The nuclear industry family study: methods and description of study population

London School of Hygiene and Tropical Medicine, University of London, London WC1E 7HT

Noreen Maconochie,
lecturer

Pat Doyle,
senior lecturer

Graham Davies,
research fellow

Peter Smith,
professor

Leukaemia Research Fund, Institute of Epidemiology, University of Leeds, LS2 9LN

Eve Roman,
reader

Imperial Cancer Research Fund, University of Oxford, Radcliffe Infirmary, Oxford OX2 6HE

Valerie Beral,
professor

Correspondence to: Dr N Maconochie n.maconochie{at}lshtm.ac.uk

Abstract

Objective To construct a comprehensive database linking nuclear industry employment and dosimetry records to information on employees’ reproductive health and the health of their children.

Design Retrospective cohort study of reproductive outcomes reported by nuclear industry workers of both sexes. Postal questionnaires were used to collect information on pregnancies, children, and periods of infertility. Information on employment and monitoring for ionising radiation was supplied by the employing nuclear authority and linked to pregnancies by using unique personal identification numbers.

Setting Nuclear establishments in Great Britain operated by the Atomic Energy Authority, the Atomic Weapons Establishment, and British Nuclear Fuels Limited.

Subjects 46 396 current and former employees.

Results When undelivered mail was excluded, the response rate was 82% for male workers and 88% for female workers. 67% of men and 58% of women reported at least one pregnancy or attempt for a pregnancy. A total of 43 710 pregnancies (39 557 live births) were reported by male workers, and 9962 pregnancies (8883 live births) by female workers. 41% of pregnancies reported by men and 10% of pregnancies reported by women were conceived after the parent had been monitored for exposure to ionising radiation. Among these, median lifetime cumulative whole body (occupational) dose before the estimated date of conception was 12.6 mSv in men and 4.2 mSv in women.

Conclusion This is the first UK study to link detailed reproductive history data to occupational information held by employers. The design and conduct of the nuclear industry family study resulted in high quality data on a representative population of the Atomic Energy Authority, Atomic Weapons Establishment, and British Nuclear Fuels workforces. The response to the survey was extremely good, and a unique relational database has been created. This has enabled infertility, pregnancy, and child health outcomes to be examined with respect to the employment and dosimetry characteristics of parents. The methods used in this study have been adapted for use in other important occupational investigations such as the study of reproductive outcome of veterans of the Gulf war.

Introduction

Generally, when an occupational hazard is suspected, the health of the workers in the industries concerned is studied directly. Within the nuclear industry, several cohort studies have investigated the relation between cancer and exposure to ionising radiation at work. ^{(14) (15) (16) (17) (18) (19) (20) (21)} The nuclear industry family study takes this one step further, linking occupational exposure of workers to cancer in their children. Further, since parental occupational exposures could, in theory, result in anything from infertility through to cancer in grandchildren and beyond, ⁽²²⁾ information on a range of other adverse reproductive outcomes was collected.

Methods for doing such studies are not well developed. The nuclear industry family study is the only study to approach nuclear industry employees directly and ask them about the health of their children. The study design that was used developed from a survey of reproductive health in British radiographers, ⁽²³⁾ which showed that a postal method could be used to obtain full reproductive histories from a large study population. The design of the nuclear industry family study, together with response rates and description of the study population, is described and discussed in this report.

Methods

Reproductive outcome and child health

Following advance publicity at individual sites, in industry newsletters, and in extensive meetings with workers, management, and trade union representatives, each employee was sent a package containing a covering letter from management and trade union representatives, an information leaflet describing the study, a questionnaire (numbered with the subject’s unique personal identifier) and a reply paid envelope addressed to the researchers. Packages to current workers were delivered in batches to individual sites, and those to past workers were sent to the home addresses supplied by the pensions administration office. Reminders were sent at intervals of three to four weeks, study subjects being requested at this stage to return the questionnaire blank if they did not wish to participate. A freephone helpline giving direct access to the researchers was available to answer any queries about the study, and this was well used.

