in Figure 7.4 7.2.5 Calculation of Indirect Costs of Fatalities Indirect costs of fatalities represent the value of labor lost to society because of premature death. The flow of information used to perform this cal-culation is pictured in Figure 7.5. Data' from intermediate projections of cohort survival probabilities, of labor value over time, and of fatalities by age cohort, type of death and sex are used to calculate indirect costs.

"^

CosiGROb DISCOUNT A TI ANO RATE . RATE TREATMENT COST PROJECTION OF EGON OF PR S (FIG. 7.10)

OtRECT COST  :

of TREATING GENETIC EFFECTS FIGURE 7.4 HECOM Calculation of the Direct Costs of Genetic Effects 7.8

DISCOUNT ' DATA /

RATE PARAMETERS l

PROJECTION OF PROJECTION OF PROJECTION OF COf f0RT SURVIVAL LABOR VALUE FATAUTIES INTERMEDtATE PROBA88UTIES OVER TIME OVER TIME PROCESSES (FIG. 7.10) (FIG. 7.6) (FIG. 7.7)

INDIRECT COSTS OF COST

FATAUTIES CALCUt.ATION I

DIRECT AND INDIRECT COST AGGREGATION (FIG 7.1)

FIGURE 7.5. HECOM Calculation of Indirect Costs of Premature Mortality, by Age, Sex, and Cause of Death The labor value lost because of a fatality is the sum of projected annual labor values from the year of death to the maximum specified age of the indiv-idual. The calculation of labor value lost is based on projections of fatal-ities in each year by age category, cause of death and sex. These labor value

-losses are discounted back to the base year to approximate the indirect costs of fatalities. These indirect costs are then aggregated by age cohort, sex and cause of death.

7.2.6 Calculation of Indirect Costs of Illness Indirect costs of illness represent the value of labor lost due to ill-ness. The flow of information in this calculation is presented in Figure 7.6. Data from intermediate calculations of projected cohort survival prob-abilities, labor value over time.and fatalities over time are combined with data for the fraction of the population in each age cohort, radiation injury

. incidence, weeks of work missed, treatment duration, and the discount rate to calculate indirect illness costs.

Labor productivity loss is assumed to occur in the year prior to death.

Projections of fatalities in each year are combined with incidence to fatality ratios, labor value projections in each year and the number of weeks of work 7.9

CANCER INCL-POPULATION RADIATION WS OF OtSCOUNT TREATMENT OENCE P R ' DATA /

p,U FRACTIONS MATE OUMAIBUN PARAMETERS INC D NCE M S I I I PROJECTION OF PROJECTION oF PROJECTION OF PROJECTION OF GENETIC COHORT SURVIVAL LA8OR VALUE FATAUTIES INTERMEDIATE EFFECTS PROSA84UTIES oVER TIME OVER TIME CALCULATIONS (FIG. 7.81 (FIG 7.10) (FIG. 7.9) (FIG. 7.7) l I f

INDIRECT I

COSTS OF COST lLLNESS -

CALCULATION I

DIRtCT AND INDIRECT COST AGGREGATION (FIG. 7.1)

• FIGURE 7.6. ,HECOM Calculation of Indirect Costs of Illness, by Age, Sex, and Cause of Death lost for each type of cancer to calculate labor value loss by age category, year of illness, type of health effect and sex. Projections of cohort survival probabilities are used to adjust these loss estimates for the possibility that an individual will die from causes other than radiation-induced cancer.

Radiation injuries are assumed to occur in the base year only. Radiation injuries, allocated by ser., are apportioned to each age cohort according to its population fraction. The estimate of work weeks missed due to each type of l radiation injury is applied to the value of labor for each cohort to calculate labor value lost due to radiation injuries.

Indirect costs attributable to genetic effects represent the lifetime I

productivity loss for each person born with a severe genetic defect. Projec-l tions of persons born with several genetic effects in each year are combined t

with labor value projections to estimate the expected value of genetic effect productivity loss.

The indirect costs associated with genetic effects, cancers and radiation injuries are discounted to the base year using the discount rate. In the final step, indirect illness costs for cancer are summarized by sex, cause of death, and age category.

l 7.10 l-

7.2.7. Projection of Fatalities .

The" information used to project fatalities over time and the subsequent use of the fatality projections is presented in Figure 7.7. Input data to the calculation include population fractions with and without the in-utero age cohort, projections of cohort survival rates over time, fatality incidence from CRAC2, period of risk estimates, risk weighting factors, median survival times after diagnosis and minimtsn latency periods -for each type of cancer. The

. fatality projections are used to calculate indirect illness and fatality costs and direct cancer costs. -

The CRAC2 cancer fatality estimates are apportioned to age categories Dased on each cohort's fraction of the total population and each cohort's risk weighting factor. Acute fatality estimates from CRAC2 are apportioned to age categories using population fractions excluding the in-utero age category. All

. acute fatalities are treated as occurring in the first year after exposure.

Using the absolute risk model option, cancer fatalities are distributed so that the annual fatality rate is constant over each age cohort's years at risk. The first fatality is projected to occur in the year after the end of both the latency period and the mean survival period. The last fatality occurs in the year that the age cohort reaches the maximum age specified or the end of the period of risk. The end result of this process is a matrix of fatality projec-tions by age cohort, year of death, cause of death and sex.

RS WEl HTING FACTORS FRACTIONS POPULATION WITHOUT FATAUTY MNO DAW

$ gg FRACTIONS IN-UTERO INC10ENCE PERIOOS PARAMETERS COHORT TIMES l

N

+ ,, ,, , ,F PROJECTION oF  ;

COHORT SURVIVAL PROJECTION oF INTERMEDIATE PRotASIUTIES E^ ^

PROCESSES (FIG. 7.10)

OVER l

l l I

4 ,,

INDIRECT INDIRECT CANCER COSTS OF COSTS OF DIRECT COST ILLNESS FATAUTIES COSTS CALCUL$TIONS (FIG. 7.6) (FIG. 7.5) (FIG. 7.2)

FIGURE 7.7. HECOM Projection of Fatalities, by Age, Sex, and Cause of Death 7.11 L

l 7.2.8 Projection of Genetic Effects Figure 7.8 presents the flow of information into, and out of, the genetic effect projection process. Genetic effect incidence estimates, institution-alization rates, and genetic effect elimination rates are inputs to the pro-cess. The genetic effect projections are used to calculate both the direct and the indirect costs of illness due to genetic effects.

Institutionalization rates are used to determine the number of genetic effects that are so severe they will require treatment and institutional care. The elimination rates are used to allocate these genetic effects to each af fected generation. The incidence in each generation is then allocated equally to each year within the generations.

7.2.9 Projection of Labor Value Over Time The flow of information into, and out of, the labor value projection pro-cess is presented in Figure 7.9. Inputs to the proc'ss e are the rate of labor productivity growth, median earnings or other labor value data for each age category and the median age of each age cohort. The labor value projections by sex and age category are used to calculate indirect illness and fatality costs.

Labor value projections for each year after exposure are calculated for each age cohort by sex. Median labor value in each future year, for each age cohort, is calculated from base year median earnings by age cohort, the rate of real income growth, labor force participation rates and the earnings categories the original cohorts will belong to in each year after the base year. When a INCIDENCE INSTITUTION-ELIMINATION OF GENETIC ALIZATION RATE EFFECTS RATE i r PROJECTION OF  :

GENETIC C EFFECTS  :

, r i r DIRECT COST OF GENETIC EFFECTS

[OSTS OF ILLNESS (FIG. 7.4) i (FIG. 7.6)

FIGURE 7.8. HECOM Projection of Genetic Effects 7.12

v LA80"

,,,yc FRC' "g*CoME m, MEDIAN MED:AN DATA /

TION RATE - RATE IN oW AGE PARA M RS I

[ l l "j"o,CQ" uE ' INTERMEDIATE oVER TIME PROCESSES r

INDIRECT INDIRECT COSTS oF COSTS OF COST .

ILLNESS FATAUTIES CALCULADONS (FIG. 7.8) (FIG. 751

~ FIGURE 7.9. HECOM Projection of Labor Value, by Age and Sex cohort ages over a ti.ne interval (i.e., five years),'it is assigned the median earnings level of the cohort five years older with five years of labor productivity growth applied.

~

7.2.10 Projection of Cohort Survival Probabilities i Figure 7.10 presents the flow of information for projection of cohort survival probabilities and the cost calculations which use this infonnation.

Data on annual survivial probabilities by sex and the median age of each cohort are inputs to the process, which produces an array of life probabilities by age category, sex and year after the base year.

Data on annual survival probabilities (the probability that a person of ,

any age and sex will . live to the subsequent year) and the median age of each L cohort are combined to calculate the probability that a person in each cohort at the time of exposure will live over subsequent years.

7.3 MODIFICATION OF CRAC2 OUTPUT FOR USE AS HECOM INPUT

-Since the CRAC2 output was not designed'to facilitate calculation of health effects costs, some intermediate steps are required to create compatible health effect and cost categories. The definitions of health effects projected by CRAC2 and the steps required to use them are described below.

7.3.1 Acute Effects The 'CRAC2. projection of acute fatalities includes all deaths due to bon.e, lung, or gastrointestinal tract exposure. The projection is available as an aggregate, not by organ ' involved. The CRAC2 categories of acute fatalities and acute injuries are mutually exclusive and individuals are not double-counted within either. category.

7.13

- , --- --a-r,, m,e- - -- , ,.--,..!, ----+,,.m,,-s+~,-,- ,,n,-- ,y n , ,,m-~w,, -, ,-snge 4e n -,n ,- --.--,-

ANNUAL LIFE MEDIAN DATA /

PROBALITIES AGE PARAMETERS l PROJECTION OF 0 INTERMEDIATE COHORT SURVIVAL PROCESS PROBABILITIES  :

i r ,

r INDIRECT INDIRECT COSTS OF COSTS OF COST ILLNESS FATALITIES CALCULATION (FIG. 7.6) (FIG.7.5)

FIGURE 7.10. HECOM Projection of Cohort Survival Probabilities, by Age and Sex Since all acute fatalities occur within the first year, the indirect costs due to a fatality do not depend on the type of injury. Therefore, the HECOM indirect cost computation is based directly on the CRAC2 acuto fatalities esti-mate and is an aggregate for fatalities resulting from all of the types of radiation injuries.

Calculation of treatment costs for acute injuries is less straightforward because of the aggregate nature of CRAC2 injury estimates. CRAC2 does not provide estimates of serious injuries by type so that approximate treatment cost estimates can be applied. CRAC2 estimates do not include those who are injured (thus incurring costs) but die. Since all injured people would require treatment, this total is needed as the basis for the direct cost estimates.

Projections of acute radiation injuries produced by CRAC2 represent the number of persons likely to have either prodromal symptoms, gastrointestinal syndrome or lung impairment. These effects are not double-counted, though in actuality, people may have multiple injuries. Bone marrow and prenatal injuries are not included in the CRAC2 projection of acute injuries.

Since the effects of radiation injuries range from minor to life-threaten-ing, their treatment costs also vary considerably. To weight these costs, estimates are needed of the incidence of each type of radiation injury. This is calculated internally by CRAC2 for all injuries except bone marrow and pre-natal injuries, but disaggregated output is unavailable as an option.

7.14

PNL nas modifled the standard CRAC2 code to provide disaggregated esti-mates of acute injuries and fatalities. The modifications use the CRAC2 health effects data set for acute exposure in its present form. Fatalities are esti-mated for each exposure type as follows:

Fj = PE PBj where Fj = fatalities due to exposure type i PE = population exposed,"as calculated by CRAC2 PBj = fatality probability given the exposure level

.i = exposure type (i.e., bone marrow, etc.).

This modified calculation is performed for each exposure type for each evacuation scenario. Total fatalities for each start time are calculated as a weighted average over each evacuation scenario (as CRAC2 does currently for other early effects).

Injuries occurring in the population with exposures exceeding the fatality threshold are estimated as follows:

Ij = PE (1.0 - P84) where -

-19 = injuries of type i for people who are exposed above the fatality threshold but do not die.

The. injuries are only calculated if P8j is greater than zero. (If equal to zero, the fatality threshold was not reached.) The injuries are weighted by each evacuation scenario probability to estimated total injuries for the start time.

The calculation of injuries occurring in the population exposed to less than the fatality threshold also excludes people who die from fatal effects.

The calculation is:

Ij = PEI + PBj where Ij = nonfatal injuries of type j PEI = population exposed above the injury threshold P8j = injury probability given the exposure level.

7.15

a

- Total, injuries (by type) are estimated as ;a weighted ~ sum over all evacuation

-scenarios.

_ ~ To project prenatal-injuries, the assumption is made that the distribution of population age groups exposed to greater than 200 rems is the same as their proportions in the general population. The proportion of the general popula-tion "in-utero" is multiplied by the number of individuals with an exposure of

- over 200 rem to estimate the size of the group at. risk for prenatal injury.

. Based on the Reactor Safety Study (NRC 1975, Appendix VI o. F-21) an -incidence s rate for' prenatal injuries of 50 percent is applied to the group at risk.

7.3.2 Cancers CRAC2 estimates of' cancers are available in the form of fatality projec-

, tions for. leukemia, lung . bone, breast, gastrointestinal, and other cancers.

These fatality estimates are used directly in the HECOM calculation of indirect cancer costs. .To calculate direct costs, the CRAC2 fatality estimates must first be converted to estimates of can'cGr incidence. This conversion is car-

~

ried out within HECOM using the fatality / incidence relationships documented .in-

, - the Reactor. Safety Study (NRC 1975, Appendix VI). These ratios are listed in

Appendix A, -Table A.6. The resulting incidence estimates provide the basis for

. the, direct cost projections.

i

'The thyroid effects projected by CRAC2 are an incidence,' rather than fatality, estimate that-includes both benign and malignant nodules. Since the costs of treating these nodules differ, the CRAC2 thyroid projection is allo- i cated by HECOM to the two-types of effects in proportion to the relative spon-taneous -incidence of benign thyroid nodules and malignancies in the population

'(NRC 1975, Appendix VI p.- 9-27).

7.3.3 Genetic Effects L

When the option of calculating genetic effects with CRAC2 is implemented, the-resulting projection is an aggregate of all types of genetic disorders.

Since different types.of effects are eliminated from the population at differ-ent rates HECOM. allocates each type _of genetic effect across generations <

separately. To accomodate this level of disaggregation, the CRAC2 estimates must be allocated between genetic effect types before being input to HECOM. r p Currently, we are assuming two types of effects and an allocation of 50 percent

-of _ effects to each type.

7.4 HECOM SENSITIVITY ANALYSIS Some level of uncertainty exists in each of the input variables used to

- , estimate health effect costs. -A sensitivity analysis was performed to provide -

an indication of the significance of these uncertainties; it gives an illustra-tion of how costs would change in' response to variation in input estimates.

E  !

7.16

x. a . . ._ - _ _ . . . . . . _ . _ - - . . . _ . . _ , _ _ . _ _ _ _ . _ _ . _ _ _ _ . . . _ . _ _ _ -

l l

Sensitivity is generally measured by systematically varying the_ value of one input variable within the bounds of a range of uncertainty, while holding all other input variables constant. To measure sensitivity in HECOM we exam- l ined seven variables: the discount rate, rate of labor productivity growth, l rate of real growth in treatment cor.ts, base year earnings, base year treatment costs, weeks of work missed due to illness, and labor force participation rates.

The model-appears to be most sensitive to changes in the discount rate, the rate of treatment cost escalation, and the rate of labor productivity growth. The effect is most significant when the discount rate is assumed to be equal to the real growth rates of labor productivity and treatment costs. In i that case there is effectively no discounting of costs over time.

Regarding HECOM sensitivity to cost input data, both variations in earn-ings and treatment costs cause substantial changes in the HECOM cost estimates; i variations in the weeks of illness cause almost no effect. In the sections l that follow, the results of each sensitivity test are examined separately.

7.4.1 Sensitivity to the Discount Rate Table 7.1 shows the effect of different discount rates on the indirect, direct and total costs calculated by HECOM. The ten percent discount rate is mandated for use by the U.S. Office of Management and Budget and is used by the Nuclear Regulatory Commission. A four percent discount rate is used to repre-sent the social rate of discount. Rates of seven percent and one percent are also tested to explore fully the sensitivity of HECOM. As shown in Table 7.1, the model is clearly sensitive to the discount rate. Use of a seven or ten percent rate significantly lowers total costs because costs occurring in the years after initial exposure are given much less weight than similar costs occurring in the base year. A discount rate of one percent substantially increases costs because costs in the more distant future are given nearly the same weight as costs in the near future. All cost categories are strongly affected by use of a one percent discount rate because the costs of treating genetic effects over 300 years become relatively large.

7.4.2 Sensitivity to Labor Productivity Growth Rates The effect on indirect and total costs of varying the rate of labor pro-ductivity growth from its base case rate of one percent to a rate of three per-cent is shown in Table 7.2. The three percent rate results in a more than TABLE 7.1. Sensitivity of HECOM Estimates to the Discount Rate Discount Indirect Cost Direct Cost Total Cost Rate %  % a From Base  % a From Base  % a From Base

. 10 -73.1 -44.0 -58.4 7 -53.7 -31.7 -42.6 4 (Base) 0.0 0.0 0.0 1 299.8 264.6 282.0

! 7.17

l I

TABLE 7.2 Sencitivity of HECOM Estimates to the Rate of Labor Productivity Growth Rate of Labor Productivity Indirect Cost Total Cost Growth (%)  % a From Base  % a From Base 3 114.0 56.4 1 (Base) 0.0 0.0 100 percent increase in indirect costs and more than a 50 percent increase in total costs. The higher labor productivity growth rate causes the share of indirect costs as a percentage of total costs to rise substantially, 7.4.3 Sensitivity to Treatment Cost Escalation Table 7.3 shows relative costs of treatment calculated using the one percent base rate, and alternative rates of three. percent and five percent for real treatment cost escalation. Increasing the rate to three percent raises direct costs by over 80 percent, and further increasing the rate to five percent results in an increase of over 1400 percent for direct costs and 700 percent for total costs. This dramatic increase occurs because the rate of treatment cost growth exceeds the base case discount rate of four percent, resulting in very large genetic effect treatment costs over the 300 years following exposure. Because there is significant uncertainty regarding the future rate of growth for real treatment costs, HECOM estimates must be inter-preted carefully. Over the_ modeled period of 300 years, real costs of medical care for genetic disorders could either rise or fall and may well have a complex pattern of change.

