ML20247Q977
| ML20247Q977 | |
| Person / Time | |
|---|---|
| Site: | Fermi  |
| Issue date: | 06/30/1986 |
| From: | Lewis P, Swaim D Battelle Memorial Institute, PACIFIC NORTHWEST NATION |
| To: | |
| Shared Package | |
| ML20247Q960 | List: |
| References | |
| PNL-SA-14093, NUDOCS 8908070299 | |
| Download: ML20247Q977 (14) | |
Text
[
@((6fENll f_
^
[,..,
., j f, j
~~
.-a s.
It\\
us.>s.La ssH on
~
4% s report o.s sweperr1 as an account of work spaneered by an agency as the L:cated States.
I
/
',Is, f
[,"#
,,, d. " ~7 overr.snent.
'G
~
Neither the United States Goveranient mer any agency thereof, nor any of their
'e esapt9yees. makes say' warroaty, espress or implied, or manusus any legal Isabdin or respunne.
- bilsey for the accuracy, ownpletemens or eerfulmens of any informistaun, apparetus primivet, w process diecimmed. er represents that ise one would not safrispe priestely v ned rights. Refee-
,) C D
- 8 " **g g* *
- e o "'"
ence heresa so any specific commercial product, prwens. m serviw by trade man.c. trademark, i
D' menefacturer. eer asherwine dnes not necessarily ainstitute w imply sta endurwmeat. rnoni, mr.adetaan. or Tesaring by the United States Government er any agency thereal. The siens and opensans of surhars empresand hereia de ses escusurely sta.e or reflect slune uf the S,
" ', g United States Guiernment er any apeary thernet.
~
PNL-SA *14093 EVALUATION OF A 12-HOUR / DAY SHIFT SCHEDULE DE87 000821 P. M. Lewis (a)
D. J. Swaim 7
June 1986 Presented at the 1986 Annual Meeting of the Human Factors Society Dayton, Ohio September 29 - October 3, 1986 1
Work supported by the U.S. Department of Energy. under Contract DE-AC06-76RLO 1830 Pacific Northwest Laboratory Richland, Wa >iinpr 99352
("IWestinghouse
...iford Company Richland, Washington d
ook$$o$$43 NN W MS NM 4 MM PNU
.).h
1 I
s' i
_ EVALUATION OF A 12-HOUR / DAY SHIFT SCHEDULE l
l in April 1985, the operating crews at the Fast Flux Test Facility near Richtand, Washington, changed their rotating shift schedule from an 8-to a 12-hour'/ day work schedule.
The primary purpose of the change was to reduce the attrition of operators by increasing their job satisfaction.
Eighty-four per-cent of the operators favored the change. A program was established to evalu-ate the ef fects on plant performance, operator alertness, attrition, sleep, health, job satisf action, and off-tne-job satisf action.
Preliminary results from that evaluation program indicate that the 12-hour shift schedule is a reasonable alternative to an 8-hour schedule at this facility.
INTR 000Citot The Fast Flux Test Facility (FFTF) is a nuclear reactor owned by the U.S.
Department of Energy and operated by the Westinghouse Ha.1 ford Company.
Since commencement of ooerations in 1977, the annual attrition rate for operators generally averaged from 151 to 20%. Although this attrition rate is not high by industry standards, such attrition does increase recruiting and training costs, and reduces the average ~1evel of experience and qualifications of the Crews.
Although shif t work is unavoidable for the operating crews, it ha? been a major cause of attrition. FFTF's original shift schedule was an 8-hour / day schedule, with weekly rotation. After seminars and discussions on alternative shift schedules, Westinghouse management and operators decided upon a short list of reasonable siternative schedules.
In February 1985, 84% of the opera-tors expressed their preference for a 12-hour / day schedule that is basically A consecutive 12-hour days (o. nights) of work followed by 3 days off, with a 7-day long change every 5 weeks. Based in part on this statement of prefer-i ences Westinghouse management adopted the 12-hour schedule for.a one-year trial period, beginning in April,1985. The effects of the change were monitored and evalufted by several organizations, Westinghouse conducted its own internal L
monitoring and evaluation program. The U.S. Department of Enerd funded (00E Contract DE-AC06-76RLD 1930) a program conducted by Westinghouse and pacific l
1
'.l
~;
k Northwest Laboratory; preliminary results from that program are reported in j
9 this paper.
In addition, the National Institute for Occupational Safety and l
Health, in cooperation with the former organizations, conducted computerized tests of alertness and issued daily questionnaires; preliminary results from that program are abstracted in this paper and sumarized in another paper in this volume (Rosa, Colligan, and Lewis).