The questionnaire requested details of all liveborn children, including name, sex, date and place of birth, birth weight, and any serious medical conditions ever experienced. Also requested were details of any adverse pregnancy outcomes (miscarriage, stillbirth, ectopic pregnancy, hydatidiform mole, missed abortion) and of terminations of pregnancy for medical reasons, including date of pregnancy end, gestation, whether any abnormalities were detected in the fetus, and sex of fetus if known. Questions relating to infertility were also included.

Parents who reported outcomes of interest were sent a second questionnaire in order to obtain further information about the condition and to request permission to consult relevant medical records. With appropriate maternal consent, general practitioners’ or obstetric notes, or both, were consulted to verify the details of all reported deaths of fetuses of 16 weeks and over where abnormalities in the fetus were detected, and for all reported terminations for medical reasons. With written permission from parents (or the children themselves if aged 18 or over), clinical notes were consulted to check the date and nature of the diagnosis where major congenital malformations, cancer, or death were reported in liveborn children. An additional check on the validity of reports of cancer and death was made by requesting a copy of the cancer registration or death certificate, or both, from the NHS central register. At the time of survey, however, cancer registration details were available only for the period 1971-93. All adverse health outcomes were double coded to the 10th revision of the international classification of diseases ⁽²⁶⁾ by a research nurse and an experienced epidemiologist, both of whom were blind to the exposure status of the worker. For comparison with national mortality and cancer incidence rates, reported cancers and deaths among children were "bridge coded" to the 9th revision of the international classification of diseases, ⁽²⁷⁾ using the coding rules of the revision in effect at the time of diagnosis or death.

An estimated date of conception was calculated for each reported pregnancy. This was taken to be the date of pregnancy end, minus gestational age, plus 14 days. Gestation was unknown for the majority (99%) of liveborn children, and was estimated as 40 weeks (36 weeks or 28 weeks if prematurity was indicated). Gestational age was also estimated for 69% of reported terminations, using the median reported gestation for all other terminations (16 weeks). For the majority (93%) of fetal deaths and other incomplete pregnancy outcomes gestational age was known, and for the remaining 7% it was estimated using the median reported gestation for all other reported pregnancies of the same type.

Employment and monitoring data

Information about employment in the nuclear industry and potential exposure to ionising radiation at work was obtained from the personnel, health physics, and epidemiological databases maintained by the three authorities. Data were supplied for all those responding who had ever had, or who had ever attempted to have, children, and were linked by computer to the cleaned, coded, and validated questionnaire based data by using the subjects’ unique personal identifiers. Thus the researchers were blind to the occupational exposure status of the workers when collecting data on reproduction and child health outcomes. All three authorities provided dates of first and last employment within the industry, site of current (or most recent) employment, and classification as an industrial or non-industrial worker. In addition, the Atomic Energy Authority and British Nuclear Fuels provided an employment history for each worker, giving details of all employment periods (by site) at any of the three authorities. Care was taken to avoid duplication and to ensure the consistency of employment histories when employees had worked at more than one site or authority. Where individuals had worked for more than one authority, all records were collated and supplied by the authority of most recent employment. For non-responders, information on date of birth was supplied by the Atomic Energy Authority and the Atomic Weapons Establishment; British Nuclear Fuels declined to supply data on age for any individual who had not returned a completed questionnaire, for reasons of confidentiality.

All employees who are potentially at risk of exposure to external ionising radiation are required to wear personal dosimeters (film badges or thermoluminescent devices) to record the amount of exposure, and for regulatory and managerial purposes nuclear establishments maintain records of all such monitoring. The three authorities supplied radiation records for all workers who responded and either had children or had attempted to have children. Workers who had at least one such radiation record are referred to in this study as "monitored workers." For each calendar year that they had worn a film badge, the industry supplied a record of the worker’s total whole body (occupational) dose accumulated over the year at each establishment, expressed in millisieverts (mSv). (A worker could have more than one record in a year if they had worked at more than one establishment.) Actual dates of first and last monitoring were provided in the personnel record. Natural background radiation was excluded and appropriate adjustments were made for factors such as the minimum detection dose value ("threshold dose") of the dosimeter and for missing film badges. ^{(20) (28)(29)}Preconceptional doses were calculated by pro rata scaling of the annual doses.