TABLE 7.3 Sensitivity of HECOM Estimates to the Rate of Treatment Cost Escalation Rate of Treatment Direct Cost Total Cost Cost Growth %A From Base  % a From Base

5% 1,446.5 742.3 l

3% 81.4 41.3 l 1% (Base) 0.0 0.0 l 7.4.4 Sensitivity to Earnings Levels l

Table 7.4 shows the effects on indirect and total costs of a 20 percent variation in base year earnings levels. Indirect costs change in direct proportion to levels of base sear earnings. The potential error in total l health effect costs resulting from uncertainties in base year earnings esti-mates is approximately 10 percent.

l l

l 7.18 i

I

- ~. - - , , ~ . , , .,_L. .-. .~ . . , . , , , . - _ - - - - , . - - . . - . . . . - - - - - - . - . . . - -

4 TABLE 7.4 Sensitivity of HECOM Estimates to Earnings Levels Indirect Cost Total Cost Earnings  % a From Base  : a From Base Base Income plus 20% 20.0 9.8 Base Income 0.0 0.0 Base Income minus 20% -20.0 -9.8 7.4.5 Sensitivity to Treatment Costs The effect of uncertainties in treatment cost estimates is presented in Table 7.5. The range of uncertainty in treatment costs is estimated to be 30 percent. Varying treatment costs by 30 percent results.in an identical percentage change in direct costs and a 15.2 percent variation in total health effect costs.

TABLE 7.5 Sensitivity of HECOM Estimates to Treatment Costs Direct Cost Total Cost Treatment Costs  % a From Base  % a From Base Base plus 30% 30.0 15.2 Base 0.0 0.0 Base minus 30% -30.0 -15.2 7.4.6 Sensitivity to Weeks of Illness The uncertainty in estimates of weeks of work missed due to illness is estimated to be about 50 percent. Table 7.6 presents the effects on indirect and total health effect costs of a 50 percent variation in estimated weeks of illness. The results indicate that this variable is of only minor importance in determining indirect costs and that the high level of uncertainty in this variable leads to only a 1.9 percent margin of uncertainty in total health effect cost estimates.

TABLE 7.6 Sensitivity of HECOM Estimates to Weeks of Illness Indirect Cost Total Cost Weeks of Illness  % a From Base  % a From Base Base plus 50% 3.9 1.9 Base 0.0 0.0 Base minus 50% -3.9 -1.9 7.4.7 Sensitivity to Labor Force participation Rates Estimates of labor force participation rates are used by HECOM to deter-mine the expected value of population earnings. Labor force participation 7.19

3

' rates for each cohcrt were analyzed at a 100 percent level, and at the base case values given in Appendix A, Table A.4. The results of this variation on cost estimates are shown in Table 7.7. The results indicate that 100 percent participation in the labor force would increase indirect costs by about 20 percent and total costs by slightly over ten percent.

TABLE 7.7 Sensitivity of HECOM Estimates to Labor Force Participation Rates Labor Force Indirect Cost Total Cost Participation Rates  % a From Base  % a From Base 100% 20.6 10.2 Base

• 0.0 0.0 7.4.8 Comparison of Median and Interval Data Results Table 7.8 compares HECOM estimates based on median and interval case data. The median case represents the national median income, while the interval data case uses 18 age category-specific income estimates. Total costs in the median case are eight percent higher than the interval case. Direct costs in both cases are almost equal. Most of the difference in cost estimates occurs in the estimation of indirect costs where the median case estimate is 16 percent higher than the interval-estimate.

TABLE 7.8 Comparative Results of Median and Interval Data Cases Indirect Cost Direct Cost Total Cost Case  % a From Interval  % a From Interval % a From Interval Median 16.2 0.0 8.0 Interval 0.0 0.0 0.0 l-l 7.20

w-7.0 REFERENCE U.S.' Nuclear Regulatory Comission. 1975. Reactor e 'ety Study.. Appendix VI.

WASH-1400, National Technical Information. Service, ,, ingfield, Virginia.

k' e

k h .'

I i

[

b l

{ l l

l i-

\-

7.21 ewF-- '-We g e g -g gm ,= q r- - w -h ywweq--M-mm-.wege,ywe-gc,,-,g.myseg- i+ ,ry-w9-eywe+w g eg er eyerMTMw ew w m gr e w Mpr-e- w We tr y-4Myve 6-99---Wr*-rm'*w'%ummve'v*me

{

8.0 HEALTH EFFECT COSTS'FOR A HYP0THETICAL' REACTOR ACCIDENT HECOM has applications in various types of siting analyses, evaluations of safety goals and standards, and many. decisions related to the management of nuclear power. -An example of the output of HECOM for use in these types of applications is provided in this section. The numbers shown are derived from only one hypothetical accident scenario atia representative reactor, and until further-research is undertaken, it cannot be determined whether the order of magnitude of the costs is typical for other reactors, or even other accident scenarios' at the same reactor. .Thus, the estimates given should be treated .

only.as illustrative examples of HECOM's calculational capabilities.

8.1 HEALTH EFFECT ESTIMATES Probabilistic es' timates of health effects from a CRAC run for a given hypothetical reactor accident scenario were used as inputs to the HECOM cost

. calculation. The projected' numbers of each type of health effect are shown in Table 8.1.. These estimates are based on PNL's modification of the standard CRAC code;=the modification provides estimates for each of the categories of acute fatalities and injuries. Bone marrow injuries are included in the PNL modification, although omitted from standard CRAC analyses. An explanation of the code modification is provided in Section 7.3.1. Prenatal effects are  ;

, - calculated as a function (explained in Section 7.3.3) of the number of people exposed to over 200 rem, in this case, 3360. . Genetic effects are estimated on the basis of 260 per million rem of population exposure.

8.2 COST ESTIMATES-To project health effect costs,- the HECOM user must specify the desired real growth and discount rates. The following estimates assume one percent-annual growth rates for real income and health care costs and a discount rate

'of ten percent. Health effect costs are all shown in 1981 dollars.

HECOM projects costs for direct (treatment) costs and indirect (lost productivity) costs. These are shown for each category of health effect in Table 8.2j Health effect costs for this reactor accident scenario total

$7.6 x 10 . Other categories of cosgs for accident consequences at this- j reactor are: evacuat

~

reduction,8.3.64

$1.80 x 10 x 10gon,

land $3.14 interdiction, $1.53 xx 10 10; and

agricultural decontamination,-

Thus, the health costs are a substantial portion of the total 1 sses, $1.56 potential economic impact of a reactor accident. ,

e

• 8.1

., _ _ . . _ , _ . _.-.......~__m . - _ . . _ .___ . _ _ _ _ . _

1 TABLE 8.1. Project Numbers of Health Effects for One Reactor

- Accident Scenario' Used ac Input to the Sample HECOM Calculation Health Ef fects Number Cancers:

Leukemia +

18.8 Lung 23.5 Breast 69.2 Bone 5.9 Gastrointestinal 6.0 Other 16.3 Thyroid 43.7 Acute Fatalities:

Bone Marrow 331.2 Lung 13.8 Gastrointestinal 0

- Acute Injuries:

Bone Marrow 198.8 Prodromai 222.6 Lung 564.5(a)

Gastrointestinal 8.0 Prenatal 20.2 Genetic 616.2

-(a) The relatively large proportion of actue lung injuries is due to l

meteorological conditions in the singio scenario analyzed.

8.2

l TABLE 8.2. Projected Health Effect Costs for Sne Reactor Accident Scenario ( Thousande f 1981 $)

Cost Component Health Effect Direct Indirect Total Cost Cancers 404 -1,056 1,460 Acute Fatalities 34,223 36,535 70,758 and ' Inj uries Genetic 3.414 584 3,998 Total 38,041 38,175 76,216 O

+

8.3

APPENDIX A HEALTH EFFECTS COST MODEL

- _ - _ - - _ _ _ _-_ ___ _ _ _ . _. _ -. . l

APPENDIX A HEALTH EFFECTS COST MODEL The Health Effects Cost Model (HECOM) calculates the direct and indirect costs resulting from radiation-induced health effects. The model is written in FORTRAN-77 and is being maintained on a Digital Equipment Corporation VAX 11/780. An overview of HECOM is given in Section A.1, a description of input data and input file structure in Section A.2, a detailed explanation of subrou-tine and function operation in Section A.3 and a description of outputs in Section A.4 The FORTRAN source code is listed in Appendix B.

A.1 OVERVIEW This section presents an overview of HECOM. A description of each file within the model is provided along with a description of the process used in calculating health effect cost estimates.

A.1.1 HECOM Files HECOM consists of seven files: three FORTRAN files (source, object and executable), three data files and- a FORTRAN control file. The function of each file is described below:

e HECOM.FOR. FORTRAN source code containing all subroutines and func-tions.

e HECOM.0BJ. Object code produced by the FORTRAN compiler.

e- HECOM.EXE. Executable file produced by linker.

o HECOM1.DAT. Data file containing median case data.

o HECOM18.0AT. Data file containing interval case data by five-year age Cohorts.

e INDIST18.0AT. Data file containing risk weighting factors by five-year age cohorts.

• INDIST1.DAT. Data file containing risk weighting factors for the median case.

e DIST.DAT. Data file containing sr:vival probability data by year and sex.

CONTROL.FOR. Source code containing PARAMETER, C0!iMON, REAL, INTEGER and CHARACTER statements. This file is used to control execution of the model and is accessed by each subroutine. It is incorporated by the FORTRAN compiler into HECOM.0BJ.

l A.1 l

A.1.2 - Description of Program Operation Execution of HECOM is controlled by the file CONTROL.FOR. Parameters are

-assigned in the first two; lines ~of this file and are used.to dimension all arrays and to control processing of all loops within the main program and its subroutines. Health effect'. data inputs, except for cancer fatalities, are identified by specifying-the array index of either acute and thyroid death types or prenatal radiation injuries.- The parameters that must be set by the user ~are listed in Table A.1. All input data must be consistent with these

-parameters. The file CONTROL.FOR is accessed by each subroutine using the system command " INCLUDE '. CONTROL.FOR'".

Figure A.1 shows the structure of the model. The main program consists of a series of sequential subroutine call statements. The program begins by read-ing in data and ends by writing out health effects costs to an output file.

The functions of each subroutine and function shown in Figure A.1 are briefly described below. A detailed description of each subroutine and function is given in Section A.3 and the program listing in Appendix B.

• READER. Reads in input data from DIST.DAT and either HECOM1.0AT or HEC 0til8.DAT depending on the number of age categories set in CONTROL.FOR.

e SPROB.- Calculates the probability a person of either sex and a given age at time of exposure will be alive in any year after exposure.

Based on the survival probabilities contained in DIST.DAT.

• LATENCY. Calculates minimum latency periods for each cancer type.

TABLE A.I. HECOM Parameters Parameter Definition AC Number of age categories IC Number of income categories

~

GAC Number of genetic effect age categories IINC Income data interval (number of years)

AINC Age data interval (number of years)

GINC Genetic effect age category data interval (number of years)

YEARS Maximun age affected population can attain GYEARS Maximum years to project genetic effects NGEN Number of generations DTYPES Number of death causes SEX Number of sex categories (1 or 2)

HTYPES Number of cancer types r

RTYPES Number of radiation injuries ACUTE Set to acute death type THYROID Set to thyroid death type PRENAT Set to prenatal radiation injury type A.2

i MAIN  :

l CONTROL.FOR l f INDIST1.

DATf fHECOM1. I DATf

: READER  : '

j HECOM18.DATf ,

f DIST. DAT f fiNDIST18.DATf SPROB

; LATENCY I  ; FATAL

; DEATH

? PV RADCOSTh FV

CANCOST I INCCAT

LVALUE

: LOSTLV +-

+ PV T A WORK +

, GENDIST + INCCAT

GENCOST + + PV

; SUMUP + FV L WRITEUP

+

,, fOUTPUTf t FIGURE A.1. HECOM Structure t

( A.3 L

. _ . . _ _ _ . _ . . . _ . _ _ _ _ _ ._ _- ,, ...c._ . . _ _ _ _ _ _ _ . _ . _ - _ - _ _ _ . _ _ . . _ . _ . _ . . . _ . - _ _ . . . . . . -

e FATAL. Distributes fatalities to age categories.

e DEATH. Calculates fatalities per year by cause of death using optional constant absolute risk distribution model, e RADCOST. Cost of treating radiation injuries is calculated based on incidence and population fractions by age cohort.

e CANCOST. Cost of treating cancer is calculated based on the inci-dence to fatality ratios and the number of fatalities of each cancer type per year calculated in OEATH.

e LVALUE. ' Labor value by age cohort and year after radiation exposure is calculated based on wages by age cohort and the rate of real income growth.

n

,

• LOSTLV. Lifetime discounted earnings loss is calculated for a person dying in each year after exposure. Earnings calculated in LVALUE are multiplied by the probability that a person will be alive in any year and summed over the time period between the year of death and the maximum age specified in CONTROL.FOR.

• WORK. Calculates the value of work lost due to illness. Cost is based on fatalities per year calculated in DEATH, life probabilities calculated in LIFE, earnings calculated in LVALUE, incidence to fatality ratios, and weeks of work lost due to illness by health effect type.

e GENDIST. Distributes genetic effects to the years after radiation expusure.

e GENCOST. Calculates the present value of treating and institutional-izing individuals with genetic defects.

• SUMUP. Calculates total income loss based on fatalities calculated in"UTATH, and income loss in LOSTLV. This subroutine also calculates summary arrays used in printing results.

e WRITER. Prints out summary variables calculated in SUMVP.

e PV. Calculates present value of a number given the discount rate and

'Tiie number of years to be included.

e Fj[. Calculates future value of a number given the growth rate and number. of years to be included.

e -INCCAT. Determines earnings category of any age cohort.

A.4

4 A.2 0ESCRIPTION OF INPUT DATA FILES The data used as input to HECOM can be grouped in four general categories, including-information on population characteristics, healtn effects incidence, direct costs of treatment, and indirect costs of lost productivity. In this section, we describe sources for data in each of those categories, explain some of the merits and the limitations of particular data sources and describe the structure of HECOM input files.

A.2.1 -Population Characteristics HECOM er%oys descriptions of the population at risk: by age and sex categories, by mean earnings for the two sexes at each age, and by life expec-tancy at different ages.

-e Population by age and sex: Population counts, distributed into cohorts by sex and by age intervals, are used both in the allocation of health effects and in the estimation 'of indirect costs. On the national level, the most recent data of this type are from the 1970 Census. These data can be used if detailed local or regional data are unavailable. See Table A.2 for the proportional distribution of the US 1970 population by age and sex cohorts.

TABLE A.2. U.S. Urban and Rural Population Distribution by Age and Sex, 1970 ISercent3ce) O Aces Total Maie f* male All ages 100 48.7 51.3 In utero 1.1 0.6 0.6 1-4 7.3 3.3 . 3.5 5-9 9.9 5.0 4.3 10 - 14 10.3 5.1 5.1 15 - 19 9.4 4.3 4.7 20 - 24 7.9 3.8 4.1 25 - 29 6.6 3.2 3.4 30 - 34 5.5 2.3 2.9 35 - 39 5.5 2.7 2.3 40 - 44 5.9 2.9 3.0 45 - 49 6.0 2.9 3.1 50 - 54 5.5 2.6 2.3 55 - 59 4.9 2.3 2.6 60 - 64 4.3 2.0 2.3 55 - 69 3.4 1.5 1.3 70 - 74 2.7 1.1 1.5 15 - 79 1.9 0.3 1.1 80+ 2.1 0.1 1.2 Source: U.S. Bureau of the Census. 1973.

Co sus of population: 1970 Oetaileo Characteristics. Final Report Pcttj-01 United States SJunary. U.S. Government l- Printing Office. Wasntngton, D.C.

Table 1, p. 591.

l (a) Percentages by age for each sex represent j the age distribution for taat su within the total population i

1 A.5 L

Statistics from the 1980 Census are not yet available to describe the characteristics of the US population as a whole. A compendium of " provisional estimates" is now available, but tnose i ' estimates describe the population in grosser schema (for example, by 10-year age increments instead of 5-year) and the estimates include little information on income characteristics. The HECOM model can easily be run with information from the 1980 Census when appropriate data are available.

• Mean earnings by age and sex: This information is used in the esti-mation of lifetime expected earnings and thereby in the calculation of indirect costs. The input data may either be median or for any width age interval. In addition, either site-specific or national data can be used in HECOM. Table A.3 provides an example of the most recent national earnings data available. Mean earnin9s figures in 1981 dollars are listed by 5-year age increments and by sex (U.S.

Dept. of Commerce 1982.) We use the Consumer Price Index for "all items" for all urban consumers, to inflate the 1980 earnings esti-mates to 1981 dollars. (US Dept. of Labor, selected years.)

