ML20236P178

Undated AEOD/T711, Review of Data on Teletherapy Misadministrations Reported to State of Ny Subj of PNO-I-87-074A

ML20236P178
Person / Time
Issue date:	11/13/1987
From:	Pettijohn S NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD)
To:
Shared Package
ML20236P175	List: ML20236P173 ML20236P178
References
FRN-52FR36942, RULE-PR-35, TASK-AE, TASK-T711 AC65-1-095, AC65-1-95, AEOD-T711, NUDOCS 8711170184
Download: ML20236P178 (8)

Text

_ - _ _ _

i l

AE00 TECHNICAL REVIEW REPORT UNIT: N/A TR REPORT NO.: AE0D/T711 DOCKET N0.: N/A DATE:

LICENSEE: N/A EVALUATOR / CONTACT: S. Pettijohn 4

SUBJECT:

REVIEW 0F DATA ON TELETHERAPY MISADMINISTRATION REPORTED TO THE STATE OF NEW YORK THAT WERE THE SUBJECT OF PN0-I-87-74A Introduction The State of New York Bureau of Environmental Radiation Protection in early August informed NRC (RI) by telephone of the discovery that 22 or more therapy misadministration involving teletherapy treatment had occurred at several hospitals in the State. The State of New York does not have a specific requirement for hospitals to report misadministration *. The Bureau of Environmental Radiation Protection learned of the misadministration through another state agency, the Office of Health Systems Management, to which the misadministration were reported.** Most of the misadministration met the criteria set forth in 10 CFR 35.2 " Misadministration."

In follow-up reports to NRC the State of New York revealed that three hospitals were involved. The misadministration that occurred at each of the hospitals were found to have resulted from errors made by a single dosimetrist during various phases of dose calculation for patient treatment. The dosimetrist was employed by a consulting firm that provided radiological physics services including " treatment planning" to the hospitals involved. Most of the errors appeared random except that the errors involved treatment plans that were generated by a computer treatment planning system. There was no indication that the treatment planning system was not functioning properly or other indications of difficulties with the system. Instead it appears that the dosimetrist was not sufficiently familiar with the use of.the system to obtain consistent results.

Our review of these events is primarily based on infomation on the events

. compiled by the State of New York, Department of Health, Bureau of Environmental Radiation Protection. This information includes:

Treatment records for 22 patients who received misadministered doses at one of the hospitals. The records were reviewed and annotated by a consulting radiological physicist who was engaged by the hospital to review patient records for errors subsequent to the discovery of the misadministration.

• The requirement to report medical misadministration involving licensed material was made an item of compatibility for Agreement States. The effective date of the requirement was April 1,1987; however, states have three years from the effective date in which to implement the requirement.

• • The misadministration apparently were reported to this agency under a requirement to report unusual events.

8711170184 871113 PDR ORG NEXD

3.% ; ,

i -

sc x ,

l n x '

y - .;

4 LSummaries of findings on.the causes of misadministration'for the

' three hospitals' involved prepared by' the same consulting radiological- 1 physicist.-

I'nformation'obtained from discussions with Ms. Rita Aldrich,

, Chief, Radioactive _ Materials Section, Bureau of Environmental Radiation 0 Protection, New-York Department of Health.

.The New York State Department of Health Bureau of Environmental Radiation Protection (R. Aldrich). advised us that the Radiological Health Advisory q Committee to'the state has been requested to review the patient records i referenced above; and that a NRC consultant on radiological physics ,

Lhas also been retained to review the data. As of the date of this report, 1

.results from the above reviews are not'available for our consideration in H

completing this report. We expect that the findings resulting from the above i reviews will be available for consideration in an analysis to be performed by AE0D/NAS:at a later ~date (described below).

l This TR was undertaken in response to the memorandum dated August 26, 1987 from 4 V. Miller to K. Black, " Request.for AE0D Study," that requested'AE00-to review

. data on the occurrence of 22 or more teletherapy misadministration at several  !

' hospitals in the state'of New York (subject of PNO-I-87-74).