Toward the end of the trial period, in April 1986, 84% of the operators expressed their preference to continue with the 12-hour schedule. Based in part on-this preference, and based on the results to-date of the evaluation programs, Westinghouse management decided to adopt the 12-hour' schedule.
EFFECT ON PLANT PERFORMANCE The potential effect of the schedule'caange on plant performance was evaluated primarily with documented reports of unusual.or off-normal events, called Event Fact Sheets, and with dats feca logs of plant parameters, called Technical Specification Compliance Logs.
pentFactSheets
~
Management routinely writes reports on any unusual events, even those of a minor nature.
The summary raport of such an event, called an Event Fact Sheet, also records the date and hour of the event. For this evaluation project.
Westinghouse personnel reviewed each report in the evaluation period in order to:
- 1) determine whether the event was caused in part by operator error. 2) determine whether the event might have been mitigated by improved Shift turn-over communication, and 3) determine the severity or safety significance of the event.
Thirty-two events with operator error occurred during 485 days on the 8 hour9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> shift (an average of 24 events per yeer); *1 events with operator error occurred during 365 days on the 12-hour shif.t.{an average of 21 per year). The rate for the 12-hour shift is slightly lower, but the difference is not statis-tically signif1 cant.
The probability of an event occurring may be affected by the work load and work requirements. For example, events may occur more or less frequently in 2
i each of the four dif ferent modes--shutdown, refueling, startup, and opera-tions. Therefore, the number of events per unit of time the p1:!nt was in each of the four modes was compared for the two shift schedules.
Again the differences were net statistically significant.
Whenever an unusual event is caused by operator error, the operator's manager is asked whetner fatigue might have contributed to the error.
In no case has fatigue been reported to be a contributory factor.
Additional data are now being collected that will allow a more refined analysis of the Event Fact Sheets. The severity of each event is being rated independently by three evaluators, one each from the FFTF Operations, Safety, and Operations Analysis departments. Four aspects of severity are being rated:
- 1) economic costs (inc!3 ding schedule delays); 2) personal it. jury and radiation exposure; 3) potenttai safety effects (if something else had failed concurrently); and 4) whether, and the degree to which procedures were viola-ted. The evaluation of severity is igortant, because, a large number of Event Fact Sheets are written on events of little consequence.
Technical Specification Compliance Logs Operators record in log books about 2,000 plant parameters (e.g., tempera--
tures and pressures) every day.
If an operator notices that he has made'an error, he must cross out the error and write the correct entry. "ne logs are checked by the shtft supervisor and some are later checked against coguterized records. The Technical.Spect fication Compliance Logs are an important measure of plant performance: a missed surveillance could result in a technical spect-fication violation that could cause a plant shutdown.
The Technical Specification Cog 11ance Log data were reviewed for the period 2/1/1985 to a/28/85 for the 4-hour shift, and 5/31/85 to 8/25/8.5 for the 12-hour shift.
Forty three parameters that require hourly log entries were examined, which means that, approximately 1,000 log entries were examined for each day, and approximately 180,000 entries were examined over the entire 6-month evaluation period. The date, hour, crew and operator for each corrected entry were recorded.
3
l' l
s The major conclusion from this analysis was that the proportion of cor-rected entries was actually lower on the 12-hour shift than on the 8-hour shift (0.0030 versus. 0.0036), although the difference was not statistically signifi-cant (p =.30).
The statistical test was a T-test on the number of corrected entries each hour. The average error rate was 0.175/ hour on the 8-hour shift, and 0.153/ hour on the 12-hour snif t.
Tr.e Technical Specifiestion Compliance Log data were examined to deter-mined whetner errors increased as the number of consecutive days of work increased.
No such trend occurred. Also, the, nighttime error rate was not lower than the daytine erro.' rate on either the 8-or 12-hour shift.
Questionnaires In Decenber,1985, operators were asked in a questionnaire how eff'ectively their crews worked on the 12-hour schedule.
Of the respondents who had worked both. schedules, 52% snid their crew worked nore ef fectively on the 12-hour senedule. Only 12% saio their crew worked less effectively on the 12-hour shift. Questionnaires were also given to the fif teen shift managers.
Literally all of the on-shift managers reported that their crews were generally more productive.
When asked why operations is more efficient on the 12-hour shif t, many operators said that when they are only away from the plant for 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />, and when very little of that time off is spent on other distractions, it is easy to
~
get back into the pattern of the work.
Another major reason for the perceived increase in work accomplishment is that the 12-hour schedule allows operators to supervise the crafts personnel
.more effectively. On the 8-hour shift, the operators and crafts arrived at the same time, and the craf ts often simply waited while the operators rere making preparations.