TABLE A.3. Mean Earnin3s of Employed Persons, by Age and Sex (19815)(a)

Ages Male Female 18 - 24(b) 7,431 5,211 25 - 29 15,696 9,542 30 - 34 19,833 9,898 35 - 39 23,173 9,892 40 - 44 23,597 9,975 45 - 49 24,445 9,921 50 - 54 23,570 9,979 55 - 59 23,055 9,844 60 - 64 19,205 9,443 65+ 9,080 4,590 Source: US Department of Commerce, Bureau of the Census. 1982.

Money Income of Households and Families and Persons in the United States: 1980. Current Population Reports, Series P-60, No. 132. U.S. Government Printing Of fice, Washington, D.C.

(a) 1980 incomes inflated to 1981 dollars by "All Items" index, Consumers Price Index for all urban consumers, as published by Bureau of Labor Statistics.

(b) Income for 18-24 year-olds was allocated to 15-20 and 20-25 year-olds based on the population weighted relationship between these categories and 18-24 year-olds' income in 1969. The same procedure was used ta compute income for 65-69, 70-74, 75-79, 80-85 age categories.

A .6

• Li fe expectancy:. Data on cumulative life probabilities are used to describe the life expectancies of individuals in the unexposed popu-lation, in cohorts distinguished according to age and sex. - Annual life probabilities are computed from the data in Table 5-1, Vital Statistics of the United States 1978, Volume II-Section 5, " Life Tables," p. 5.9. (National Center for Health Statistics 1980.) The 1978 life tables are the most recent currently available; the vital statistics life table data are typically 2 to 3 years old at the time of publication.

• Labor force participation rate: Based on an analysis by Hartunian, et. al. (1982, p. 49) these data are the average of employment and housekeeping pa'rticipation rates for 1970 and 1975 published in Employment and Earnings by the Bureau of Labor Statistics. Because 1970 was a high employment year and 1975 was a post-recession year, the average of the two years' rates was used to estimate expected labor force participation rates. The computed rates are listed in Table A.4.

TABLE A.4 Enployment and Housekeeping Participation Rates by Age and Sex (in %)

Ages Male Female 16 - 19 49.4 49.3 20 - 24 76.6 84.0 25 - 29 89.7 93.0 30 - 34 92.9 93.6 35 - 39 93.4 94.0 40 - 44 92.7 94.5 45 - 49 91.6 94.6 50 - 54 88.5 94.2 55 - 59 84.2 94.0 60-- 64 68.5. 91.8 65 - 69 35.8 88.3 70 - 74 17.9 78.0 75 - 79 9.3 74.5 80 - 84 5.3 73.4 85+ 3.5 73.0 Source: N. S. Hartunian, C. A. Smart and M. S. Thompson.

1982. The Incidence and Economic Costs of Major Health Impair-ments. Lexington Books, Lexington, Massachusetts, p. 49.

A.2.2 Health Effects Calculation of health effects costs requires data on incidence, latency periods, survival times, period of risk and relative risk by age and sex. Data for cancers and for radiation injuries and fatalities are described first.

A.7

~ - _ - - - . - . - .- - - - - - - - - -

I l

l l

• Incidence: HECOM requires incidence data for mortality and morbidity  ;

for each type of health effect. These data must be entered by sex if j the model is run using two sex categories.- Proportional allocation of the data by sex is computed prior to data entry. Incidence data for fatalities, injuries and cancers are taken directly from CRAC2 output, when availab.le, and calculated based on CRAC2 output in the remaining cases. Table A.5 shows the source and method used for each portion of the data. The incidence to fatality ratios applied to the CRAC2 cancer fatality estimates to calculate total cancer incidence are shown in Table A.6. These are the same ratios that are assumed by CRAC2 in projectin.g fatalities. The process by which CRAC2 acute injury estimates are'disaggregated by type is described in Section 7.3.2.

TABLE A.S. Health Effects Incidence Data Sources latent Effects Cancers Thyroid Acute Effects Morbidity Computed by HECOM from CRAC 2 output Computed by a modified Data CRAC2 output using inci- CRAC2 process dence to fatality ratios Mortality CRAC2 output None CRAC2 output Data TABLE A.6. Incidence / Fatality Ratios Applied to CRAC2 Fatality Projections Cause of Death _ Ratio Leukemia 1.00 Lung 1.00 Breast 2.00 Bone 1.25 Gastrointestinal 1.20

.Othe:* 2.00 Acute 1.00 Source: U.S. Nuclear Regulatory Commission. 1975.

Reactor Safety Study. Appendix VI. WASH-1400.

National Technical In formation Service, Washington, D.C., pp. G18-G23.

• Latency periods: Appendix VI of the Nuclear Regulatory Commission's Reactor Safety Study (Section G, p. G-23) is the source of data on minimum latency periods, by cancer site, for the population in utero and for all other ages. (NRC 1975.) See Table A.7 for a listing of the values used. There is no latency period for acute health effects.

A.8

TABLE A.7 Period of Latency for Selected Cancer Types Cancer Type In Utero All Other Leukemia- 0 2 Lung 15 15 Gastrointestinal 15 15 Breast 15 15 Bone 10 10 All Other 0 15 Thyroid 10- 10

. Source: US. Nuclear Regulatory Commission. 1975. Reactor Safety Study, Appendix VI. Wash-1400. Government Printing Of fice, Washington, D.C., p. G-23.

• Survival time: Median survival times are calculated, by cancer site from data in summary tables 1 and 2 in _ Cancer Patient Survival, Report Number 5. (National Institutes of Health 1976.) The median survival times input to the HECOM model are averages of the data reported for black and white population subgroups, weighted by the proportion of each cancer type attributable to that subgroup. The NIH data do not distinguish survival times by sex. Median survival times are presented in Table A.8. For radiation injuries the sur-vival time for all fatal. cases is less than a year.

TABLE A.8. Median Survival Time, 1960-1973 Median Survival (a)

Cancer Type Time (years)

Leukemia 1 Lung 1 Gastrointestinal 2 Breast 6 Bone 2 All Other 4 Thyroid 15 Source: National Institutes of Health, USOHHS. 1976.

Cancer Patient Survival, Report Number 5. NIH Publication No.81-992. Government Printing Of fice, Washington, D.C.

.(a) Averages for data for blacks and whites, weighted by proportion of each cencer type attributed to the racial subgroup.

A.9

L e Period of Risk: Estintates of the time period an individual exposed to radiation would be at risk for cancers are listed in Table A.9. These risk periods are~ used to allocate fatalities to the years after radiation exposure.

TABLE A.9. Period of Risk of Incurring Cancer After Exposure Cancer Tyoe Period of Risk Leukemia 30 years Lung li fetime Breast lifetime Bone .

30 years Gastrointestinal lifetime Other lifetime Thyriod 1ifetime Source: Committee on the Biological Effects of Ionizing Radiation. 1980. The Effects on Pooulations of Excosure to_ Low Levels of Ionizing Radiation. National Academy of Sciences, Wasnington, D.C., p. 243.

• Risk Weighting Factors: Risk weighting factors are used in conjunc-tion with population fraction data to allocate cancer incidence to each age and sex cohort. These data are BEIR III estimates'of excess cancer incidence resulting from radiation exposure. Values used are listed in Table A.10.

TABLE A.10. Risk Weighting Factors by Age and Sex f1 ALES aQe at Eteosure cancer Tvoes 0-9 to - u D - 34 35 - 49 so-Leutenta 3.38 1.85 2.60 1.92 4.32 Lung 0.00 0.54 2.45 5.10 6.79 Gastrointestinal 0.33 0.33 0.55 1.06 2.79 Breast 0.00 0.00 0.00 0.00 0.00 Bons 3.98 1.85 2.60 1.92 4.32 Otner 0.62 0.38 1.12 1.40 2.30 Thyroid 2.20 2.20 2.20 2.20 2.20 FEMALES Leukeinta 2.54 1.19 1.67 1.24 2.76 Lung 0.00 0.54 2.45 5.10 6.79 Gastrointestinal 0.33 0.33 0.65 1.06 2.79 Breast 0.00 7.30 6.60 6.60 6.6~

Bone 2.54 1.19 1.67 1.24 2.76 Other 0.62 0.38 1.12 1.40 2.30 Thyrofd $JO 5.80 5.80 5.80 5.80 Source: Comittee on the Biological Effects of Iontzing Raciation. 1990.

The Effects on Pooulettons of Eroosura to Lew tev.ls of itst rino aa oistien.

National <acemy of 5ciences. aasnington. J.c. Isole d-i4 ano e- U . pp. 50 and 256.

A.10

HECOM requires estimates of the total number of genetic effects of each type considered. Currently the base case treats only antosomal dominant and multifactorial" defects as -shown in Table A.11. An estimate.of the proportion of cases that are-severe is input to assign costs. The rate per generation at

which genetic defects are eliminated from the population must also be input to distribute incidence over time.

TABLE A.11. Genetic Effects Incidence Percentage of Total Elimination Rate Genetic Effect type- that are Severe Per Generation Autosomal-dominant 50% 20%

Multi factorial 50% 10%

A.2.3 Direct Costs Input data for direct costs of cancers and radiation injuries are required by cancer and injury type. . Methods used to develop the HECOM data base shown in Table A.12 are describ!d for radiatian injuries in Section 5.1 and for can-cers in Section 5.2. We have inflated estimates to 1981 dollars using The Consumer Price -Index " hospital room" component for hospital costs and the more general " medical care" component for other treatment costs. (U.S. Department

. of Labor, various years.) See Table A.13 for relevant components of the Con-sumer Price Index, for selected years.

, TABLE A A.12. Direct Costs of Health Effects (1981 $)

Cancers Treatment Cost

. Leukemia 16,300 Lung 17,400 Breast 9,400 Bone 37,600 Gastrointestinal 14,000 Other 11,400 Thyroid-benign 7,700 Thyroid-malignant 8,400 Radiation Injuries Prodromal 1,000 3one marrow 56,000 Gastrointestinal 28,000 Pulmonary 3,600 Prenatal (in utero) 100,000 A.11

TABLE A.13. Consumer Price Index, All Urban Consumers (1967 = 100) 1970 1975 1980 1981 CPI, all items 116.3 161.2 247.0 272.3 CPI, all services 121.6 166.6 270.9 306.2 All Medical Care 120.6 168.6 267.2 L295.1

.(services + commodities)

Medical care services - 124.2 179.1 288.9 318.6 Hospital Room 145.4 236.1 416.3 476.8 Source: Bureau of Labor Statistics, US Department of Labor, Consumer Price Index: Detailed Statistics. Published monthly.

U.S. Government Printing Office, Washington, D.C.

A.2.4 Indirect Costs The HECOM model treats both cancers and radiation injuries as resulting in indirect costs related to lost productivity for periods of morbidity and because of premature mortalities. Calculation of indirect costs due to morbid-ity requires data on work weeks lost. Data presented in Hartunian et al.

(1981, p. 236) are shown in Table A.14. The value associated with those lost weeks is computed using earnings data by age and sex category.

TABLE A.14. ' Mean Number of Work Weeks Lost by Cancer Patients During First Year' After Onset of Illness Number of Work Cancer Type Weeks Lost Leukemia (a)- 16.3 Lung 19.9 Gastrointestinal (D) 18.5 Breast 17.2 Bone 23.3 Al1 Others(D) 16.1

' Thyroid 5.9 Source: Hartunian N.S., et al. 1981. The Incidence and L Economic Costs of Major Health Impairments. Lexington Books, Lexington, Massachusetts, p. 236 l

(a) Simple mean for all types of leukemia.

(b) Mean for gastrointestinal and "others," weighted by relative share of component cancer sites.

A.12

.- - :. -z. -

For radiation injuries, the period of productivity' loss due to morbidity is estimated using information from Prasad 1974;_ Dalrymple 1973; and Blakely 1968 One week of work loss is assumed for prodromal injuries'and 26 weeks each for bone marrow, gastrointestinal and pulmonary injuries.

-A.3 DESCRIPTION OF SUBROUTINES AND FUNCTIONS HECOM contains 15 subroutines and 3 functions as well as the main pro-gram. The main program contains only subroutine call statements. A descrip-

-tion of each subroutine and. function follows; including the calculations used, a brief description of subroutine or. function operation and a listing of all variabir.l. and array's.

The subscripts used in the calculations follow the following conventions:

'a = age category of an individual in the year of exposure s =. sex n = number of years after base year d = cause of death (cancer and acute radiation injuries) c = cancer type r = radiation injury type t = genetic effect type The subroutine and function descriptions that follow are in the same order as used within HECOM (see Figure A.1). Section A.1 explains how each subrou-tire and function fits into the HECOM structure.

.A.3.1 READER This subroutine reads in input data. Lifetime. probability data are read from the file DIST.DAT. If the median data case is being run, data are read from HECOM1.DAT and risk weighting factor. data from INDIST1.DAT. If the inter-val data case is being run, data are read from HECOM18.DAT and risk weighting factor data from INDIST18.DAT.

i A.3.2 SPROB This subroutine calculates the prior probability that a person of a given

, sex and age category in the year of exposure would live to any subsequent

_ year. The only inputs to the subroutine are the probabilities that a person of

age a will live to age a + 1. The probability of living to any given year n is 1 calculated as follows

P(n)a,s = P(a + 1) s P(a + 2)a+1,s ... P(a + n)a+n-1,s t' where P(n)a,s = prior probability that a person who was of age a in the year of exposure and of sex s will be alive after n years l

l A.13 I

probability that a person of age a and sex s will live to P(a+1)a,s = be age a + 1.

The subroutine processes nested loops for age category and sex. Within the inner loop the median age of the age category is determined and used to calcul ate ' remaining fli fe period. A third loop processes years, and, nested within this loop is another that calculates the product of the conditional probabilities. This subroutine is used to compute the vector for a newborn infant. A separate vector of probabilities is calculated for use in determin-ing.the costs of' genetic effects.

.A.3.3 LATENCY, ,

This subroutine generates a matrix of latency periods by age category and cause of death. Latency periods are assigned identically to each age _ category for each cause of death except for the in _ utero age category that is assigned separately. The latency period information is constructed from input data.

A.3.4 FATAL This subroutine distributes fatalities by age category, cause of death and sex. Input items include population fractions by age category and sex (calcu-lated' for all. age categories and again for all age categories except in utero) risk weighting factors by cause of death, age and sex, and total fatalities by cause of death and -sex. Cancer fatalities are distributed based on both age category population fractions and risk weighting factors. - Acute fatalities are distributed based on age category population fractions excluding in-utero. The in-utero category is treated separately. While in-utero fatalities are not currently estimated in the HEC 0ti calculation, indirect costs of prenatal radia-tion exposure are calculated in this subroutine since these victims are too disabled to earn any income. Fatalities are calculated according to the fol-lowing equations:

Cancer Fatalities 2 AC i

F a,s,d = Fd(PF a,s RWF a,s,d)/{ [ (PF a,s RWF a,s,d) s=1 a=1 Indirect Costs, In-Utero Age Category F(in utero),s,( Acute) = RIF(Prenatal),s + RINF(Prenatal),s Acute Fatalv fes Fa ,s,d = F d(acute) PAa ,s A.14

.where Fa ,s,d = death dfatalities for age category a, sex s and cause of Fq= total fatalities by cancer fatality type d PFa ,s = population fraction of age category a and sex s RWFa ,s,d = risk weighting factor for age category a, sex s and cause of death d AC = number of_ age categories

.PAa ,s

• Population fraction (excluding in-utero)~ of age category a and sex s-l RIF(prenatal),s = fatal prenatal radiation injuries for sex s l

l RINF(prenatal),s = nonfatal prenatal' radiation injuries for sex s.

l A.3.6 DEATH (OPTION)

This subroutine calculates fatalities in each subsequent year by age cate-gory tause of death and sex. A provision exists within the subroutine to distribu.e deaths using alternative epidemiological models. At present, a I

constant absolute risk model is used. Inputs to the subroutine are the distri-i bution model option; fatalities by age category, cause of death and sex; life probabilities by age category and sex; periods of risk by cause of death; l latency periods by age category and cause of death; mean survival times by l cause of death and the median age of each age category. Except for acute fatalities, deaths are allocated to each year of a cohort's lifetime from the l end.of the minimum latent period and mean survival time to either the end of

l. the risk period or the maximum age attainable. Thus, cancer fatalities are allocated to the years within this period, weighted by the probability that an

, individual will be alive in each year. Acute fatalities are allocated equally l from the year;after the latency period to the year of the latency period plus i

mean survival time. (In application, all acute fatalities occur in the base year.) The calculations used to allocate fatalities are Acute Fatalities F(n)a,s,d = Fa,s,d /(MSd )

subject to LP 3,q < n ,$,MSd l'

A.15

All Others - ,

M l F(n)a,s d =Fa,d s

• b ")a-n,s/ a,s n=L5,d subject to LP ,d'+ MSd 4"' Ma ,s a

and LM a ,d = LPa,d + MSd where:

F(n)a,s,d = fatalities in year n, for age category a and sex s due to cause of death d P(n)a,s = probability that an individual of age category a and sex s will_be alive in year n Fa ,d.s = total fatalities for age category a and sex s due to

- cause of death d MSd = median survival time for cause of death d LP a ,d.= latency period for age category a and cause.of death d M remaining life period _or period of risk, whichever is a's = less, for age a and sex s at time of exposure a (M ,s = A -

median age a ,s)*

.The subroutine DEATH processes nested loops for sex, cause of death and age category. Within the-innermost loop the cause of death is checked to determine whether or not it is acute. If it is, fatalities are distributed based on the equation described above for acute fatalities. If the death type

-is not acute, fatalities are distributed to each year using the constant abso-lute risk model shown above for all others.