. Background ,

.f Teletherapy treatment. involves delivering a radiation dose to a tumor.by i Jdirecting a beam of ionizing radiation from an external source toward the. f tumor. region. The prescribed radiation dose is delivered in a series of q

treatments (fractions), usually one per weekday for several weeks. Teletherapy j treatment is a complex procedure that involves several professionals.and many 1

. steps. Figure 1 is a simplified' flow chart showing the key staff functions in- i radiation therapy.: The teletherapy misadministration: reviewed in this report  ;

resulted from errors-made during the treatment planning (dosimetry) phase.of therapy. i Determining-the distribution prescribed treatment by plan (dosimetry the therapist (phase) to physician) is produce' nonnallythe theradiation tumor dose task of the radiotherapy physicist or dosimetrist. Treatment plans are developed manually or by  !

computer treatment planning systems that may involve some manual calculations.

The treatment: plan defines the parameters used for the actual treatment of the

' patient by the therapy technologist in the treatment phase. Typical parameters are source to skin distance (SSD); treatment time; beam modifying device, if

< .any, to be used-(e.g., beam blocks, wedge filters, etc.); the number and angle  !

of ~ beams to use; or any special treatment method to be employed (for example, the l use of skip-scan treatment procedures). Small changes in these parameters can lead to a large' variation in the dose to the patient.

While these parameters can be computed manually or with aid of computer treatment j planning software, computer treatment planning systems vary between vendors both  !

in the input parameter requirements and the output produced. In one case the  !

actual: time that the source is to be exposed may be the output of the treatment 1 plan whereas in'another case some intermediate value may be generated that is L then used in a hand calculation to produce the actual time that the machine is  !

I to be on.

l l

+. , ,

6 L

q  :.3-- ,

r ,

LJust asitreatment planning methods' vary between institutions, the make up of

the! staffs:that perform the.above task also vary (see' Figure 1). . In'some cases, j all_-personnel are-employed by the institution; in others, such as the. events-1 that are the subject of this review, itinerant personnel may, be' involved in -

treatment planning.

' I Most . treatment regimes are modified after weekly examination 'of the patient..

One method of modifying treatment regimes as well as detecting errors is to

~

. have important calculations and measurements (e. g., dose calculations, tumor E

depth measurements,;etc.) checked by a . qualified individual who did not do the-original calculations'or. measurements. Another method.of treatment

modification and detecting errors is through periodic reviews of the patient treatment chart..

Description of the' Events Three hospitals (designated hospital A, hospital 8 and hospital C) were identified where the .dosimetrist prepared treatment plans. The current l information'available indicates that he prepared the treatment plans without assistance from or supervision of other members of the radiological physics consulting group. The consulting radiological physicist retained by the i hospitals-to review patient records subsequent to the discovery of the misadministration of misadministration reported at thethe following involved findings hospitals (see regarding)the table' 1 : occurrence forhospital.A,22of249patienttreatmentplans(about9%) reviewed

' contained errors in calculations or in dose administration (e.g., incorrect beam geometry). .In 21 of these cases, the errors led to the patient being treated with a dose that different from the prescribed dose by more than 10% (the maximum difference was 138% and 10 of the 22 patients were treated with doses that differed from the prescribed dose by more than 40%). Twenty ~one of the 22 patients received a. dose that differed from the prescribed dose by greater than 10% (criteria for reportable misadministrationunder10CFR35.2),

forhospitalB,13of30patienttreatmentplans(about43%) reviewed contained.significant errors in calculations and or in dose administration.

However, in only two cases did treatment doses resulting from i the errors meet the criteria for a misadministration. This was not a reflection of the relative seriousness of the errors but rather the treatments involved generally were for " boosting doses" where the bulk of the treatment had been given so that even a 50 percent error in the

" boosting" part of the treatment would not in all cases result in a 10 % difference between the prescribed and administered dose to the patient (criteriaforreportablemisadministrations). Also, in one of the cases the dosimetrist made a second error that reduced the severity of the first error.