On the 12-hour shift, operators arrive at work before the crafts persennel, and so the operators are prepared to assign tasks to the: crafts per-sonnel as soon as they arrive at work. Similarly, at the end of the work day on the 8-hour shif t, operators often,had the craf ts personnel stop york early, 4
i i
so that the operators themselves could prepare for shif t turnover. On the 12 hour1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> shif t, however, operators can have the,craf ts personnel continue to work to the end of their work day.
In other respects, however: crews seemed-to be less productive. For example, operators are encouraged to study for qualificatir,ns ex.minations dur-ing slack periods. Many operators, however, found it more difficult to main-tain study efficiency for the longer periods availabir on the 12-hour day.
1 Effect on Performance Summarized Performance is the most important criterion by which to evaluate the effects of the 12 hour1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> shift. Both " negative" performance (i.e., error rates) and " positive" performance (i.e., work accompitshment) were evaluated.
Event Fact Sheets and the Technical Specification Compliance Logs both measured error rates. Highly quantified data from both of these measures indicate that there is no statistically significant of fference in error rates between the 8-and 12-hour shifts. On the other hend, most operators and managers believe that work performance generally improved on the 12-hour shift.
The following sections are analyses of factors related to shift scheduling that may affect plant performance:
alertness, shift turnover, and operator qualifications, attrition, recruiting, sleep, and health.
EFFECT ON ALERTNESS Particular attention was paid to alertness, because operators may become less alert after working 4 consecutive 12-hour days with less time off for rest and sleep. On the other han.1, with more days off, operators may be able to recuperate more fully.
The Technical Specification Compliance Logs are both a measure of perfor mance and a measure of alertness.
They can be considered a measure of alert-ness because filling out the 1.ogs is routine and repetitious. As w6s mentioned aboys, this error rate actually declined on the 12 hour1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> shift, which would l
indicate that alertness increased, not decreased, on the 12-hour shif t, although the amount of the increase was not statistically significant.
5 i
L__-.__________.
.v Computerized Test of Alertness Computerized tests of alertness also provided highly quantified data on-alertness.
In contrast to the Technical Specification Compliance Logs, how-ever, the computerized tests indicated somewhat decreased alertness on the 12-hour shift.
Four computers were distributed among the control rooms and other rooms of the FFTF plant during the last 5 weeks of the 8-hour shift (April 1985), during the first 5 weeks of the 12-hour shif t (May), and during 8 weeks in November and December,1985.
Operators were scheduled to take the tests once or twice during their work day. Taking the tests was voluntary; about 34 operators participated consistently during both shift schedules.
The computerized tests of alertness included tests of di;f t addition, grammatical reasoning, auditory reaction time, anti hand steadiness.
The test scores were slightly "better" (i.e., in the direction of greater alertness: eore responses and fewer errors) on the 12-hour shift.
- However, much of the improvement can be attributed to an increase in practice. When the effect of practice is accounted for statistically, the resulting analysis indi-cates that the operators were slightly less alert on the 12-hour shift than on the 8-hour shift. - For example, after accounting statistically for the practice effect, operators performed the gramatical reasoning test slower (about 2.8 versus 2.6 seconds),' although their accuracy remained unchanged. Also, opera-l tors' reaction time to an audio signal that they heard while they were simul-taneously conducting the gramatical reasoning test increased on the 12-hour shift from about 130 to 170 milliseconds.
~
Details of the computerized tests of alertness are presented in another paper in this volume (Rosa, Colligan, and Lewis).
Questionnaires j
Several questionnaires asked,about al&rtoess. Here, too, the data from different sources are not entirely consistent. Most, but not all, of the questionnaires indicated that alArtness was generally somewhat lower on the 12-hour shif t.
These data also indicate, however, that f atigue does not accumulate significantly over 3 or 4 consecutive 12-hour work days (or nights).
t 6
,[*h i
j x,'
j Operators were asked to fill out questionnaires at the beginning of each shift; these' questionnaires asked for the operators' subjective rating of I
alertness.
These subjective ratings did not indicate a significant difference in fatigue between the 8. and 12-hour shifts.
By contrast,'the 15 on-shift manag' rs reported that their personnel were e
clearly more fatigued at the end of the-12-hour workday than they were at the end of the 8-hour workday.
The dif ference between the managers' observations and the operators' daily subjective reports may be that the operators' reports came at the beginning, not the end, of the work day.
The operators' reports, however, indicate that there was no significant accumulation of fatigde over consecutive 12-hour work days.
Operators were also given comprehensive questionnaires for the 3-hour shif t in May, and for the 12-hour shif t in December,1985.