A.3.6' RADCOST This subroutine calculates the cost of treating radiation injuries. Input items are the cost of treating a radiation injury, the incidence of radiation injuries and the population fractions. Costs for all radiation injuries, except prenatal, are allocated to age categories based on their- population

-fractions.' Prenatal injuries are allocated entirely to the in-utero age cohort (evcept when running the median age case). The calculations for radiation i.. Jury treatment costs are Direct Costs, All Acute Radiation Injuries but Prenatal RC = (RIFp.+ RINFp) CPRp + PF a,r,s a,s A.16 l

Direct Costs, Prenatal -

RINFp = EREM - (PFp, male + PFp, female) 0.5 2

RC 1,p,s = (RIFp + RINFp ) CPR p

+ PF p,s/[ PF P'S s=1 where RCa ,r,s = cost of treating a radiation injury of type r, for _ age category a and sex s RIFp = fatal radiation injury incidence of type r RINF p = nonfatal radiation injury incidence of type r CPR p = treatment cost per case radiation injury r PFa ,s = population fraction for age category a, and sex s EREM = population exposed to over 200 rem p = prenatal.

A.3.7 CANCOST This subroutine calculates the cost of treating cancers. Inputs to the subroutine include the incidence to fatality ratio (upon which CRAC2 fatality estimates are based) for each cancer type; the cost to treat each type of cancer; fatalities per year for each age category, death type and sex and an array for associating cancer treatment costs with death types. The equation used to calculate health effects is:

A CH c . IPFc F(n)a,s d . (1 + T/100) '~1 HC(n)a,s,c = [

n=1 (1 + R/100)"~l where HC(n)a,s,c = present value of the cost of treating age category a and sex s for cancer type c in year n CHg = cost of treating one person for cancer type c IPFe = incidence to fatality ratio for cancer type c F(n)a,s,d = fatalities in year n for age category a, sex s and death type d A.17

R = real discount rate T = rate of treatment cost growth.

~

The subroutine processes nested loops for sex, age category, cancer types,

.and years. Within these loops real cancer treatment costs for each year after exposure-are calculated. In the next statement, these costs are. discounted and added to the cost accumulator for each age category, sex and cancer type. The function FV is used to calculate the -future value of base level ~ treatment costs. The function PV is used to calculate the present value of future treat-ment costs.

A.3.8 LVALUE This subroutine calculates labor value by age category and sex in each year after the base year. Inputs to the subroutine include income in the year of_ exposure by age category and sex, the median age-of each cohort, labor force participation rates and the rate of labor productivity growth. The following equation is used to determine labor value in each year:

L(n)a,s

• bII)b s . PR a,s(1 + W/100)n-1 where L(n)a',s = labor value in year n for age category a and sex s PRa ,s = labor force participation rate of age category a and sex s

' ~

W = rate of real earnings growth. .

This calculation is controlled by three loops which process sex, age category and years, respectively. Within these loops the age category of the group

being processed is determined using the function INCCAT, and real income in the l year being processed is calculated using the function FV, A.3.9 LOSTLV

.This subroutine calculates the present value of total lifetime labor value

( lost due to a premature fatality occurring in each year after population expo-l sure. Inputs to the subroutine are median age by sex and age cohort; annual j labor value by age category, sex, and number of. years after exposure; life probabilities in each year by age and sex; and the discount rate. Li fetime labor value loss is calculated using the following equation:

M a,s L(n)a,s

• P(n)a-n,s LL(n)*** =n=1 [ (1 + R/100)"~

l A.18

.where

-LL(n)a,s = 11_fetime labor value loss for a person dying n years after exposure of age category a in year of exposece, and sex s fia,s = less, for age category a and sex sremaining life period or period of risks, w 4_

L(n)a,s = labor value in year n for a person of age category a in the year of exposure, and sex s P(n)a,s = be alive in year nprobability that a person of age category a and se R = real discount rate y = number of years after base year to year of Dath.

.The subroutine processes nested loops for sex, age category and year after

-exposure. Within the inner loop the remaining life period is calculated and

- used as the termination year for accumulating real income loss. The function PV is used to discount each real income- figure .to the year of exposure as the income is accumulated. A person dying in a given year is assumed to lose ~ all income in that year and all subsequent years.until he would have reached an age

-equal to the maximum considered (A).

A.3.10 WORK This subroutine calculates the value of lost work time. Inputs to the subroutine include fatalities per year, incidence to fatality ratios, the num-ber of weeks-of work missed for each cause of death, income for each age cate-

' gory and the prior. probability that a person would be alive in each year after the year of exposure. The values of work lost due to cancers and due to radia-tion injuries are calculated separately using the following' equations:

Radiation ~ Injuries RINF LW PA -

! r e a.s L(1)a.s P(1)a.s RWC(1)a,s,r =

l 52

' Cancers Td F(n)a.s.d IPFd LNd .L(n)a.s - n)a,s

!. WC(n)a,s ,d =[Ty=1 52 - (1 + R/100)" Ty L

i where' o

-RWC(1)a,s,r = value of lost work in year 1, for age category a, and sex s for radiation injury r A.19 l

l

. .. _ . _ _ ..a._..____...._.2._, ._..,____,.._...__ _ _ _ , _ . _ . _ , . . - . - _ _ , _ . . . , . . . _ , . _ , , , _ - _ _ . .

- ~

- ..z ,,

I RINF r

= incidence of nonfatal radiation injury r LWp

• weeks of work lost for each type of radiation injury r PAa ,s = population fract' ion, excluding fr. utero, for age cata-gory a and sex s L(n)a,s = income in year n for age category a, and sex s P(n)a,s = probability that an individual in age category a, and sex s will be alive in year n WC(n)a,s,d = value of lost work in year n, for age category a, and sex s for cause of death d Td = years of treatment for cancer type d ,

Ty = sequential year of treatment F(n)a,s,d = fatalities in year n for age category a, and sex s for cause of death d IPFo = incidence to fatality ratio for cause of death d LW o = total weeks of work lost for each cancer type d R = real discount rate.  ;

The cost of lost work is calculated by processing nested loops for sex, age category, death type, treatment time and years. Treatment The time is inc1uded function PV is to spread lost work costs to more then one year if desired.

used to calculate the present value of income in any year. Radiation injury costs are calculated for the first year only.

A.3.11 GENDIST This subroutine distributes genetic effects to each year afterInputs radiation to exposure until the end of a user-specified genetic effect period.

the subroutine include total genetic effects by type, incidence fractions by The num-sex, and genetic effect elimination rates by type of genetic effect.

ber of generations is user specified. The first step in the subroutine is to calculate first generation effects according to the following equation:

G GE g,s,t

=(TG g .S,IR)/{g(1-O) t g

t where GEg .s,t = major genetic effects of type t occuring in generation g for sex s A.20

7

TGt = total genetic effects of type t Ss _= fraction of-genetic effects allocated to sex s

~IRt = institutional 1zation rate for genetic effect t.

-g = number of generations to a maximum of G Og = elimination rate for genetic effect t The second step is to project major genetic effects for each remaining generation according to the following equation:

GE g.s,t

= GE g- 1, s ,t (1-dt)

The final step is to allocate the effects for each generation equally to the years within the generational period. The subroutine performs each of these steps separately. The output of the subroutine is a matrix of genetic effects by year, type of effect and . sex.

A.3.12 GENCOST This subroutine calculates the present value of direct and indirect costs attributable to the genetic effects allocated in GENDIST. Inputs to the sub-

. routine include incidence of major genetic effects, median income, the cost of treating genetic effects, labor force participation rates and survival prob-abilities. Direct and indirect costs are calculated according to the following equations:

Di rect max

.y ars IC, GE - (1 + T/100)y+a-1 P(a)s y.s,t y,s,t

+

y+a=1- (1 + R/100)p a-1 Indirect max ye rs L(1)a,s

• PR,,, P(a) GE - (1 + W/100)y+a-1 y,3,g y,s,t y+a=1 (1 + R/100)y+a-1

.where DGy ,s,t = present value of direct costs of genetic effect type t, for a person born in year y, of sex s y = year of birth after year of population exposure A.21

\;

ICa = cost of institutionalizing and treating genetic effects of an individual in age category a P(a)s = probability _that an individual of sex s will live to age a GEy ,s,t = genetic effects of type t, occuring in year y-to sex s T = rate of treatment cost growth R = real discount rate y+a = years after birth in year y IGy .s,t = present value of indirect costs of genetic effect type t, in a person born in year y, of sex s L(1)a,s = median earnings in the base year for age category a and sex s PR a ,s = labor force participation cate for age category a and sex s W = cate of real earnings growth The subroutine processes nested loops for sex, genetic effect type and number of years in which effects occur. Within the innennost loop another loop accum-ulates the lifetime direct and indirect costs occurring to an individual born in each year. The subroutine assigns income and labor force participation rates in two ways, depending on the number of income categories being used.

A.3.13 SUMVP This~ subroutine calculates total income loss and summarizes direct and

. indirect cost data for reporting purposes. Income loss due to fatalities is determined for all causes of death except thyroid causes. The equation used to -

calculate income loss due to fatalities is FC(n) a,s ,d = F(n)a,s,d

• l'I"Ia .s where FC(n)a,s,d = lifetime, discounted real income loss in year n due to fatalities of type d, for age category a, and sex s F(n)a,s,d = fatalities occurring in year n for age category a and sex s due to cause of death d LL{n)a,s = discounted lifetime labor value loss of an individual in age category a and sex s dying n years after exposure.

A.22

+

Summary information is calculated by sex, age category, and health effect type, for direct and indirect costs.

A.3.14 WRITER This subroutine prints out ,results. The following summary tables are

' printed with subtotals for each sex.

e indirect cost due to fatalities by age cohort e indirect cost due to fatalities by cause of death e indirect cost due to fliness by age cohort e- indirect cost due to illness by cause of death e- indirect cost summary e direct cost of radiation injuries by injury type e direct cost of cancer by cancer type e direct cost sunmary e total cost simnary.

A.3.15.E This function determines the present value of a number based on the dis-count rate and number of years to be included. The followir.g equation is used:

py , V(nl (1 + R/100)"4 where PV = present value of number -

V(n) = value of number in year n ,

R = real discount rate n = years to be discounted back to the year of exposure.

l- A.3.16 E i

This function calculates the future value of a number given the initial

value, rate of growth and nunber of years in the future. The value is calcu-lated using the following equation ,

FV = V(b) - (1 + G/100)"-1 l where FV = future value of a number A.23 L. )

t

. V(b) = base level value of. a number G = real rate of growth n = year of future value determination after the year of exposure.

A.3.17 INCCAT This . function determines the income category of each age cohort. The function checks-first to see if the median data case is running. If it is, it returns an incc:ne category-of one (1) since this case has only one income cate-gory. Otherwise, it checks to see if the current age of the cohart is zero (in utero). If it is, it also returns an income category of one (1) since in utero is the first income category. If the age is not zero the income category is determined by comparing the age to the upper age boundary of each income cate-gory. When the age is determined to be greater than the upper age boundary of an income category, the function returns the number of that income category. A separate comparison is made for the in-utero income category because, unlike the other age categories, all its members are of the same age and do not change income categories in the same years as the other age categories.

A.4 OUTPUT Tables A.15 through A.24 provide samples of HECOM output for one case of each type of health effect. The health effect costs shown are those associated with the national data samples described in Section A.2. All costs are in 1981 dollars. The number of age categories; incidence of fatalities, radiation injuries, and cancers; rates of real income and health cost growth and discount rate are displayed, where appropriate, in the header above each table. Resul ts are printed for each sex and for-the entire exposed population. Indirect cost

, estimates are disaggregated by causes of death and age category. ' Direct costs l are disaggregated by type of illness only.

I A.24

TABLE A.15. HECOM Output: Indirect Costs Due to Fatalities HEALTHh EFFECTS COST MODEL NUMBER OF AGE CATEGORIESs18.0 RATE OF INCOME'GR0dTHs1.0 OISCOUNT RATES 10.0 l FATALITIES:

LEUKEMIA 1.0 LUNG 1.0 GI TRACT 1.0 BREAST 1.0 SONE 1.0 i

ALL OTHERS 1.0 THYROIO 0.0 ACUTE 3.0 PRENATAL 1.0 INDIRECT COSTS OUE TO FATALITIES DEATH CAUSE MALE FEMALE TOTAL

. LEUKEMIA 21597. 8227 29824 LUNG 2123. 1890 3813.

GI' TRACT 1799 1366. 3155.

l BREAST 0. 2651. 2651.

BONE 12862. 4715. 17577.

ALL OTHERS 1777. 1236, 3013.

THYROIO 0 O. 0

! ACUTE 192163. 111064 303227.

TOTAL LOSS

232312. 130949. 363261.

i I

l l

l 1 l l A.25

- _ . . . . __ . . _ . . _ . _ . . , - - __. _ . ._,_...._ _ _ _ _ - - . _ . . . = . - . - _ _ _ . _ _ . - - . _ . . . _ . _ _ . _ - _

TABLE A.16. HECOM Output: Indirect Costs Due to Fatalities by Age Category HEALTH EFFECTS C037 MODEL NUMBER OF AGE CATEGORIE$s18.0 RATE OF - INCOME growths 1.0 OISCOUNT RATES 10.0 FATALITIES:

LEUKEMIA 1.0

, LUNG 1.0 ' '

GI TRACT 1.0 SREAST 1.0 BONE 1.0

• ALL OTHERS 1.0 THYROID 0.0 ACUTE 3.0 PRENATAL 1.0 INDIRECT COSTS DUE TO FA7ALITIES AGE CATEGORY MALE FEMALE ,

TOTAL 1 16295 9469 25765.

2 7337 3226, 10563 3 13721. o321. 20042 4 15400 8654, 20054 5 20732, 11410 32141.

6 24324 14043 38367.

7 24075 12S53, 36928 8 22430, 10855, 33285 9 21078 10114 31192 10 20693 10265 30958 11 18098 9773 27871.

12 13864 8523 22387.

l 13 8794 6660 15455.

l 14 8082. 4477 8559 I 15 982. 2267 3249 l 16 309 1237 15d6 17 83 596 679 18 15. 204 219 TOTAL LOSS 232312. 130949 363261.

A.26 l

ey wy +,o.- ,-p -- ,g-g--y .- -r-----y-y-9w- p---vm e,ygymgy-- y

, TABLE A.17. HECOM Output: Indirect Costs Due to Illness HEALTH EFFECTS COST'MODEL NUMBER OF AGE CATEGORIESz18.0 RATE OF INCOME GR0wTHat.0 OISCOUNT RATES 10.0 ILLNESS AND INJURY INCIDENCE CANCER $t LEUKENIA 1.0 L

LUNG 1.0 GI TR'ACT 2.0 BREAST 1.3 BONE 1.2 ALL OTHERS 2.0 THYROIO.8ENIGN 0.6 THYROIO. MALIGNANT 0.4 RADIATION INJURIES:

PRODROMAL 1.0 BONE 1.0 LUNG 1.0 GI TRACT. 1.0 PRENATAL 1.0 INDIRECT COSTS DUE TO ILLNESS HEALTH EFFECT NALE FEMALE TOTAL CANCERS LEUKEHIA 631. 252. 883.

LUNG 145. 110 255 GI TRACT 197 167. 364 BREAST 0. 156 156.

BONE 628 254 882.

ALL OTHERS 165. 119 284 THYROIO 15. 22, 36.

RADIATION 3 PRODROMAL 87. 51. 138.

BONE 2266. 1316 3583.

LUNG 2266 1316, 3583 GI TRACT 22n6 1316 3563 DRENiTAL 0 O. O.

TOTAL LOSS 8666. 5080, 137'47.

A.27

TABLE A.18. HECOM Output: Indirect Costs Due to Illness by Age Category HEALTH EFFECTS COST MODEL NUMBER OF AGE CATEGORIESs18.0 RATE OF INCOME G90aTNat.0 OISCOUNT RATES 10.0 ILLNESS AND INJURY INCIDENCE CANCER $t LEUKEMIA 1.0 LUNG 1.0 GI TRACT 2.0 BREAST 1.3 SONE 1.2 ALL OTHERS 2.0 THYROIO. BENIGN 0.6 THY 40!O. MALIGNA.'47 0.4 RADIATION INJURIES:

PRODRONAL 1.0 SONE 1.0 .

LUNG 1.0 GI TRACT 1.0 PRENATAL 1.0 INDIRECT COSTS OUE TO ILLNESS AGE CATEGORY MALE ~ FEMALE TOTAL 1 7. 4 11.

2 47 22 69 3 123. 53. 176 4

111. 70 181.

5 269 175 444 6 691. 477 1167. -

7 895. 552 1447.

8 972 490, 1463.

9 1059 480 1539 10 1120 515. 1635.

11 1097 518 1615.

i 12 955. '

506 1460 13 745 449 1194 14 425, 368 795.

l 15 99 183 282 16 35 104 138 17 10 60 71.

18 4 Se, 60 TOTAL LOSS 8666 5080 13747.

! A.28

. l l

l l

TABLE A.19. HECOM Output: Indirect Cost Summary HEALTH EFFECTS COST MODEL

......................... l l

NUMBER OF AGE CATEGORIESalS.0  ;

NUMBER OF DEATH CATEGORIES 6.0 l NATE OF INCOME GROWTHS 1.0 OISC0ijNT RATES 10.0 INDIRECT COST

SUMMARY

HEALTH EFFECT MALE FEMALE TOTAL CANCERS a1928 20966 62895.

RAD INJ+ FATAL 199050 115063. 314113 GENETIC 596 351. 9a7.