L. .

(

3

, I l

, - 4-

<  ?

i l

Figure 1 Key Staff Functions in Radiation Therapy l . . a r , P, s., a wa ,a*.

iCu=.,.. . i a ....oao=* L e

i ne,;, c o.c, a i n.... . o~, al

= 1

. j iume, Lac fla.c.a e.i..a. Coa.o.,

, ca;m*';;'l,,,

,, j t.<,. vo..m. o.em.b.a _ no.....oa om . 1 seas. .e o. sa. o. hoa  : Phys.c Ch e v a c.,,c.

.....t. onato,.st oo,,,a. ,,,

, Semot.to, f.ch ,

i E ev peeat C.hbest oa 1 P*F*'

l aa i f.et I . Phys.cret 1

1 Aeco*es 1 Phys.c.I i . . .e , ~ , a.

$.m.ieuo. v Phys.cist. 1 S'.a i.io, toen . i i

C- ,.t.. O, ..ms M e..t o,a, O,acologist ,D,esimete.se oos. ,sec.s.

1 t,".a m , ',*',0;"'"' a.......a omeo,* co..m , .

Physicist '

Physc'st l Cho.c.ofRa Pi.a Phys.cist.

1 i

p.o,.i.ca oncolog st .Dosim.ter.t 7ilms to, v.e.fic.hoa . , 3.m.seio,7.ch . A.di.hoa Do.e C.rcut.hoa. . R.d.. ben O-o8 Map

, g pmys,c,3, Oncologase ; Dosimet,.se C.-peas.ier. ' 3a o, en,,,,,c.  ; m.o . a SNeidiag Stock. I po,,,ac,,g. . oncosogist ob,I .', Wowed Room Tech. Physac.sf

, , , W.che#wst Wem i  !

R.di.hoa oncologist. Phys.cist Dose f me Dosimetry C.lcu.i.hoa.

Ch.ct Ther.py R.c.ositeoamq fechaologests Loc ba.s.oa Fdas

^ 0 Porteens Ev.s.se '

-ef f vence Respea .ad ' R.d .hoa Oncosogest. Therapy T.ae< ac. 0.eia, m. s eses 1ochao6ogists i s ,0 i e C.,. i_ _ ___ ' a .. a o~

l

g. -

i ,ea.. u, e..W. i ei o- g.

i i

I .a , .

i From: " Criteria for Radiation Oncology in Multidiciplinary Cancer Management,"

Report to the Director of the National Cancer Institute of Health by the Coninittee for Radiation Oncology Studies, February 1981.

1 I

i

____._____._________j

ms mgo . .,'r

>j;)

y

, 4, ,s

- f . _ .

A[n _ n.

~

. 3p t

b4 5

,; . _5 Jc

, + .

for hospital _C,-one outlof eight patient treatment . plans- (about 12%)l reviewed  ;

was Lfound to 'contairi errors in calculations that led to' thelpatient being U treated in a way not? prescribed. For example, one' prescribed treatment!

Lplan called for treatment with two beams weighted by different factors.

' The patient was treated with both' beams weighted the same;-

although the correct prescribed dose was administered , it' wasLnot

• administered in the way indicated in the treatment ~ plan. ]

U o TABLE 1

!y 1 Hospital No. of. Records No. of Records. Percent . No of.

Reviewed: W/ Errors. Misadministration *

.A 249 22 9 21* 1

B' 30 13 43 2 H

.C 8 l' 12 0-p ..__. ...... . _________..........______..........__.......................__

L*as: defined in 10 CFR 35.2. - Another event, considered by New York to be.a misadministration, did not meet the criteria of 10CFR 35.2.

l 4

, Causes of Events iThe report of the consulting radiological physicist who reviewed patient W - treatment records at the involved hospitals characterized the type of errors i

-that led to the misadministration as follows- I i

As far as could be ascertained from.the treatment record, all errors 1 resulting from incorrect calculations were made by the dosimetrist. In no case did the technologist deliver a different time from that calculated, ]