In addition to being monitored for external ionising radiation, employees may also have been investigated for possible internal contamination if they were considered to be potentially exposed through working with radioactive substances. These investigations consist of bioassays or whole body monitoring (available since the 1960s). The type of information provided on this "internal monitoring" varied among authorities, but all employers provided the calendar year in which an employee was monitored for the first time for potential internal contamination by any radionuclide. All workers for whom such a date existed are referred to in this study as internally monitored workers.

Extensive checking of the computerised radiation and employment records supplied was carried out by the researchers, checking against the primary (industry) data source where necessary. It was found, however, that certain inconsistencies could not be uniformly resolved, and such data were tagged as "uncertain": for example, if the date of first employment in the personnel (summary) data predated the information from the employment history data, or if a radiation record of monitoring at one of the three authorities in the study predated the date of first employment in the personnel data, this intervening period was classified as a period of "uncertain" employment. For monitoring, "uncertain" status was assigned where data from the personnel records predated the year of first monitoring in the annual dosimetry records.

Further inconsistencies existed within the dosimetry data with "transfer" doses. These arise when a monitored worker moves between places of employment, either within or outside the authorities that make up this study cohort. When this happens either all annual radiation records for that worker are transferred to the new employer’s radiation records, or a single summary "transfer" dose record is appended to the new records. Although in many cases single transfer doses could be associated with a known period of employment (hence a period of "uncertain" monitoring), for some subjects there was no way of knowing what period the transfer dose covered. In this situation they were designated as being of "uncertain" monitoring status from age 16. Among workers who had, or had ever attempted to have, children 164 (0.9%) men and 9 (0.2%) women were allocated "uncertain" monitoring status back to age 16 in this way. On some occasions a transfer dose was found to be replicated by annual doses—that is, we had both the annual dosimetry data for the period and a transfer dose, in which case the latter was removed.

Overall, among workers who had, or had ever attempted to have, children 942 (5%) men and 51 (1%) women had at least one period of "uncertain" monitoring prior to the date of survey, although only nine men and one woman had a period of "uncertain" employment. In terms of the pregnancies these workers conceived, 255 (1%) of pregnancies reported by men and 11 (0.1%) reported by women were conceived during a period of "uncertain" monitoring, with just five pregnancies reported by men and none by women being conceived during a period of "uncertain" employment.

Results

Table 1 Composition of occupational cohort: employees of the Atomic Energy Authority (AEA), Atomic
Weapons Establishment (AWE) and British Nuclear Fuels (BNFL) aged under 75 years at survey

Fig 1 Geographical location of principal establishments of the Atomic Energy Authority, Atomic Weapons Establishment, and British Nuclear Fuels

Response

Seven per cent (3068) of all questionnaires (excluding reminders) were returned undelivered. As expected, past workers were harder to trace, with 1596 (10%) questionnaires being returned because the addressee had moved and a further 176 (1%) because the addressee had died or was too sick to participate. Among workers assumed to have received a questionnaire, the response rate for completed questionnaires was 82% for men and 88% for women (representing 78% of the total survey population of 46 396 male and female workers originally identified for the study). This adjusted response rate was uniformly high across authorities, and by employment status. Only 3% (1276) of workers whose questionnaires were assumed to have been delivered chose the option of returning a blank questionnaire, or stated that they did not wish to participate (table 2).

Table 2 Response rates and proportion of workers who had ever attempted to have children, by
authority, employment status and sex among workers aged under 75 yearsa at time of survey

Characteristics of non-respondents

Among current Atomic Energy Authority and Atomic Weapons Establishment workers, responders tended to be younger than non-responders, but the difference in mean age was not large, ranging from 1.3 years to 4.2 years (table 3). By contrast, past workers at the Atomic Energy Authority who responded tended to be older than those not completing a questionnaire, with a difference in mean age of 10 years for men and 8.3 years for women. Although no data on age were available for non-responders at British Nuclear Fuels, the distribution of age among those returning completed questionnaires was similar to that of Atomic Energy Authority workers for both current and past employees.