TOTAL LOSS 24157a. 136360 37795a.

3 A.29

TABLE A.20. HECOM Output: Indirect Cost Summary by Age Category HEALTH EFFECTS COST MODEL NUMEER OF AGE CATEGORIESzi8.0 NUMSEE OF DEATH CATEGORIESs 8.0 9 ATE OF INCOME GPO4TH21.0 OISCOUNT RATE =10.0 INDIRECT COST

SUMMARY

AGE CATEGORY HALE FEMALE TOTAL 1 16303. 9473 25775.

2 7384 3248 10632, 3 13844, 6375 20219 4 15511. 6724 24235 5 21001, 11585 325a6 6 25015 14520, 39535 7 24970 13405 35375.

$ 23402 11345, 34747 9 22137, 10595 32731.

10 21812 10780 32593 11 19195 10290 29486 12 14819, 9028 23647 13 0540 7109, 16649 14 4510 4845 9355 15 1081 2450 .3532.

16 '

343 1341. 1684 17 93. 656 749 le 20 260 279 TOTAL LOSS 240978 136029 377007.

A.30

- _- . -. .. . . . . _ _ . , _ . . . . , - . . -. ~.

TABLE A.21. HECOM Output: Direct Costs of Radiation Injuries HEALTH EFFECTS COST N00EL NUMBER OF AGE CATEGORIESs18.0 RATE OF HEALTH COST GROWTHw1.0

,0ISCOUNT RATES 10.0 RADIATION INJURY INCIDhNCE '

PR00n0 MAL 1.0 SONE 2.0 LUNG 2.0 GI TRACT 2.0 PRENATAL 1.0 DIRECT COSTS OF RADIATION INJURIES INJURY TYPE MALE FEMALE TOTAL

'PRODRONAL u86 514.- 1000 SONE 54432. 57568 112000

' LUNG 3499 3701. 7200.

GI TRACT 27216. 2S784 56000

' PRENATAL 50001. 50001. 100002.

TOTAL LOSS 135634 140568 276202.

A.31

TABLE A.22. HECOM Output: Direct Costs of Cancers HEALTH EFFECTS COST MODEL NUMBEP OF AGE CATEGORIESale.0 RATE OF HEALTH COST.GR0aTHs1.0 DISCOUN,7 RATES 10.0 CANCER INCIDENCE:

LEUKEMIA 1.0 LUNG 1.0 GI TRACT 2.0 BREAST 1.3 SONE 1.2 ALL OTHERS 2.0 THYROID =8ENIGN 0.6 .

THYROIO. MALIGNANT 0.4 OIRECT COST OF CANCERS CANCER TYPE MALE FEMALE TOTAL LEUKEMIA 3755, 2529 6284 LUNG 1147 1213. 2360 GI TRACT 1701, 1826. 3528 i

SREAST 0 660 860 l SONE 5935, 3997. 9932.

l ALL OTHERS 1025 1057. 2082

THYROID.8ENIGN 63. 162. 229.

l THYROIO. MALIGNANT 46, 118. 164 TOTAL LOSS 13672. 11763. 25435 l

l A.32

TABLE A.23. HECOM Output: Direct Cost Summary i

HEALTH EFFECTS COST N00EL NUMBER OF AGE CATEGORIES =18.0 NUMBER OF CANCER TYPES = a.0 NUMSER OF RADIATION INJURIESs 5.0 RATE OF HEALTH COST GR0aTHs1.0 OISCOUNT RATE =10.0 l

l OIRECT COST

SUMMARY

)

HEALTH EFFECT MALE FEM 4LE TOTAL CANCEPS 13672. 11763. 25435.

RAD INJURIES 135634 140568, 276202.

GENETIC 2759 2781 5541. !

TOTAL LOSS 152066 155113 307176 l

l i

l A.33 t

TABLE A.24. HECOM Output: Total Cost Summary HEALTH EFFECTS COST H0 DEL NUMBER OF AGE CATEGORIEss18.0 NUMBE9 0F CANCER TYPES 8.0 NUMBER OF RADIATION INJURY TYPES 5.0 RATE OF INCOME GR0aTHs1.0 RATE OF HEALTH COST GR0dTHs1.0 OISCOUNT RATES 10.0 TOTAL' COST

SUMMARY

HEALTH EFFECT ~ MALE FEMALE TOTAL CANCERS 55601. 32729, 68330 RA0 INJURIES 334684 255e31. 590315 GENETIC 3355. 3132. e488

......... ....c.... .........

TOTAL LOSS 393640 291493. 685132.

s

, A.34

1

' CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC CC CC HAIN PROGRAM CONTAINING SUBROUTINE CALL STATEMFNTS CC

' CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC C

C-------------------------------------------------- ---------- ------------

C INSERT COMMON BLOCK AND PARAMETER INFORetATION FROM FILE

C-------------------------------------------------------------- -l------------

CONTROL.FOR8 INCLUDE ' CONTROL.FOR'

! C---------------------------------------------

t C CALL SUBROUTINE READEH TO READ IN DATA

C==-------------------------------------------

CALL READER l C-----------n-------------------------------------- ------

! C CALCULATE LIFE PRuBABILITIES FOR EACH AGE CATEGORY

! C-------------------------------------------------- ------

j CALL SPROB j C--------------------------------

4 C CALCULATE LATENCY PERIODS r m C--------------------------------

l -

CALL LATENCY

C------------------------------------------------ - -

1 i C CALCULATE CUST OF TREATING RADIATION INJURIES

C-------------------------------------------------- -

CALL RADCOST

! C-------------------------------------------------- -------

C CALCUALTE FATALITIES BY AGE CATEGORY AND DEATH TYPE C-------------------------------------------------- ------

CALL FATAL j C------------------------ ------------------------- ----------- -

C CALCULATE FATALITIES IN EACH YEAR USING SUBROUTINE DEATH.

j C-------------------------------------------------- ---------- -

CALL DEATH (1)

! C------------------------------------------------

i C CALCULATE COST OF TREATING HEALTH EFFECTS

! C------------------------------------------------

! CALL CANCOST

C-------------------------------------------------- ------------

} C CALCULATE LABOR VALUE FOR EACH AGE CATAGORY IN EACH

C-------------------------------------------------- -------YEAR ---

l 100 CALL LVALUE t

i i

i I. _

________---_-__A

C-------------------------------------------------- ---------- --------

t C CALCULATE LABOR VALUE. LOSS PER PERSON IN EACH YEAR.A DEAIH OCCURS l C-----------==---------------------------=------------========== --------

, CALL LOSTLV C=====----------===-----------------------------------------

C CALCULATE COST DUE TO WORK LOSS IN YEAR BEFORE DEATH

C==========-----------========----~~--------------- -=====--

CALL WORK C-----------=====---------------------*---

C CALCULATE GENETIC EFFECTS PER YEAR i C----====-------=-------------------------

CALL GENDIST

! C--------------------------------------------------

C CALCULATE DIRECT AND INDIRECT GENETIC COSTS C-=-----==------=====-==========------=------------

r CALL GENCOST C----==---==---=---------

C -

SUMMARIZE RESULTS 4 C---==-------====-----===

l o, CALL SUMUP

! k>

4

, C==----------------- ----

C PRINT 00T RESULTS 1

C------------------------

CALL WRITER l 200 STOP END CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC CC SUBROUTINE GENDIST '

1 CC j CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC 1

C C THIS SUBROUTINE DISTRIBUTES GENETIC EFFECTS TO EACH YEAR AFTER C EXPOSURE C

4 INCLUDE ' CONTROL.FOR' INTEGEH YPG I

i e

i-l l

C---------------------------------------------------

C CALCULATE THE NUHdER OF YEAHS PER GENERATION C===------====--------------------a===--------------

YPGaGYEARS/NGEN C---------==========---==-----------------

l C CALCULATE FIRST GENERATION EFFECTS

! C-----------------------------------------

l 00 2000 Is!,$EX l 00 2000 Jul,GTYPES DO 1000 Ka!,NGEN ~

C-=-------------------------

C SUM UP DECAY DIVISOR C==-------------------------

8UHsSUH+(1-DRATE(J))**(K-1)

~

i 1000 CONTINUE l C-------=--------------====------------------

C USE SUM TO COMPUTE FIRST YEAR EFFECTS C======---------=----------------------------

GEPG(J,1,1)s(G1(J)*GSRATE(I)*INRATE(J))/ SUM

' ~

C-------------------

C ZERO OUT SUN P' C-------------------

I SUHs0 2000 CONTINUE

! C---=====-----------------------------------------

} C CALCULATE EFFECTS IN REMAINING GENERATIONS j C-------------------------------------------------

00 3000 !al, SEX

DO 3000 Ja!,GTYPES 1 DO 3000 Km2,NGEN

! GEPG(J,K,1)sGEPG(J,K-1,1)*(1-DRATE(J))

} .

3000 CONTINUE C-------------------------------------------------- ----------- -----------

t C ALLOCATE GENEHATIONAL EFFECTS EQUALLY To EACH YEAR HIIHIN GENERATION j C====-------------------- ------------------------- ----------- -----------

t DO 4000 Is!, SEX i 00 4000 Jst,GTYPES i 00 4000 Kai,NGEN

} .

00 4000 Lal,YPG

GEPY(J,((K-1)*YPG)tL,1)sGEPG(J,K,1)/YPG

{ 4000 CONTINUE

RETURN i END 2

i

CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC-CC SUBROUTINE GENCOST CC CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC C

C THIS SUBROUTINE CALCULATES'THE DIRECT AND INDIRECT COST u,F C GENETIC EFFECTS C

INCLUDE ' CONTROL.FOR' REAL LABOR C

00 2000 Iat, SEX 00 2000 Jst,GTYPE8 00 2000 Kst,GYEARS DO 2000 Lal, YEARS-P' C---------------------------------------------

• C DETERMINE AGE CATEGORY BEING PROCESSED C---------------------------------------------

AGEaFLOAT(L-1)

KKaGINC KATsINCCAT(AGE,1,KK)

C-------------------------------------------------- -------

C SUM UP LIFETIME DIRECT COST OF INSTITUTIONAL AZATION C-------------------------------------------------- -------

DGCOST(J,K,1)aDGCOST(J,K,1)+PV(FV(INCOST(KAT),RHG,KtL-1) s *GLPROB(L,I)*GEPY(J,K,1),H,KtL-1)

C-------------------------------------------------- ---------- ---

C GET ST UP To CALCULAT INDIRECT COSTS. IF GAC AND AC ARE C EQUAL DETERMINE EARHINGS (EARN) AND LABOR F0HCE PARTICIPATION C (LABOR) BASED ON GENETIC. CATEGORIES.

C-------------------------------------------------- ----------- ----

IF (AC.NE.GAC) GOTO 1000 AGEmFLOAT(L) -

KATsINCCAT(AGE,1,KK)

EARNsHI(KAT,1)

LABORaLFPR(KAT,1)

GOTO 1800

3

C=====-------------- ---===------------------------ -===------ -----

C IF GAC AND AC ARE NOT EuuAL DETERMINE EARNINGS AND LABOR FORCE C PARTICIPATION RATES BASED ON INCOME CATEGORIES.

C-=---------------------------------------------------==------- -----

j 1000 CONTINUE l KKaIINC l AGEmFLOAT(L)

LLsINCCAT(AGE,1,KK)

EARNmMI(LL,1)

LA80RsLFPR(LL,I) 1800 CONTINUE C========-------------=---------------------------- -====----- -

! C SUM UP LIFETIME INDIRECT COST OF EXPECTED LABOR VALUE LO C-=---==------=--===----===--------~====----------- -==------- !S-

} IDGCOST(J,K,1)sIDGCOST(J,K,I)+PVCFV(EARN,HIG,K+L-1) j &

• LABOR *GLPHOB(L,1)*GEPY(J,K,1),R,K+L-1)

! EARNS 0 4

LABORS 0 2

2000 CONTINUE j RETURN i

." END CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC

. CC j SUBROUTINE RADCOST CC CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC C

i

C C

THIS SUBROUTItJE CALCULATES THE COST OF TREATI.NG RADIAYION INJURIES C

INCLUDE 'CONTHOL.FOR' 4

C-----------==------------------------------------- ---------- --------

C THE PRCNATAL RADIATION INCIDENCE VALUE REPRESENTS THE NUMBER OF i C PEOPLE EXPOSED TO OVEH 200 HEMS. THIS NUMBER HUST BE ADJUSTED TO C REPRESENT ONLY PRENATAL ItJURIES.

C==------------------------------------------------ ----------- --------

IF (PRENAT.NE.0) HADINF(PHENAT) RADINF(PHENAT)e

! & (PDPF(1,11+POPF(1,2))a0.5 4

l

4 4

4

! C======------=-====------------------------------ ---===----- ---------

C . COST IS CALCULATED BY HULTIPLYING COST PEN CASE BY THE NUMBER

! C. OF CASES AND DISTRIeUTING THESE COST 5 TO AGE CATEGORIES SASEO C' ON POPULATION FRACTIONS. IF THE RADIATION INJURY IS PRENATAL THEt, C ALL INJURIES ARE ASSUMEU T:0 SE IN UTERO.

4 C======---------=------=------------------------===---====----== ---------

l 00 100 Ist,RTYPES l DD 100 Jal,AC

00 100 Ks1, SEX IF (1.NE.PRENAT) RCOST(J,I,K)s(RADINF(I)+NADIF(I!)

• CPRADCI)*POPF(J,K)~

i &

! IF (I.EG.PRENAT.AND.J.EQ.1) RCOST(J,I,K)m(RADIF(I) i & 4RADINFCI))*CPRA0(1)*POPF(1,K)/(POPF(1,1)+POPF(5,2)) '

100 CONTINUE

l. RETURN END i CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC

! CC SUBROUTINE CANCOST

, CC

[ CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC C

C THIS SUBROUTINE CALCULATES THE PRESENT VALUE COST OF TREATING l C CANCERS RESULTING FROM RADIATION EXPOSURE.

! C

INCLUDE 8CONTHOL.FOR8 i C-------------------------------------------------- ----------- -----------

! C TREATMENT COST IN A YEAN IS CALCULATED 8Y HULTIPLYING THE FATALITIES

C IN THE NEXT YEAR HY THE INCIDENCE PER FATALITY AND THE REAL COST PER

, C INCIDENCE. THIS COST IS THEN DISCOUNTED TO THE YEAR OF {XPOSURE.

C-------------------------------------------------- -----=---- -----------

00 1000 Is!, SEX 00 1000 Jul,AC DO 1000 Kai,CTYPES l DD 1000 Lal, YEARS-1 j ZsFV(CPI (K),RHG,L)*IPF(K)*FPY(CFCONV(K),J,Lt1,1)

CCOST(J,K,I)sCCOST(J,K,1)tPV(Z,R,L) i Zs0 i

1000 CONTINUE l RETURN

! END l

i

C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C L C C C C C C C C C C C C C C C C C C C C C C C C C C C '; C CC SuoROUTINE SUMUP

CC .

CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC l C i C THIS SUBROUTINE SUMMARIZES FATALITY AND WORK LOSS DATA C

INCLUDE ' CONTROL.FOR8 C-------------------------------------------------- ----------- -------------

f 4

C CALCULATE TOTAL LA80R VALUE LOSSES AND SUMMARIZE LOSSES 48Y DEATH TYPE, C AGE CATEGORY AND YEAR.

C------------------- ------------------------------ ---------- -------------

} 00 2000 Lal,8EX 00 2000 Ja!,AC 00 1000 Ia!,YEANS 00 1000 Ka!,DTYPES i

I C---------------------------------------------------=---------- ------------

C CALCULATE TOTAL LABOR VALUE LOSS DUE TO FATALITI!S. IF DEATH C TYPE IS THYROID THEN 00 NOT PEHF0HM COMP i C-------------------------------------------------- -- UTATION. ------- ------------

l C------------------------------------------,-,h,L)zLYLOSS(J,I,L)*FPY(K,J,I,L) IF (K.NE. THYROID) TLYLOSS(J I j C SUMMARIZE LABOR VALUE LOST BY DEATH TYPE

C------------------------------------------------ -- ---

SDTLOSS(K,L)aSDTLOSS(K,L)tTLVLUSS J 1 K

! C-------------------------------------------------(-, ,-,L)

C SUMMARIZE LABOR VALUE LOSS BY AGE CATEGOR

---

Y C .,------------------------------------------------

SACLOSS(J,L)sSACLOSS(J,L)tTLVLOSS

! C------------------- -----------------------------(J,1,K,L) l C l C CALCULATE

SUMMARY

STATISTICS FOR LOSI WORK

C C SUMMARIZE LOSS BY DEATH TYPE il c-------------------------------------------------- ------

t i SDTLM(K,L)aSDTLh(K,L)+LWCCUST(J,I,K,L) l i <

I

)

1

C=-----------------------------------------

C SUMMARIZE LOSS by AGE CATEGORY C----------------------===--------------------

, SACCLWCJ,L)sSACCLW(J,L)+LHCCOST(J,1,K,L)

, 1000 CONTINUE C------------=------------------------------------=-----

, C CALCULATE

SUMMARY

STATISTICS FOR. RADIATION TREATHENT l

C AND ILLNESS C-------------------------------------------------- -------

DO 1400 Kst,RTYPES

, C---------------------------------------------

i C SUMMARIZE COST SY AGE CATEGORY i C==-------------------------------------------

SACRAD(J,L)sSACHAD(J,L)+RCOST(J,K,L)

SACRLH(J,L)sSACRLh(J,L)tLHNCOST(J,1,K,L)

C-------------------------------------------------- ---

C SUMMARIZE COST HY RADIATION INJURY TYPE

, C---------=-----------=---------------------------- --- ,

SRTRA0(K,L)mSRTRAD(h,L)+RCOST(J,F,L) l 8RTLW(K,LisSRTLN(K,L)+LWHCOST(J,1,K,L) t 1400 CONTINUE I 1" C---==== ------------------------------------------ -

l C CALCULATE

SUMMARY

STATISTICS FOR CANCER COSTS j C------------------- ------------------------------ -

1 00 1500 Kal,CTYPES

\ C------------------- :------------------------

C SUMMARIZE COST bY AGE CATEGORY C---------------------------------------------

8ACCAH(J,L)mSACCAN(J,L)+CCOST(J,K,L)

C--------------------------------------------

C SUMMARIZE COST BY HEALTH TYPE

C--------------------------------------------

SCTCAN(K,L)aSCTCAN(K,L)+CCOST(J,K,L)

~~

1500 CONTINUE l 2000 CONTINUE t

i i

I C---------------------===-------------------------- ----------- --------

C COMPUTE TOTALS FOR PRINT OUT FOR AGE CATECOR,4,ES AND TOTAhS BY SEX C AND FOR SEXEx COMBINED j C------------------------------------------------------n------- --------

4 0 .