. or. in a different manner. -

' The-dose per fraction and point, or isodose level, at which the dose was cprescribed was, in most cases, clearly written by the radiation therapist on the treatment plan, and written again by the dosimetrist at the top of the calculation sheet. Therefore the errors were not caused by miscommunication between the therapist and the dosimetrist, or misunderstanding of the prescription, i

I

> c

. y. v p - ,

h >

t ,

In a' group' of about: 7 patients treated with' " skip-scan" rotation, the dosimetrist calculated the time'for one' arc, and failed to multiply by-Ltwo..which would be-the. correct machine setting for treating two

! lateral arcs -in the. " skip-scan" mode. This resulted in' delivery of. J

% ' half the desired dose. However, because this type of treatment is. l 4 normally used as.a boost delivering only part of the patient's total J

.. ' treatment.this reduced the percent underdose. l

.. L A large proportion of:the oatient charts contained the type of errors l1 which demonstrated a misunderstanding of the meaning of beam weighting in >

the treatment planning: system being used.

l E In some calculations the"" wedge transmission factor" was wrongly included.in the timer setting equation.- In the computer treatment q E planning system being used this factor was already-included.  ;

1 In some calculations the numbers used in the timer setting equation ' '!

. were taken from the wrong part of the computer printout.- i L

]

lIn general, it appears that the errors that led to the misadministration were '

random errors made by the.dosimetrist in various phase of dose calculations.

"While'the type of errors appeared' to be random all of the errors were associated with treatment plans that used computer treatment planning systems. The plans {

that were done manually were done correctly. j 2

.No ' specific reasons were determined for the dosimetrist' making the number and type

'of errors identified except that the dosimetrist appeared to lack sufficient ]1 knowledge of the operation of the computer treatment plan being used to produce i consistent accurate results. The consulting radiological physicist who reviewed the salient treatment plans for errors following the discovery of the l ,

misadministration summarized her findings regarding the errors that were made -

by the dosimetrist as follows: "The mistakes were 'of different kinds,' and ' l were not made consistently. They seem to demonstrate a fundamental lack of Li understanding of the MDDCS computer system."

Hospital therapists appeared not to have checked on the accuracy of the treatment planning being done by this dosimetrist. In the opinion of the radiological physicist who reviewed the patient treatment plans for errors, independent checks of the dose calculations would likely have prevented most of the errors.

While the errors that led to the misadministration generally can be ascribed i to'a dosimetrist making random errors involving dose calibrations, the facts -

of the misadministration highlight several general issues that we believe lshould be considered for further review. These issues are: l Training requirements for personnel involved in treatment planning, dose l ll calculations, dose delivery, etc., in radiotherapy facilities. Lack j L

of appropriate training appeared to be the root cause of the misadmini- )

strations that were the subject of our review. j l

i I

l

n

...m3= , ,

1

, 1 ,

7 ,

f .d 'kh L ,

L - 7>-

t t ..

Definition of the relationship between radiotherapy' consultants, licensees

~

Land -NRC in regard to training and other regulatory, requirements- . The.

" desimetrist who.made the errors that led to the. misadministration, a; member of a' consulting group retained by the hospitals, would not be J prohibited by any NRC regulations from continuing to prepare treatment iplans,'although the nature of the errors indicated that the individual

' appeared to lack 'a' fundamental' understanding of the work that' he was performing.

2 -: Definitive ~ requirements for periodic patient chart. reviews to detect errors:

in the patient treatment- )lan. Phile the_ proposed rule on basic quality,

. assurance addresses a wee cly. accuracy check of daily arithmetic calculations,.

'it ~ appears that more extensive checks such as a: check of correction factors and patient setup, the_ use of beam modifying devices, etc. may airo be called .

.for .to assure a highl probability of detecting errors.before they lead to .

misadministration. For the misadministration that were the subject of our review (23 events) none were detected by radiotherapy personnel. reviewing patient treatment progress.