On both of.these questionnaires, operators were asked, "About how often do you feel tired or sleepy. aft &. completing ycur work?" The responses indicate increased f atigue on the 12 hour1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> shift:
the percentage of respondents that answered "4-7 times a week" rose from 13% to 251, and the percentage answering "1-3 times a week" rose from 42% to 50%. The difference is statistically significant (p =.02).
. In the December questionnaire, operators were asked to compare the " worst" (i.e., den they were the most tired) days of the 8-hour shift with the worst days of the 12-hour shift. Forty-five percent of the respondents who had worked both shif ts said that they felt more tired on the worst day of the 12-hour shift. Most operators (705) reported that the worst work days of the 12-hour shift were not the last but rather the first two night shifts. Asked to explain this result, operators said, in effect, that circadian desynchronosis caused more sleepiness than fatigue accumulated over several days. This explanation is consistert with the fact that most operators (58%) also said that the worst days of the 8-hour shift were also the first two night shifts.
Presumably the f;1rst 12-hour night shif ts are worse than the first two 8-hour night shif ts because they are 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> longer. But the. fact that the succeeding 12-hour night s_h_tf ts are generally perceived as being better suggests that the primary determinant of fatigue on night shifts is circadian rhythms, not tie number of hours of work.
7
s.
I;,
3
~
Effect on Alertness Summarized The. average level of alertness probably declined somewhat on the 12-hour shift, although not all of the indicators of alertness point to this conclu -
sion.
If.the average level of alertness did decline, it apparently had little 4
or no adverse affect on plant performance.
EFFECT ON SHIFT TURNOVER COMMUNICATION Because the on-coming crew must learn the details of the plant status before asstaning control of the plant, shift turnover communication is impor-tant.
In the December comprehensive questionnaire, operators were asked to coripare snif t turnover communication between the two shift schedules.
Eignty percent of the respondents who nad worked botn schedules said tnat turnover communication was easier af ter returning from a 12-hour absence on the 12-hour snitt than af ter returning from a 16-hour absence on the S-hour shift..Much of the igrovement was due to the fact that on the 12-hour shift, the on-coming crew receives the plant from the same crew it gave the plant to 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> earlier.
The on-coming crew knows what the plant status was when it left, so if it has any questions concerning what happened at the plant'in the mean time, the of f-going crew ought to have the answers.
During shift turnovers on the 8-hour shift, there is always a third crew, not at the plant at the time, that might have the answers, or that can at least get stuck with responsibility for the answers.
On the other hand, 42% of the respondents that had worked both schedules said turnover communications were more difficult after returning from the 7-day long change on the 12-hour schedule, compared to returning from the 4-day long change on the S-hour schedule. One operator wrote that after a 7-day long change "it takes 2 or 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> to get back in the groove on what has happened In order to ensure proper shift turnovers, Westinghouse instituted new turnover procedures whereby the Shift Operations Mahager from the eff-going crew briefs the crew that is returning from long change.
8 e
__._____.__n___---.___
.f During 16 months on the 3-hour shif t and 12 months on the 12-hour shif t there was a total of only three unusual events (documented in an Event Fact Sheet) that were judged to have been caused in part by shif t turnover consnuni-cation problems.
EFFECT Oft 00ALIFICAT10fl5 From management's point of view, one of the objectives of the change to a 12-hour snift was to increase gradually the average level of qualifications, primarily by reducing attrition. Westinghouse has a highly quantified measure of qualifications, which is derived from the frequent examinations that opera-tors take in orcer to qualify for various watch stations. As of May, 1986, ten months after the schedule change, the average level of qualifications was sligntly higher than at the time of the schedule change.
EFFECT ON ATTRITION AND RECRUITING The improvement in qualifications was expected to result paimarily from reduced attrition.
The reduction in attrition was expected to re wit from increased satisf action with the shift schedule. Because most aerators are.
more satisfied with tne 12-hour schedule, good reasons exist for believing that the change will tend to reduce the attrition rate.
In the short run, however, the attrition rate has increased, for reasons entirely unrfiated a the 12 hour1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> shift: by coincidence, at the time FFTF changed its shift schedule, FCF was also in the middle of a planned ramp down in total staff. Only very selective outside hiring was permitted throughout the company. This increased the opportunity for internal transfers within the company as other departments could not fill their positions. Because the Operations personnel are highly regarded, they became prime candidates for transfers or promotions.
Although attrition during the trial period was 18%, 70% of the attrition was due to' pro-motions and transfers.
In the.y. ear prior to the trial period attrition had been 12%, with promotions and transfers responsible for only 40% of the attrition.