DD 3000 Isi,AC 00 3000 Jan, SEX TACLOSSCIJsTACLOSS(I)+3ACLOSS(I,J)

TACCLW(!)=TACCLH(I)+8ACCLW(I,J)

TACRLWCI)sTACRLW(I)+SACRLNCI,J)

SACLW(I,J)sSACCLW(I,J)+SAthLh(I,J)

SLWCOST(JjaSt.WCOST(J)+SACCLW(I,J)tSACRLM(1,J)~

SCOST(J)msCOST(J)tSACLOSS(I,J)

TACLW(IJsTACLN(I)+SACLWCI,J)

TACRAD(I)sTACRAD(I)tSACHAD(1,J)

TACCAN(I)sTACCAN(I)+SACCAN(I,J) 80 RAD (J)s3DRAD(J)+SACRAD(I,J)

SDCAN(J)s$DCAN(J)+SACCAN(1,J)

~ ~

3000 CONTINUE C-------------------------------------------------- - --------- ------

C CALCULATE TOTAL LOSSES FOR FATALITIES AND ILLNE8S SY DEAIH TYPE l C-------------------------------------------------------------- ------

00 3500 Jai, SEX 00 3600 Is!,01YPES  ;

o, TDTLOSS(IJsTDTLOSS(I)+SDTLOSS(I,J) l L3 TDTLd(I)sTDTLNCI)+SDTLHCI,J) l 3600 CONTINUE C-------------------------------------------------- ----------- --------

C CALCULATE TOTAL LOSSES FOR ILLhESS AND RADIATIDN INJURY THEATHENT l C-------------------------------------------------- ----------

l DD 3700 Is!,RTYPES TRTRAD(I)=TkTRAD(I)+SRTRAD(1,J)

TRILWCI)sTRTLw(I)tSRTLh(I,J) 3700 CONTINUE C-------------------------------------------------- ----------- --------

C CALCULATE TOTAL LOSSES FOR ILLNESS AND RADIATION INJURY THCATHENT l C-------------------------------------------------- ----------

! 00 3800 Is!,CTYPES TCTCAN(I)sTCTCANCI)+SCTCAN(I,J) 3500 CONTINUE C---------------------------------------------

C CALCULATE TOTAL CANCER TREATMENT COSTS C--------==-----------------------------------

DO 3900 !al, SEX TCOSTalcobT+SCOST(I)

TLdCOSTsTLHCOST+SLHCOST(I)

TORADaTDRA9tSDRAD(I)

TDCAN=TDCANtSDCANCI) 3900 CONTINUE C************************=*******

C .

4 C CALCULATE GENETIC

SUMMARY

l C

C*******************************a DO 3950 Int, SEX DD 3940 Jat,GTYPES 00 3930 Kai,GYEARS 3DGTGEN(J,1)sSDGTGEN(J,I)tDGCOST(J,K,I)

SIDGTGEN(J,1)aSIDGTGEN(J,1)+IDGCOST(J,K,1) 3930 CONTINUE 80GENCI)s4DGEN(I)t3DGTGEN(J,I)

. SIDGEN(I)a810 GEN (I)+SIDGTGEN(J,1) l i

1

^

1 o3 3940 CONTINUE

_. TDGENsTDGENtSDGEN(I) c' i

TIOGEN=TIDGEN+SIDGEN(I)

~

3950 CONTINUE C************************************

1 C

] C COMPUTE INDIRECT COST

SUMMARY

! C Ca***********************************

C 4 C------------------- ------------------------------ ----------- ----------

C CALCULATE INDIRECT CANCER COSTS BY SEX AND TOTAL. ACUTE DEATHS ARE C NOT INCLuoED AS PART OF CANCER TOTAL.

C--------------------------------------------------- ---------- ----------

1 00 4000 I 1,DTYPES j IF (I.EQ. ACUTE) GOTO 4000 i TIOCANETIOCANtTDTLH(I)+TDTLOSs(I) -

4 00 4000 JuleSEX

SIDCAH(J)*SIDCAN(J)tSDTLW(1,J)tSOTLOSS(I,J) -

1 4000 CONTINUE i C==----------------- -==--------------------------- -

! C ADD ILLNESS COSTS TO INDIHECT RADIATION COSTS

! C=------------------------------------------------- -

l j

i _ _ _ _ _ _ _ _ _ _ _ _ _ . _ . .

T i

! 00 4050 Is!,RTYPES

TIDRADaTIDRADtTNTLWCI) l DO 4050 Jst, SEX SIDRAD(J)=SIDRAD(J)+SRTLWCI,J) i 4050 CONTINUE C-------======------------------------------------- --=--------

C ADD IN ACUTE FATALITY COSTS TO INDIRECT RADIATION COSTS C==------------------------------------------------ -----------

. TIDRADaTIDRA9tTDTLOSS(ACUTE) i DO 4060 1:1, SEX l SIDRAD(IjaSIDRA0(I)+SDTLOSS(ACUTE,1) j 4060 CONTINUE C-------------------------------------------------- --

C CALCULATE TOTAL INDIRECT COSTS BY AGE CATEGORY C-------------------------------------------------- --

00 4100 Is!,AC

!' TIDACCI)sTACCLW(!)+TACRLW(1)+TACLOSS(I) ,

00 4100 Ja!, SEX I

6IDAC(1,J)aSACCLN(I,J)+SACRLW(I,J)+SACLO! SCI,J)

P 4100 CONTINUE

~

C-------------------------------------------------- --

l C CALCULATE TOTAL AND SEX SPECIFIC It4 DIRECT COSTS

. C-------------------------------------------------- ---

! 00 4200 Int, SEX l SACID(I)sSIDRAD(I)+SIDCAN(I)

SID(I)aSIDHAD(13tSIDCAN(I)+SIDGEN(I)

TACIDsTACIDtSACID(I)

TIDaTID+SIDCI) 4200 CONTINUE j Ca*****************************>***********************

j C

C COMPUTE DIRECT COST SUMMA 4Y i C j Ca*****************************************************

i C 1 C-------------------------------------------------- -

! C CALCULATE TOTAL AND SEX SPECIFIC DIRECT COSTS 4 C--.---------.--------------------.---------------- -

l 00 5000 Iz1,AC 4

l

TDAC(I)sTACCAN(I)tTACRAD(I)

DD 5000 Jal, SEX SDAC(I,JJaSACCAN(I,J)+SACRAD(I,J)

SD(J)sSD(J)+SDACCI,J) 5000 CONTINUE' 00 5400 Ia!, SEX SD(I)aSD(I)+SDGENCI)

TDaTD+SD(I) 5400 CDNTINUE Ca**************************************************

! C C FINALLY, CALCULATE TOTAL HEALTH EFFECT COSTS C

I Ca**************************************************

DO 5500 Im1, SEX SCAN (1)mSIDCAN(!)tSDCAN(I) l SRAD(I)sSIDRADCI)+SDRAD(!)

SGEN(!)aSIDGEN(I)+SDGEN(I)

ST(I)aSDCI)+SID(!)

.I TCANsTCAN+ SCAR 4(I)

!= TRADaTRADtSRAD(I)

" TGENsTGEN+SGEN(I)

, TTalitST(1) 5500 CONTINUE

RETURN 5

END CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC CC SuuROUTINE WRITER CC CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC C

C SU8 ROUTINE TD WRITE OUT SUNHARY DATA C

C FIRST INCLUDE COMMON hLOCK AND CONTROL PARAHETERS j C i INCLUDE 'C0:4 TROL.FOR' C

C PRINT OUT HEADER C

t CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC CCCCCCCCCCCCCCC TEST WRITE SECTION CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC l

3

i GOTO.1234

DO 666 Int, SEX i DO 666 Jal,AC WRITE (6,1002) J, (PICK,J,I),Ka!,8) 666 CONTINUE MRITE(6,1101) l 00 667 Ia!, SEX

DO 667 Jat,AC l WRITE (6,1002) J, (F AT (K,J,1), Ks1,8) i 667 CONTINUE l GOTO 4321

, WRITE (6,1101) '

DO 771 Iai, YEARS 771 WRITE (6,1001) I,(LV(J,I,1),Jst,9)

, 1001 FORMAT (1X,12,9F10.0) i WRITE (6,1101)

! 00 772 Ia!, YEARS

772 WRITE (6,1001) I, (LvtJ,I,1),Js10,18) '

1002 FORMAT (1X,12,8Flo.6)

P'

! C '-

f WRITE (6,1101) ,

, DO 773 !al, YEARS 773 WRITE (6,1001) I,(LVLOSS(J,1,1),Jat,9) 3 WRITE (6,1101)

00 774 Isi, YEARS .

. 774 WRITE (6,1001) I,(LVLOSS(J,1,1),Ja10,18) ,

00 775 Ist,YE4HS

775 WRITE (6,1003) 1,LIFEP(I,1),(LPROB(J,1,1),Jst,8)

1003 FORMAT (1X,14,9F10.6) l DO 776 Isi,AC j NHITE(6,1101) j DO 776 Jst, YEARS

776 WRITE (6,1004) I,J,(TLVLOSS(1,J,K,1),Kat,8)

4321 WRITE (6,1101)

~

DO 779 Ist,AC NRITE(6,1101)

• I 00 779 Jat, YEARS

, 779 WHITE (6,1006) I,J,(FPY(K,1,J,1),Kai,8)

GOTO 5556 i '

j 1

4 WHITE (6,1101) i 00 777 Jst, SEX 00 777 Is!,AC

777 WRITE (6,1005) I,J,(LWRCOST(I,1,K,J),K=1,5)

HRITE(6,1101)

GOTO 5556

00 778 Int,AC i

778 WRITE (6,1009) I,(CCOST(1,J,1),Jst,CTYPES) 1009 FORMAT (1X,II,8F10.2) 1004 FORMAT (1X,214,8F10.2) 1005 FORMATI1X,214,5F10.2) i 1006 FORMAT (1X,2I4,8F10.5)

! 1101 FORMAT (////)

5556 CONTINUE 1234 WRITE (6,5) CASE C******************************************************

C

! C WRITE OUT FATALITY

SUMMARY

i C

C******************************************************

a3 WRITE (6,10) AC, RIG,R

_. WRITE (6,610) i

• DD 1000 Ist,DTYPES '

l IF (I.EQ.THYRUID) WRITE (6,811) OTNAMES(I),0.0

IF (I.NE. THYROID) WRITE (6,811) OTNAMES(I),CFCI) j 1000 CONTINUE 4 WRITE (6,811) RNAMES(PRENAT),(RADINF(PHENAT)+RADIF(PREHAT{}

j nRITE(6,20)

{ C------------------------------------------

i C WRITE OUT COSTS FOR EACH DEATH TYPE j C------------------------------------------

00 1100 Is!,0 TYPES

] WRITE (6,30) DTNAMES(I), (SDTLOSSCI,J),Jst, SEX),TDTLUSS(I{

j 1100 CONTINUE C=----~~------------===-

i C WRITE OUT TOTALS C==---------------------

WRITE (6,40) SCOSTC1),5 COST (2),TCOST ,

C------------------- ------------------------------ ----

j C NOW hRITE OUT COST BY AGE CATEGORY - FIRST H4ADER C-------------------------------------------------- -----

WRITE (6,10) AC, RIG,R-WRITE (6,810)

00 1150 Is!,DTYPES IF (I.EQ. THYR 0!D) MRITE(6,811) DTNAMES(I),0.0 I

IF (I.NE. THYROID) WRITE (6,811) OTNAHES(I),CF(I) 1150 CONTINUE WRITE (6,811) HNAMES(PRENAT),(RADINF(PRENAT)+NADIF(PHENAT))

! C------------------------------

C WRITE OUT TABLE CAPTION C==----------------------------

hRITE(6,50)

C======----===--------

C WRITE OUT DATA C==----------==----===

00 1200 Is!,AC

,' hRITE(6,60) I,(SACLOSS(I,J),Ja!, SEX),TACLOSS(I) 1200 CONTINUE 3

C-======----==---=======

} C WRITE OUT TOTALS

C-----------------------

l hRITE(6,70) SCOST(1),5 COST (2),TCOST

, C******************************************a*****

C 1 ui C WRITE OUT RESULTS FOR HISSED HOR'K

SUMMARY

C q C************a***********************************

WHITE (6,10) AC, RIG,R j NRITE(6,820) l HRITE(6,821) 00 1300 Im1,CTYPES j

WRITE (6,845) CNAHES(I),(CF(CFCONV(I))*IPFCI)).

1300 CONTINUE WRITE (6,825) li 00 1400 Is1,RTYPES 1 WRITE (6,811) RNAMES(I),RADINF(I) j 1400 CONTINUE

C------------------------ -----------------------------

l C WRITE OUT TABLE CAPTIONS FOR DEATH TYPE SUMMAR i C------------------------ ------------------------- --Y  :

l WRITE (e,120)

C-----------=------- ----------------------

f C WRITE OUT COSTS FUR EACH DEATH TYPE C-----------------------.------------------

WRITE (6,33) ' CANCERS: '

3 I

I

00 2000 Ial,0 TYPES IF (I.EG. ACUTE) GUTO 2000 MRITE(6,31) DTNAMES(I), (SDTLW(I,J), Jai, SEX),TOTLW(I) 2000 CONTINUE WRITE (6,33) 'HADIATION '

Do 2100 Is!,RTYPES WRITE (6,31) RNAMES(I), (8RTL%(1,J),Ja!, SEX),TRTLw(I) 2100 CONTINUE C---===-----------------

C WRITE OUT TOTALS C-----------------------

WRITE (6,40) SLWCOST(1),SLWCOST(2),TLWCOST C==========-------------------=-a------------------ -----

C NOW WRITE OUT COST BY AGE CATEGORY - FIRST HEADER C-------------------------------------------------- -----

WRITE (6,10) AC, RIG,R WRITE (6,820)

WRITE (6,821)

DO 2150 Is!,CTYPE3 WRITE (6,845) CNAMES(!),(CF(CFCONV(I))*IPF(I))

P G

2150 CONTINUE

! WRITE (6,825) l 00 2160 Ia!,RTYPES i WRITE (6,811) HNAMES(I),RADINF(I)

, 2160 CONTINUE

( C------------------------ -----

C WRITE OUT TABLE CAPTION C------------------------------

! WHITE (6,150) i C---------------------

! C WRITE OUT DATA ,

i C---------------------

DO 2200 In!,AC

WRITE (6,60) I,(SACLW(I,J),Jal, SEX),TACLW(I)

I 2200 CONTINUE C-----------------------

C WRITE OUT TOTALS

C-=------===------------

WRITE (6,70) SLWCOST(1),SLwCOS1(2),TLkCOST ,

C**************************************

C C WRITE DUT INDIRECT COST SUNHARY '

I C C**************************************

j NRITE(6,510) AC,DTYPES, RIG,R i C===----=---------==------------------------------- ---

l C WRITE OUT TABLE CAPTIONS FOR SUHHARY DEATH TYPE C------------------------------------------------------

. WRITE (6,520) l C------------------------------------------

C WRITE OUT COSTS FOR EACH DEATH TYPE C--------------==--------------------------

! WRITE (6,30) ' CANCERS ', (SIDCAN(J), Jet, SEX),TIDCAN

WRITE (6,34) 8 RAD INJ+ FATALE, (SIORAD(J),Ja!,$EX),TIDRAD l.