Table 3 Age distribution at survey by sex among workers aged under 75 years who returned a completed questionnaire, and by authority, sex and whether a completed questionnaire was received for all those posted a questionnairea

Together with the high response rate, these data show that, in terms of age, current workers who responded are representative of the total current cohort surveyed. Among past workers there is some evidence of a bias in favour of older responders. The most likely explanation is that younger ex-workers, who are not yet receiving their pension, may be less likely to have an up to date home address registered with the pensions office of their previous employer. Thus a higher proportion of the questionnaires may have failed to reach them than for older ex-workers who are receiving their pension. This age bias may have inflated the proportion of past workers reporting pregnancies and children, but it is unlikely that it would have resulted in biased reporting of adverse pregnancy outcome or childhood illness.

Characteristics of respondents reporting pregnancies (or an attempt for pregnancy)

At the time of survey 18 744 (67%) men and 4702 (58%) women who returned a completed questionnaire had attempted to have children, reporting either at least one pregnancy or primary infertility (table 2). These proportions were higher in past workers, reflecting the older age distribution, but were similar among employees of the three different authorities (table 2).

Characteristics of the workers reporting pregnancies or an attempt for pregnancy are presented in table 4. At the time of the survey male workers were older than female workers and had served longer. Further, 82% of the men had been monitored at some time compared with only 28% of women. Among employees of the Atomic Energy Authority and British Nuclear Fuels, 74% of ever monitored and 71% of never monitored men had been in continuous service at one site until either leaving the industry or date of survey. Similar proportions were observed among women: 71% of ever monitored and 72% of never monitored workers had been in continuous service at one site (data not shown).

Only 328 (2%) male workers and 123 (3%) female workers reported that they had attempted to have a child but had never had a pregnancy. The remaining men and women reported an average of 2.4 and 2.2 pregnancies respectively per worker, 96 men and 57 women reporting a first pregnancy current at the time of survey. Overall, 18 131 men and 4435 women reported at least one liveborn child, with an average of 2.2 and 2.0 liveborn children respectively per worker (table 4). The questionnaire included a question asking whether the worker’s partner had ever worked in the industry. Using this information, and a variety of methods including matching on children’s names, dates, and place of birth, it was possible to determine that 1232 male workers and 1233 female workers shared one or more liveborn children with a partner who was in the survey population. These included two male workers and one female worker who shared liveborn children with two different partners in the survey.

Table 4 Characteristics of all workers reporting ever having attempted to have childrena

Characteristics of pregnancies and offspring

The characteristics of the 43 710 pregnancies reported by male workers, and the 9962 reported by female workers, are presented in table 5. More than half of all pregnancies reported by men were conceived after first employment by the Atomic Energy Authority, Atomic Weapons Establishment, or British Nuclear Fuels. Around 40% were conceived after the father had been first monitored, the median lifetime cumulative preconceptional dose being 12.6 mSv. In only 9% of these pregnancies conceived by ever monitored fathers (4% of all pregnancies reported by men) did the father have a cumulative preconceptional dose of 100 mSv or higher (table 5). At the other end of the scale, only 716 pregnancies conceived by ever monitored fathers (2% of all pregnancies reported by men) had a cumulative preconceptional dose of less than 0.1 mSv.

Table 5 Characteristics of pregnancies

Among pregnancies reported by female workers, just over a third were conceived after first employment by one of the three authorities, and 10% were conceived by ever monitored women (table 5). The median lifetime cumulative preconceptional dose for these latter pregnancies was 4.2 mSv, with only one pregnancy being conceived by a woman who had a lifetime cumulative whole body dose of 100 mSv or more. Nevertheless, most of these pregnancies had a cumulative preconceptional dose of more than 0.1 mSv, only 62 pregnancies conceived by ever monitored mothers (1% of all pregnancies reported by women) having a cumulative preconceptional dose of less than 0.1 mSv.

In total, 39 557 liveborn children were reported by men and 8883 by women; their characteristics are summarised in table 6. Of the 46 107 children in the study, 2333 had both a father and a mother in the survey population. The average age of living children at time of survey was 22.8 years for children of male workers and 21.6 years for children of female workers. The median preconceptional dose among children whose parents were monitored prior to their conception was 12.4 mSv for fathers and 4.3 mSv for mothers. Four per cent of male workers’ children and 0.01% of female workers’ children had preconceptional doses of 100 mSv or higher, and 2% and 1% had doses of less than 0.1 mSv prior to conception. Two per cent of children reported by male workers and 6% reported by female workers had a parent who was not employed or monitored at any time either before the child’s conception or during their childhood and early adulthood (prior to death, 25th birthday, or date of survey, whichever was earliest). Parental monitoring exclusive to the preconceptional period was rare (table 6).