-- Annual audit of patient treatment records.- _0ne significant fact concerning the.22 teletherapy misadministration is that none of the misadministration were detected by hospital radiotherapy' personnel, or. personnel _ of the consulting physics group during the normal implementation of' treatment

> protocols. fThis' failure to detect errors continued over a five year period in which the average annual error rate (misadministration per patients - 1 treated);for one hospital was about 8%. The annual error rate per patient

~

-l

.given'in.AE0D. report AE0D/N505, Therapy _ Misadministration Reported to the -

'NRC Pursuant:to 10 CFR 35.'42," is .02% a factor of1400 less. An annual audit that' detected errors the~first year could have prevented the  :

-misadministration that occurred in the'later years, i 1

- Standardization of' computer treatment plans. The consulting radiological '

physicist who reviewed the salient treatment plans for errors following 1 the discovery of the' misadministration, summarized her findings regarding  !

'the errors that were made;by the dosimeters as follows: "The mistakes were of different kinds, and'were not made' consistently. They seem to demonstrate a fundamental lack of ' understanding' of the MDDCS computer system." The  ;

individual whose errors led to the misadministration may have been using i several computer treatment plans, and may have confused some of the l parameters used in the different plans. Standardization of computer treatment plans, or user-friendly interfaces, could make working with different computer treatment plans less likely to lead to errors.

- Definition of a misadministration. In six out of the 22 misadministration at _ hospital A (21 of the 22 met the criteria of a misadministration given in 10 CFR -35.2), in addition to the administered dose differing from the prescribed dose, the distribution of the dose to the tumor was found to be different than what was prescribed. While most (5 out of 6) of the events met the criteria for misadministration reporting based on the prescribed and administered dose differing by greater than 10%, the primary error for one of the events was the maldistribution of the dose. Maldistribution of the dose may be as significant as a 10% difference in the prescribed and administered dose in regard to the effect on the patient.

~ ~~

C., .1 ., .!

3  ;

r b  !

Linear Accelerators. Several of the misadministration that occurred 1 involved radiation therapy protocols using linear accelerators. Census data on demographics of radiation therapy in areas such as staffing, equipment, and new patient load found in the Patterns of Care Studies conducted by the American College of Radiology show that the number of teletherapy machines in use declined from 970 in 1975 to '744 in 1983 (equivalent to about a 3% drop per year) whereas the number of linear accelerators / betatrons increased from 407 to 1018 (equivalent to about a- i 19% increase per year) during the same period. The data show a definite trend toward the use of linear accelerators. Consequently, to be effective in. preventing all misadministration, quality assurance procedure requirements ' u for radiotherapy facilities should be applied to therapy protocols using linear accelerators as well as to therapy protocols using cobalt 60 teletherapy machines. Although NRC has no authority to regulate linear accelerators, 1 should we, or can we work to have some group or groups tedertake. the regulation of accelerators in this regard. i Conclusions  !

Based on our review of the data currently available we found that the misadmini- ,

strations essentially resulted from random errors made by a consulting dosimetrist  !

in' calculating patient doses. All of the errors were made on patient trsatment i plans involving data generated by computer (computer treatment plans). Further, i we found that the type of errors that were made (e.g., errors in dose calculation, j use' of wrong beam modifying factors) likely could have been detected if the '1 treatment plans had been checked for accuracy by another individual. We believe that, in most cases, such checks could have detected errors sufficiently early to have. prevented the misadministration. A review of the data also revealed that many of the over- or under-treatments likely could have been detected by i prudent patient chart reviews. These also could have prevented the misadmini- l strations if the over- or under-treatment had been identified sufficiently early i to adiust'the treatment plan, i In relating.the causes of the 22 misadministration events to the proposed rule

" Basic Quality Assurance in Radiation Therapy," we found that the requirement for a check of dose calculations by an individual who did not perform the ,

calculations and the requirement for weekly accuracy checks of daily arithmetic j calculations contained in the proposed rule appear to address the primary causes )

identified for the misadministration. Therefore, we do not believe that any ,

changes to the proposed rule are occasioned by our findings regarding the i causes of the misadministration.

In regard to the " general issues" discussed above, we will develop a technical )j review (TR) of these general issues. Our schedule for completion of this additional TR will allow this information to be considered in the development )

of the proposed rule for comprehensive quality assurance. ,

1 l

1 3