WRITE (6,30) 8 GENE 1.IC 8, (SIDGEN(J),Ja1,3@X),TIDGEN c-.--------------------

C MRITE OUT TOTALS C-----------------------

WRITE (6,40) SID(1),SID(a), TID i m C-------------------------------------------------- -----

l L C NOW WRITE OUT COST BY AGE CATEGURY - FIRST HEADER

C-------------------------------------------------------- .

l WRITE (6,510) AC,0 TYPES,HIG,R I

C=--===-===-=-===-===---==n----

!. C WRITE OUT TABLE CAPTION

! C------------------------------

! WRITE (6,550) i C=----------==---===-=

l C hRITE OUT DATA l C---------------------

l 00 3200 Ial,AC

WHITE (6,60) I,(51DAC(1,J),Jm1, sex),TIDAC(I)

! 3200 CONTINUE C----==-----------------

I C WRITE OUT T01ALs

C-----------------------

l HRITE(6,70) SACID(1),SACIO(2),1 ACID i C******************************************************

l C

! C WHITE OUT RESULTS FOR RADIATION TREATHENT COSTS i C l

Ca*****************************************************

i 1

WRITE (6,11) AC,RHG,H nRITE(6,830) 00 3300 Ist,RTYPE3 WRITE (6,811) RNAHES(I),(RADIF(I)tRADINF(I))

i 3300 CONTINUE C-------------------------------------------------- -------

C WRITE OUT TABLE CAPTIONS FOR RADIATION TYPE

SUMMARY

C-------------------------------------------------- -------

! MRITE(6,320) l C===------===-----------------------------a C WRITE OUT COSTS FOR EACH DEATH TYPE C------------------------------------------

, 00 4000 Is!,RTYPES '

WRITE (6,30) RNAMES(I), (SRTRAD(I,J),Jal, SEX),TRTRAD(I) 4000 CONTINUE C--------==---------====

C WRITE 00T TOTALS

! C-----------------------

i nRITE(6,40) SDRAD(1),SORAD(2),TORAD j m -

'C-------------------------------------------------- -----

C NOW MRITE OUT COST hY AGE CATEGORY - FIRST HEADER

C-------------------------------------------------- ----

, C WRITE (6,11) AC,HHG,R

'i C HRITE(6,830)

. C 00 4100 Ist,RTYPES i

C WRITE (6,811) HNAMES(I),(RADINF(!)+RADIF(I))

4100 CONTINUE i C------------------------------

i C WRITE OUT TABLE CAPTION C------------------------------

C WRITE (6,350)

C---------------------

C HRITE OUT DATA C---------------------

! C 00 4200 Iz1,AC I

C WRITE (6,60) I,(SACRAD(1,J),Jal, SEX),TACKAD(Il 1

4200 CONTINUE

C-----------------------

1 C MRITE OUT TOTALS i C-----------------------

i C WRITE (6,70) SDRAD(1),SDRAD(2),TOHAD i

l i

i __ _ _ _ _

p C***************************************************

C C hRITE OUT RESULTS FOR CANCER TREATHENT COSTS C

Ce**************************************************

WRITE (6,12) AC,HHG,R nRITE(6,840) 00 4500 Is!,CTYPEb ,

WHITE (6,845) CNAMES(I),(CF(CFCONv(I))*IPFfI))

I 4500 CONTINUE C==----===----------------------------------------- ---- -

C WRITE OUT TABLE CAPTInHS.FOR CANCER COST

SUMMARY

C-------------------------------------------------------

WRITE (6,420)

C---------------------==---------------

C WRITE OUT COSTS FOR EACH CANCEH C==-==-------------- ------------------

00 5000 Im1,CTYPES WRITE (6,430) CNAMES(I), (SCTCAN(I,J),Jal, SEX),TCTCAN(I) 5000 CONTINUE C-----------------------

G C WRITE OUT TOTALS C---------e==-----------

WRITE (6,440) SDCAN(1),SDCAN(2),TOCAN C--------------------------------------------------------

C NOM WRITE OUT COST BY AGE CATEGORY - FIRST HEADEN C -------------------------------------------------------

C WHITE (6,12) AC,RHG,R C WRITE (6,840) i l C 00 5100 Is!,Cf7 PES C WRITE (6,845) CNAMESCI),(CF(CFCONv(I))*IPF(I))

5100 CONTINUE C------------------------------

C WRITE OUT TABLE CAPTION C------------------------------

C WRITE (6,450)

C--------------------'

C WRITE OUT DATA

t C==----------====----=

C- .00 5200 Ist,AC C HRITE(6,60) I, ( S ACC AN(1., J), J u l, SEX ), T ACC AN (I) '

~

! 5200 CONTINUE C-----------------------

i C HRITE OUT TOTALS C=========.--===-========.

j _ C WRITE (6,70) SCAN (1), SCAN (2),TCAN j - C********************************************************

C WRITE OUT OIRECT COST

SUMMARY

C

C Ca*******************************************************

C

.C-----------------------

-C MHITE OUT HEADER C==---------------------

WRITE (6,610) AC,CTYPES,RTYPES,RHG,R -

C==-------------===------------

C WRITE OUT TABLE CAPTION

, C------------------------------

m WRITE (6,650) s o C---------------------

l C WRITE OUT DATA C---------------------

HRITE(6,30) ' CANCERS 8, (SDCAN(J), Jai, SEX),TDCAN WRITE (6,30) IRAD INJUHIES', (SONA0(J),Jul, SEX),TDRAD HRITE(6,30) ' GENETIC ', (SOGEN(J), Jai, SEX),TDGEN C---==------------------

C WHITE OUT TOTALS C-----------------------

I WRITE (6,40) SD(1),SD(2),TO J C-----------------------------------

j C WRITE OUT TOTAL COST SUHHARY C---------------------------------~~

i WRITE (6,710) AC,CTYPES,HTYPES, RIG,RHG,H

! C------------------------------

i C WRITE OUT TABLE CAPTION i C=~----------=------ ----------

! HRITE(6,750)

! C---------------------

1 C WHITE UUT DATA

c---------------------

J 1

> RITE (6,30) ' CANCERS .

', (SCAN (J),Jal, SEX),TCAN HRITE(6,30) ' RAD INJURIES', (SRAD(J),Jst, SEX),TRAD c

HRITE(6,30) eGENETIC ,8, (SGEN(J),Ja!, SEX),TGEN 'i C-------===========---==

C HRITE-0UT TOTALS C-=====-====-======---==

I 4 RITE (6,40) (ST(J),Jst, SEX),TT Ca***********************************

C FORMAT STATEMENTS FOR REPORTS

! Ca***********************************

I 5 FORMAT ('1',15(/),1X,40X,' HEALTH EFFECTS COST MODEL8,//,1X, 6 35X,'BATTELLE' PACIFIC NORTHWEST LABORATORIES',////,1X,35X,450) 10 FORMAT ('18,///,1X,' HEALTH EFFECTS COST MODEL,',/,1X,25(' '),//,

! & 1X,'NUNHER OF AGE CATEGORIESa',F4.1, i a /,1X,' RATE OF INCOME GROWTHa',F3.1,/,1X,' DISCOUNT RATES',F4.1) i 11 FOR'i AT ( 818, ///,1X, 'hE ALTH EFFECTS COST MODEL', /,1 X,25( 8 '

'), //,

} a2 & 1X,'NUHdER OF AGE CATEGORIESa',F4'.1,

! k>

~

& /,1X,' RATE OF HEALTH COST GR0HTHs',F3.1, l & /,1X,' DISCOUNT RATES',F4.1)

! 12 FORMAT (818,///,1X,8 HEALTH EFFECTS COST MODEL',/,1X,25(8 ,'),//,

i s 1X,8 NUMBER OF AGE CATEGORIESa',F4.1, j & /,1X,' HATE OF HEALTH COST GRONTHz8,F3.1, ,

l E /,1X,' DISCOUNT HATEu',F4..) -

L 20 FORMAT (1X,///,1X,15X,' INDIRECT COSTS OUE TO FATALITIES',//, '

& IX,'OEATH CAUSE',9X,8 MALE',7X,8 FEMALE',7X,eTOTAL'/

4 ,1X,11(8 '),9X,4(' '),7x,6(' '),7X,5(' 8))

30 FORMAT (1X,A12,2X,3F12.0) 31 FDRMAT(1X,2X,A12,jF12.0) 32 FORT 1Af(1X,/)

33 FORMAT (1X,Alo),

i 34 FORMAT (1X,A13,1X,3F12.0) i 40 FORMAT (1X,17X,9(8-8),3X,9(' '),3X,9(' '),

j s /,1X,' TOTAL LOSS',2X,3F12.0,///)

1 60 FORMAT (1X,4X,12,10X,3F12.0) i 50 F0HMAT(1X,///,'lX,15X,'1NDIRECT COSTS OUE TO FATALITIES 8,//,

] & 1x,' AGE CATEG0Hye,tox,edALE',7X,8FEHALE',7X,'T01AL8/

, & ,1X,12('='),10X,4(8 8),7X,6(8 8),7X,5(' 8))

70 FORMAT (1X,19X,9('-8),3X,9(' '),3X,9(' 8), '

i 1

i

, s /,1X,' TOTAL LOSSe,4X,3F12.0,///)

r . 120 FORMAT (1X,///,1X,15X,' INDIRECT COSTS DUE TO ILLNES$s,//,

[

& IX,' HEALTH EFFECT',7X,' MALES,7X,eFEMALE',7X,' TOTAL'/

k ,1X,13(8 8),7X,4(' '),7X,6(8 '),7X,5(' 83) 150 FORHAT(1X,///,1X,15X,' INDIRECT COSTS DUE TO ILLNESS',//,

& 1X,8 AGE CATEGORY 8,10X,8 HALE 8,7X,' FEMALE',7X,' TOTAL 8/

& ,1X,12(' '),10X,4(' '),7X,6(8 8),7x,S(8-8))

i 320 FORMAT (1X,///,1X,15X,801hECT COSTS OF RADIATION INJURIES!,//,

l 6 1X,' INJURY' TYPE',9X,' HALE',7X,' FEMALE',7X,' TOTAL'/

& ,1X,11(' '),9X,4(' 8),7X,6(' '),7X,5(' '))

I 350 FORMAT (1X,///,1X,15X,' DIRECT COST OF RADIATIDH INJURIES',//,

j & 1X,' AGE CATEGORY',10X,'NALE',7X,8FEHALE',7X,' TOTAL'/

j s ,1X,12(8 8),10X,4(8 8),7X,6(' 8),7x,5(' 81)

' 4 430 FORMAT (1X,A20,2X,3F12.0) 440 FORMAT (1X,25X,9(' '),3X,9(8 '),3X,9(' '),

& /,1X,' TOTAL LOSS 8,1oX,3F12.0,///)

l 420 FORMAT (1X,///,1X,15X,' DIRECT COST OF CANCERS',//,

, & 1X,8 CANCER TYPE',17X,8 HALE 8,7X,' FEMALE 8,7X,' TOTAL'/

4 ,1X,11(8 '),17X,4(' '),7X,6(8 '),7X,5(8 '))

450 FORMAT (1X,///,1X,iSX,' DIRECT COST 0F CANCERS',//,

i P & 1X,' AGE CATEGORY',10X,' HALE',7X,' FEMALE',7X,' TOTAL'/

y & ,1X,12(' '),10X,4(8 '),7X,6(' '),7x,5(' 83) j 510 FORMAT ('1",///,1X,' HEALTH EFFECTS COST H0 DEL',/,1X,25(8-!),//,

& 1X,' NUMBER OF AGE CATEGORIESz',F4.1,/,

& 1X,' NUMBER OF DEATH CATEGONIESa',F4.1,

& /,1X,8 HATE OF INCOME GRodTHz',F3.1,/,1X,' DISCOUNT RATES',F4.1) i 520 FORMAT (1X,///,1X,15X,'INDIHECT COST

SUMMARY

e,//,

l 6 1x,' HEALTH EFFECT',7X,' MALE',7x,8 FEMALE 8,7X,' TOTAL'/

i a ,1X,13(' 8),7X,4(' 8),7X,6(8 8),7X,5(' '))

! 550 FORMAT (1X,///,1X,15X,'INDINECT COST

SUMMARY

',//,

& 1X,8 AGE CATEGORf', lox,' MALE',7X,' FEMALE',7X,'TOTAle/

i & ,1X,12(' '),10X,4(' '),7X,6(' '),7x,5(' 8))

l 610 FORMAT ('18,///,1X,8 HEALTH EFFECTS COST MODEL',/,1X,25('-j),//,

& 1X,8 NUMBER OF AGE CATEGORIESz',F4.1,

& /,1X,'NUHdER OF CANCEH lYPESa',F4.1, j & /,1X,' NUMBER OF RADIATION INJURIESa',F4.1, j & /,1X,' RATE OF HEALTH COST GHodTHa',F3.1, j & /,1X,' DISCOUNT RATE ',F4.1) 4

650 FORMAT (1X,///,14,15X,80lRECT COST SUMHARY's//,.

& 1X,8 COST TYPE',11X,'H4LE',7X,' FEMALE',7X,' TOTAL'/

4 ,1X,9(8 8),11X,4(' '),7X,o(' '),7X,5(' 81) 710 FORMAT ('18,///,1X,' HEALTH EFFECTS COST HODEL',/,1X,25(8 '),//,

& 1X,8 NUMBER OF AGE CATEGORIESa',F4.1, 4 /,1X,8 NUMBER OF CANCER TYPESs',F4.1,

& /,1X,8NUH8ER OF RADIATIDH INJURY TYPESa',F4.1, .l

& /,1X,8 RATE OF INCUME GROWTHa',F3.1, l

& /,1X,' RATE-0F HEALTH COST GRodTHa',F3.1,  !

a /,1X,' DISCOUNT HATES',F4.1) 750 FORMAT (1X,///,1XaiSX,8 TOTAL COST

SUMMARY

',//, l

& 1X,' COST TYPE',11X,' HALE',7X,' FEMALE',7X,' TOTAL'/ (

& ,1X,9(8 '),11x,4(8 8),7X,6(' '),7x,5(s.s))

810 FORMAT (1X,/s1X,'FATALITIEst')

811 FORMAT (1X,3X,A12,10X,F6.1) 8E FORMAT (1X,/,1X,8!LLNESS AND INJURY INCIDENCE,')

821 FORMAT (1X,/,1X,'CANCERSI')

825 FORMAT (1X,/,1X,' RADIATION INJURIESt')

830 FORMAT (1X,/,1X,'RADIAT10h INJURY INCIDENCEs8).

840 FORMAT (1X,/,1X,' CANCER INCIDENCggs)

P 845 FORMAT (1X,3X,A20,)X,F5.1) y 5555 CONTINUE RETURN END CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC C

C SUBROUTINE READER C

CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC C

SUBROUTINE READER C

C THIS SUBROUTINE RpADS ALL INPUT DATA FOR EXECUTION OF HECOM C

C FIRST INCLUDE CONTROL FILE C

INCLUDE 'CONTHOL.FOR' C====-----===-------------------------------------- ----------

C DETERMINE WHICH FILE SHOULD BE OPENED DEPENDING ON THE NUnuEH C 0F AGE CATEGORIES C------------------- ------------------------------ ----------- ----

IF (AC.EG.18) OPEN (UNITz2, FILES HEC 0H18.DAT',STA10Sa'OLD')

s

IF (AC.EQ.1) OPEN (Ur11Ts2, FILES 'HEcotti .D A T ', S T ATUSs ' OLD'l

.C-----------==------------------------------------- ----------- ----------

j C READ IN DATA l C-------------------------------------------------------------- ---------- I READ (2,30) CASE READ (2,*) RHG READ (2,*) RIG READ (2,*) R i

READ (2,4) (MI(1,1),Is!,AC)

READ (2,4) (HI(I,2),Is!,AC)

READ (2,*) (LFPR (I,~ 1), Is !, AC)

READ (2,*) (LFPR(I,2),Is!,AC)

READ (2,*) (POPF(I,1),Is!,AC)

R E AD (.2, * ) (POPF(I,2),Ist,AC)

READ (2,*) (POPFACI,1),Is!,AC)

READ (2,*) (POPFA(I,2),Is!,AC) m READ (2,*) (MAGE (I,1),Is!,AC) b READ (2,*) (MAGE (I,2),Is!,AC)

READ (2,*) (LPU(I),Is!,DTYPES)

READ (2,*) (LP0(I),Ial,DTYPES) l READ (2,*) (CF(I),Isi,DTYPES) l READ (2,*) (PORCI),Is!,DTYPES) l NEAD(2,*) (MS(I),Isi,0TfPES)  !

READ (2,*) (LWORK(I),Is!,DTYPES) l HEAD (2,*) (TREAT (I),Iul,DTYPES)

C---=---------------------------

C READ IN DEATH TYPE NAhES C-------------------------------

DD 1000 Is!,DTYPES READ (2,10) DTt1AHESCI) 1000 CONTINUE C------------------------

C READ HEALTH DA!A C-----------------------

HEAD (2,*) (CPI (I),Is!,CTYPES)

READ (2,*) (IPFCI),Is!,CTYPES)

READ (2,*) (CFCONV(I),Is!,CTYPES)

D0 1100 Is!,CTYPES i

r READ (2,20) CHAMES(I')

! 1100 CONTINUE C------------------------------------

i C READ IN RADIATI0ff INJURY DATA C----------------~=====--------------

READ (2,*) (LHORKR(I),Is!,RTYPES) j READ (2,*)-(RADINF(I),Isi,RTYPES)

READ (2,*) (RADIF(1), Int,RTYPES)

READ (2,*) (CPRAD(I),1:1,RTYPES) 00 1300 Ist,RTYPES .