If promotions and transfers are deleted from the attrition figures, the attrition rate during the 12-hour shif t would be about 5.51, compared to about 7.1% for the 8-hour snift.
9
i 1
Therefore, although the attrition rate rose af ter the schedule change, the increase in' the attrition rate cannot be attributed to the schedule change.
More likely, the schedule will tend to decrease attrition in the long run, due i
to improved job satisfaction and of f-the-job satisfaction.
People who left shift work at FFTF were asked whether the shift schedule af fected their decision to leave.
The interviews were condected in confidence by an outside researcher who is independent of Westinghouse (Daul Lewis).
These interviews indicate that the dif ference between the 8-and 12-hour scredules was a minor f actor or no factor at all for most of the people who left shift work.
Some operator; did say, however, that although they were looking for a new job Defore the schedule change, they began looking harder after the schedule change because they preferred the 8-hour schedule.
Two people wno were single parents left snift work because it was very difficult
~
for them to find baby sitters or day care for their children for the unusual hours reaufred by the 12-bour schedule.
The adoption of the 12-hour shif t has had a negligible ef fect on the ability to recruit qualified new employees for FFTF Operations.
After the
_ adoption of the 12-hour shif t, nine new operators were hired. All were quality candidates and of essentially the same demographic mix a.; in previous years.
All nine constoered the 12-hour shift a positive reason for accepting the job. At the end of the trial period, 8 of the 9 new employees voted to retain the 12-hour shift permanently.
(The 9th employee had alrecdy lef t the company for reasons unrelated to shift work; in fact, he was a strong advocate of the 12-hour shift.)
EFFECT ON SLEEP That rotating shif t work seriously disrupts sleep patterns is well known. FFTF's former and current shif t schedules are both rotating shift senedules, andiboth disrupt circadian rhythms and sleep patterns. Althougn sleep patterns are difrerent on the two schedules, the evidence does not con.
clusively indicate that sleep patterns are on the average significantly better or worse on the 12-hour shif t.
However individual differences are signifi-cent:
in the comprehensive questionnaire, operators were asked to complete tne 10 i
y.
e following sentente:
" Compared to my sleep on FFTF's 8-hour schedule, my sleep on the 12-hour senedule is generally..." " Nineteen percent of those who had worced both schedules said their sleep was "better" or "much oetter;" 20s said
)
i it was " worse" or "much worse."
Data from daily questionnaires also indicate that the 12-hour shift pre-j sents a mixture of advantages and disadvantages for sleep patterns (See Rosa, Colligan, and Lewis in this volume).
1 EFFECT ON HEALTH The effects of tae schedule change on operators' physical and psycho-logical healtn was monitored by comprehensive questionnaires issued for the el-and 12-hour shif ts, by observation of '99 operators' medical recoros, and by interviews with some of tne operators.
The questionnaire on physical health was a stigntly modified version of The General Heal *h and Adjustment Question-naire used by the 1ational Institute for, Occupational Health and Safety ( fasto, Colligan, Skjei, and Polly 1978).
The change to the 12-hour shift appears to have had no significant impact either negative or positive on the physical and psychological healtn of the FFTF operatort. FFTr operators are generally a young (average age = 32) and healthy.
The vast majority of them rated their health as good or excellent before (99%) and after (971) the schedule change.
SUMMARY
Although subjective and objective measures of alertness indicate that operators are probably slightly less alert at the end of a 12-hour day, the available evidence indicates that plant performance was not adversely affec-ted. Analysis of the two highly quantified measures of plant performance, Event Fact Sheets and Technical Specification Compliance Logs, shows no statis-tical difference in pe,rformance between the 8-and 12 dour shifts. Most opera-tors ano all 15 of the on-shift managers judged that crews were generally more effective on the 12-nour shift. Furthermore, the time in the shif t schedule 11
~
1 e
when operators feel the most tired is not the last night shift, but the first night shift, a perceptien that is probably related more to circadian rhythms than to the number of hours of work.
Eighty-four percent of the operators and 100% of the new hires prefer the 12 'to the 8-hour schedule. The expectation is that this preference will reduce attrition and therefore increase the average level of experience and qualifications at the plant, which may eventually result in a measurable improvement in plant performance.
M*
D g
9 g
M ee N'
12 l
1 1
i 19' re-i g
REFERENCE Tasto, D. L., M. J. Colligan, E. W. Skjel, and S. J. Polly.
1978. Health 3
Consequences of Shift Work.. DHEW (NIOSH) Publication No.78-154.
National Institute for occupational Safety and Health, Cincinnati, Ohio.
.~
~
s
- - * - ~ '
~g g
(
~
L.-
-L
~
i l
13 l
3 i
I