READ (2,10) RNAMES(I) 1300 CONTINUE C==---------------------------------

C READ IN GENETIC EFFECTS DATA

! C-----------------------------------

READ (2,*) (INCOST(I),Is!,GAC)

READ (2,*) (GI(IJ,Isi,GTYPES)

READ (2,*) ( GS R A T E ('! ) , I s i , S E X )

. READ (2,*) (0RATh(1),Iai,GTYPES)

~

READ (2,*) (INRATE(I),Is!,GTYPES)

P DO 1400 Ial,GTYPES y READ (2,20) GNAMES,(I) 1400 CONTINUE C-----------------

C CLOSE FILE C-----------------

CLOSE (UNITa2)

C-------------------------------------------------- ----------- --

C OPEN FILE CONTAINING DEATH DISTRIBUTION DATA AND READ IT IN l C-------------------------------------------------- ----------- --

l OPEN (UNITa3,FILEz'DIST.DAT',STATUSs'OLD')

00 2000 Is!, YEARS -

i t

READ (3,*) (LIFEP(I,J),Jal, SEX) l i 2000 CONTINUE C-===-----------==

! C CLOSE FILE

C-----------------

CLOSE (UNITS 3) -

i I

C-----------------=====---------------------------- ---

C' OPEN FILE CONTAINING INCIDENT DI'STRIBUTION DATA C-===-----------------==--------------------------- --- '

IF (AC.EQ.18) OPEN (UNITa4, FILES'INDIST18.DAT',STATU$a80LDs)

IF-(AC.EW 1) OPEN (UNITa4,FILEa'INDIST1.DAT',$TATUSa'OL@')

C-----------------------------------~~-

C HEAD INCIDENT DISTRIBUTION. DATA C------------------------------------==

00 3000 Is!, SEX _

00 3000 Jal,DTYPES -

READ (4, *). (IDIST (J,k /D ,Ka t , AC)

! C WRITE (6,*) (IDIST(J,K,1),Kat,AC)

3000 CONTINUE C-----====-----==-

C CLOSE FILE

C-----------------

l CLOSE-(UNITm4).

i 10 FORMAT (A12) l 20 FORMAT (A20)

. 30 FORMAT (A50)

. RETURN l

g END CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC CC

, SUBROUTINE LATENCY CC -

) CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC

C

C THIS SUBHOUTINE. CALCULATES THE LATENCY PERIOD FOR EACH

C CAUSE OF DEATH AND AGE CATEGORY

C i INCLUDE 'CONTHOL.FOR'

C------------------- -------------------------------

l C CALCULATE LATENCY PERIOD FOR EACH AGE OPTION i C

C FIRST FOR AGE CATEGORY EQUAL 1

C---------------------------------------------------

^

IF (AC.NE.1) GOTO 2000 i DO 1000 Im1,DTYPES LP(I,1)sLP0(!)

l 1000 CONTINUE l GOTQ 3000 i

C-----------------------=------------------------8 C COMPUTE LATENCY PERIOD FOR AGE CATEGORYz1 C--------=--------========------------------------

2000 CONTINUE' DO 2500 Is!,DTYPES LP(I,1)sLPU(I) 00 2500 Js2,AC LP(I,J)sLP0(I) 2500 CONTINUE 3000 CONTINUE RETURN END CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC9CCCCCCCC CC i

SUBROUTINE FATAL CC m CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC9CCCCCCCC

'm C C THIS SUBROUTINE CALCULATES FATALITIES DY AGE CATEGORY l C AND CAUSE OF DEATH.

C INCLUDE 8 CONTROL. FORE REAL SUH(DTYPES) ~ ---------

C---==---------------===---------------------------

C ----------IDENCE FIRST, CALCULATE THE SUN OF (POPULATION FRACTION

• INC DIST)

C FOR EACH CAUSE 0F DEATH.

C--------------=====--======----------------------- -----------

00 1000 Is!,DTYPES 00 1000 Jal, SEX 00 1000 hat,AC SUN (I)sSUM(I)+POPF(K,J)*1DIST(1,K,J) 1000 CONTINUE C-------------------------------------------------- -----------

C CALAULATE PERCENT INCIDENCE OF CANCEH'F0H EACH SEX AND AGE ----TEGORY CA C--------------------------------------------------

DO 3500 Kal, SEX 00 3500 Ist,AC 00 3500 Jat,0 TYPES PI(J,I,dja(POPF(1,K)alu1ST(J,I,K))/ SUM (J{

u

9 i

l- C---------------------------- --------------------- ---------- ----------

.C CHECK TO SEE IF ACUTE IS A DEATH-TYPE. IF IT'IS C THEN OVERRIDE POPULATION FRACTION WITH NUH IN UTERO C POPULATION FRACTIONS.

C--------------------------------~~---------------- ----------- *----------

IF (J.EU. ACUTE) PI(J,1,K)a(PDPFACI,K)*IDIST(J,1,K))

& / SUM (J) 3500 CONTINUE C=======------------------------------------------- -----------

C CALCULATE. DEATHS FOR EACH CANCER TYPE Ah0 AGE CATEGORY.

C-------------------------------------------------- -----------

DO 4000 Kai, SEX 00 4000 Jm1,0 TYPES DO 4000 Isi,AC FAT (J,1,K) SPI (J,1,K)aCF(J) ,

4000 t CONTINUE C==------------------------------------------------ ----------- -------

C ASSIGN PRENATAL RADIATION INJURIES TO IN UTEHO, ACUTE FATALITES. l

. C THIS IS DONE dECAUSE PRENATAL RA0!ATION VICTIMS ARE ASSUAED TO i E$ C NEVER BE AHLE TO WDHK.

I C-------------------------------------------------- ----------- -------

i 00 5000 Is!,$EX _

IF (ACUTE.NE.O.) F AT( ACUTE,1,I) a(R AD1t!F (PHEN AT) +R ADIF (PREN AT )) *

& POPF(1,1)/(POPF(1,1)+POPF(1,2))

5000 CONTINUE i RETURN END CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC CC SUBROUTINE WORK .

CC CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC C

C THIS SUBROUTINE CALCULATES THE LAHOR VALUE LOST DUE TO C MISSED h0RK C

INCLUDE 'CONTHOL.F0H'

4 t

., C I DO S000 Lal, SEX 00 5000 Is!,AC i 00 4000 Jal,DTYPES C=------------------------------------------------------------- -------

l C IF DTYPE IS ACUTE THEN SKIP THE CALCULATION. ACUTE WORK LOSS IS

! C CALCULATED SEPARATELY FUH EACH RADIATION INJURY.

! C----------------------- a------------------------- ----------- -------

i IF (J.EG. ACUTE) GUTO 4000 C-------------------------------------------------- ---------- --------

l C DEATH TYPE IS NOT ACUTE S0 COMPUTE LOST'WORh BASED ON FAIALITIES.

i C-------------------------------------------------- ----------- --------

j DO 4000 Hal, TREAT (J)

, DO 4000 Kate.2ARS-H

! C---------------------=---------------------------- ----------- -------

} C IF DEATH TYPE IS THYROID THEN FATALITIES ARE ACTUALLY INCIDENCE.

j '

C IF DEATH TYPE IS NOT THYROID THEN INCLUDE INCIDENCE TO FATALITY C RATIO IN COST CALCULATION.

C-------------------------------------------------- ---------- -------

w IF (J.EQ.THYRUID) GOTn 3000 L LWCC03T(I,K,J,L)sFPY(J,I,K+ti,L)*IPF(J)*LHORh(J)/52.0*

• 6 PV(LV(I,K,L),H,K)*LPROB(I,K+ti,L) i

GOTO 4000 3000 CONTINUE-

! LHCCOST(1,K,J,L)sFPY(J,I,K+H,L)*LWORK(J)/52.0=

& PV(LV(I,K,L),H,K)*LPROB(I,n+H,L) 4000 CONTINUE C-------------------------------------------------- ----------- ----

C CALCULATE WORK LOSS FOR EACH RADIATION IdJURY BY AGE CATEGORY C-------------------------------------------------- ---~~------ ----

! DO S000 Jal,RTYPEb

, LMRCOST(I,1,J,L)aRADINF(J)*LwoRKR(J)/52.0*PICACUTE,1,L)*

i

& PVtLV(1,1,L),R,1)*LPR0b(I,1,L) l 5000 CONTINUE i HETURN END

, CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC

CC 1 SUBROUTIrlE SPROB

! CC CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCT,CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC -

1 i

I

i

. C

. C THIS SUBROUTINE CALCULATES THE PROHAulLITY THAT A HEMBER OF I C AN' AGE CATEG0Hf WILL BE ALIVE.IN A YEAR - INDEPENDENT OF ANY DEA 1HS C CAUSED BY RADIAiluN EXPOSURE.

C

!' INCLUDE 'CONTHOL.FOR8

~

INTEGER RLP C===========-------------------==---

C THE FIRST. LOOP PROCESSES SEX 4

C------------------- ---------------

DO S000 ISEXal, SEX 4

C,--======-------------------=---------------------

} C THE SECOND LOOP PROCESSES EACH AGE CATEGORY

! C--------------------------------------------------

! DO S000 Is!,AC

C-------------------------------------------------- ------------

C CONVERT MEDIAN AGE AND HEHAINING LIFE PERIOD TO INTEGERS i C---------------------===-------------------------- ------------

! MAsINT(HAGE(1,ISEX))

l RLPzYEARS-MA

! P C-------------------------------------------------- -------

w

, C3 i

1 4

5 i C IF IN UTERO SET MAm1 TO HATCH IT UP HITH FIRST PROB i C-------------------------------------------------- ------

! IF (MA.EQ.0) MAnt 4

C-------------------------------------

i C THE THIRO LOOP PROCESSES YEARS ,

i C-------------------------------------

l D0 5000 Jst,RLP

, C-------------------------------------------------- -

i C INITILIZE CURRENT PROGABILITY OF LIVING TO 1.0 i C====---=------------------------------------------ --

) LPROB(I,J,ISEX)=1.0 C-------------------------------------------------------------- --

! C CALCULATE PRODUCT OF CONDITIONAL PHOBAdILITIES FROM YEAR OF l C EXPOSURE TO CURRENT YEAR (J)

! C-------------------------------------------------- ----------- --

2 00 5000 LAMA,MAtJ-1

] LPR06(I,J,ISEX) LPRnb(1,J,ISEX)*LIFEP(L,I. SEX) i 1

i

5000 CONTINUE C

C 00 6000 Int, SEX GLPR08(1,1)aLIFEP(1,3) 00 6000 Js2, YEARS' -

1 GLPROB(J,1)aGLPHOB(J-1,1)*LIFEP(J,I) 6000 CONTINUE RETURN -

END CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC CC i SUBROUTINE LVALUE

! CC CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC INCLUDE 8 CONTROL.FOR' INTEGER A,8 C

00 2000 ISEXst, SEX

00 2000 Aa!,AC j m DO 2000 Is!, YEARS .

4 L C-------------------------------------------------- ----------- --------

3 C DETERHINE LA80R VALUE CATEGORY OF SOMEONE WHO IS ItHAGE Y 1

C-------------------------------------------------- ----------- , EARS OLD

BsINCCAT(ItHAGE(A,ISEX),4,IINC) l C--------------------------------------

! C CALCULATE LAB 0H VALUE IN YEAR I j C--------------------------------------

LVCA,1,ISEX)aFV(MI(8,ISEX)*LFPR(8,ISEX), RIG,1) 2000 CONTINUE HETURN -

! END j CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC

! CC j SUBROUTINE LOSTLV i CC i CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC

C INCLUDE ' CONTROL.FOR' INTEGER Y,A,RLP

C-----------------------------------------

l C A IS COUNTER FOR INCortE CATEGORIES l C---==------------------------------------

1 1

i

00 3000 ISEXm1, SEX 00 3000 Aal,AC C------------------=====---------------

C DETERMINE REMAINING LIFE PERIOD C--------------------------------------

RLPsVEARS-INT (MAGE (A,ISEX))

C-----------------------------------------------

C. DETERMINE LABOR VALUE LOSS FOR EACH YEAR C==------------------------------------==-------

DO 3000 Yst,RLP C-----------------------------------------------

j C . ADD UP PRESENT VALUE OF LOST LABOR VALUE i

C-----------------------------------------------

DO 3000 InY,RLP l LVLOSS(A,Y,ISEx)=LVLOSS(A,Y,lSLX)+PV(LV(A,1,ISEXJ,R,I) m & *LPROB(A,1,ISEX)

L 3000 CONTINUE

" RETURN END CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC l CC

SUBROUTINE DEATH (OPTION)

CC CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC C

INCLUDE ' CONTROL.FOR8 INTEGER OPTION,HLP C-------------------------------------------------- -----

C DETERMINE DESIRED METHOD OF ALLOCATING FATALITIES

C--------------------------------------------------------

j IF (OPTIDH.NE.1) GOTO 5000

! C---------------------------- --------------------- ----------- -----

I C ALLOCATE FATALITIES EQUALLY To ALL YEAHS IN HHICH DEATHS OCCUR C-------------------------------------------------- ---------'- -----

! 00 2000 Kal, sex D0 2000 Jsi,01YPES

00 2000 1 1,AC

l C--------------------==----------------------------

C CHECK TO SEE IF DEATH TYPE IS ACUTE *

, C----=----------=----------------------------------

.IF (J.NE. ACUTE) GOTO 1500 Ca****************************************************

C C COMPUTE FATALITES FOR ACUTE DEATH TYPE

~

! C C*****************************************************

C l

C====----=---==---==-=====----------------------

l C -COMPUTE FIRST YEAR OF FATALITIES

} C-----===---------------------------------------

MsINT(LP(J,1))

i C--------------------------------------=-

C COMPUTE LAST YEAR OF LIFE C=-------==------------------------------

LaMtINT(HS(J))

i

! C----==-----------------------------------

t = C CALCULATE ACUTE FATALITIES C====-------------------------------------

w IF (L.EG.M) GOTO 2000 DO 1400 12mHtt,L i FPY(J,1,I2,K)sFAT(J,I,K)/(L-M) 1400 CONTINUE GOTO 2000 3 C i

1 e

1500 CONTINUE i C-------------------------------------------------- ---------- ------

C COMPUTE FATALITIES FOR NON-ACUTE DEATH TYPE C

C

! C

• CONVENT FIRST YEAR OF DEATH TO AN INTEGER At4D CALCULATE

)

C HEMAINING LIFE PERIOD C-------------------------------------------------- ---------- ------

f MaINT(LP(J,I)+MS(J))

j RLPsYEARS-INT (HAGE(I,K))

i 1

l C-------====-====---------------------------------------------- ------

C DETERMINE NUMGER OF YEARS TO CALCULATE DEATHS. !HIS C N'JHWER IS EITHER. THE REMAINING LIFE PERIOD OR PENIOD OF i C RISK FOR A CAUSE OF DEATHp WHICHEVER IS LESS. ,

C-==------------===-------------------------------- ---------- ------

IF (RLP.LE.POR(J)) LaRLP 1

IF (RLP.GT.POR(J)) LsINT(POR(J))

IF (M.GE.L) GOTO 2000 SUHs0

DO 1550 KKsH+1,L

SUMzSUMtLPROB(1,KK,K) 1550 CONTINUE l C=======------------------------------------------- ------==

C IF YEAR IS GREATER THEN, OR EQUAL TO L START

C ON NEd DEATH TYPE

C-------------------------------------------------- -------

i 1600 IF (H.GE.L) GOTU 2000 i

! C-------------------------------------------------- ----------- -------

C COMPUTE FATALITIES FOR DEATH TYPE J, AGE CATEGORY I AND

! C YEAR H+1. H+1 REPRESENTS THE FIRST YEAR OF DEATH DURING i '

C THE FIRST ITERATION. IT IS THEN INCREMENTED BY ONE YEAR ,

4 y C UNTIL THE REMAINING LIFE PERIOD HAS EXPIRED.

C------------------- ------------------------------ ---------- -------

l IF (SUH.Eu.0) HRITE(6,*)J,1,K,H+1 i

FPY(J,1,H+1,K)sFAT(J,1,K)*LPROB(1,H+1,K)/ SUN ,

!' C-----------------------------

C INCRE!1EN1 YEAR C-==========---=====------====

HsH+1

GUTO 1600 i 2000 CONTINUE ,

5000 CONTINUE

RETURN

{ END l CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC

FUNCTION I N C C A T ( A G E , 1 C , 1 I riC )

INTEGER IINC,IC

Ks0 i C==-~~---------------==-----------------------

CHECK FIRST TO SEE IF IINC EQUALS ZERO C

i C-=-------------------------------------------

l IF (IINC.NE.0) GOTO 1000 i

i

t

,,s te

, Kal

GOTO 5000 C====------------- ---=----------- '---------------- ----

C AGE CATEGORY EQUALS 18 - CHECK FIRST FOR IN UTERO .

C---------------~---------------------------------------

1000 IF (AGE.GT.0) GOTO 2000

, Kat ~

GO TO 5000 C-------------------------------------------------------------- --------

I C NOW STEP THROUGH EACH YEAR. K WILL COUNT INCOME CATEGORY OF AGE.

$ C-------------------------------------------------------------- --------

l 2000 CONTINUE I 00 3000 Int,86,IIHC-

) KsK+1 C------------------- ------------------------------ ----------- ----------

C CHECK TO.SEE IF THIS IS IN UTERO AGE CATEGORY IF IT IS i

P' C THEN A3 SIGN IT THE PROPER AGE CATEGONY AS SOUN'AS ITS AGE

! g C REACHES THE HINIMUM SOUNDRY FOR A CATEGORY. IF IT IS NOT

, C IN UTERO WAIT UNTIL AGE Is AB0VE HEDIAN A A' CATEGORY.

C-------------------------------------------------- ---- GE FOR

---

--e-------

IF (IC.EG.1.AND. AGE.LT.(!)) GOTO 5000 IF (IC.NE.1.AND. AGE.LT.(It2)) GUTO S000 3000 CONTINUE j 5000 INCCATzK

! RETURN

! END j

CC i

CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC

~

CC

! FUNCTION FV(7,R,Y) -

j REAL P,R

INTEGER Y j

FVaPa(1+R/100)**(Y-1)

RETURN I

END i CC l CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCsc CC l

e

- ee wue 3

2 put *  %#e Z 4 m it i

2 m 1

% O 1 O he m

> \

1 E l 2+

( ZK d Q =2w M 1 taJ w Z D- LS N E (J 3 LaJ & 3 24>nWC 3 taa Z > La3 Z ta. Eo-e1 EtaJ l

l 1

[

- B.36

- , _ - _ _ _ _ _ ._ __ ,___