Table 6 Characteristics of live births

Discussion

We have assembled a database of 53 672 pregnancies, the largest occupation based study of reproductive outcome in the United Kingdom. Most of these pregnancies were reported by men, since they made up the bulk of the occupational group studied. We obtained complete employment and monitoring data from employers that could be linked to each parent’s reproductive life in an unbiased way. In particular, it was possible to link both employment and monitoring information to the estimated time of conception of each reported pregnancy. Among pregnancies reported by male workers, the fathers in just over half had been monitored prior to conception, but most had relatively low cumulative preconceptional whole body doses. As expected, preconceptional monitoring was rare among women.

There is increasing scientific and public interest in possible adverse reproductive effects of workplace exposures. Studies such as the one reported here, where we have shown the feasibility of collecting detailed information on reproductive outcome from men and women in an occupational setting, are necessary in order to investigate such effects. The methods used in this study have now been adapted for use in other important occupational investigations such as the study of reproductive outcome of veterans of the Gulf war.

We acknowledge the skills and commitment of all those who worked on the study at London School of Hygiene and Tropical Medicine, including Patricia Fraser, Juliet Jain, Patrick Sampson, Tuyet Ngyen, Mary Sewell, Margo Pelerin, Angela MacCarthy, Aurora Berra, and Amanda Thomas. We thank our scientific steering group (Eva Alberman, Tim Bishop, Martin Bobrow, and Dudley Goodhead) for their continued support and guidance. For supplying occupational data we thank Dallas Law, Will Atkinson, and Keith Bromley of the AEA; George Sallit and Pauline Johnson of the AWE; and Keith Binks, Sheila Jones, and Les Scott of British Nuclear Fuels Limited. We are also grateful to representatives of the management and workforces of the three authorities who gave permission for the study to be conducted. Most importantly, we thank the study participants themselves.

Funding: Department of Health and the Health and Safety Executive.

Contributors: NM (guarantor) participated in protocol development, data collection, analysis and writing the paper. PD participated in protocol design, data collection, analysis and writing the paper. ER initiated the research and participated in protocol design, data collection, analysis and writing the paper. GD participated in data collection and analysis. PGS initiated the research and participated in protocol design and writing the paper. VB initiated the research and participated in protocol design and writing the paper.

Competing interests: PS has received funding from British Nuclear Fuels Ltd for research on the health of Sellafield workers.

1. Black D. Investigations of the possible increased incidence of cancer in West Cumbria. London: HMSO, 1984. (Report of the Independent Advisory Group (chairman Sir Douglas Black).)
2. Urquhart J, Palmer M, Burke M. Cancer in Cumbria: the Windscale connection. Lancet 1984;i:217-8.
3. Heasman MA, Kemp IW, Urquhart JD, Black R. Childhood leukaemia in northern Scotland. Lancet 1986;i:266.
4. Gardner MJ, Hall AJ, Downes S, Terrell JD. Follow up study of children born to mothers resident in Seascale, West Cumbria (birth cohort). BMJ 1987;295:822-8.
5. Gardner MJ, Hall AJ, Downes S, Terrell JD. Follow up study of children born elsewhere but attending schools in Seascale, West Cumbria (schools cohort). BMJ 1987;295:819-22.
6. Roman E, Beral V, Carpenter L, Watson A, Barton C, Ryder H, et al. Childhood leukaemia in the West Berkshire and Basingstoke and North Hampshire District Health Authorities in relation to nuclear establishments in the vicinity. BMJ 1987;294:597-602.
7. Committee on Medical Aspects of Radiation in the Environment (COMARE). First report. The implications of the new data on the releases from Sellafield in the 1950s for the conclusions of the report on the investigations of the possible increased incidence of cancer in west Cumbria. London: HMSO, 1986.
8. Committee on Medical Aspects of Radiation in the Environment (COMARE). Second report. Investigation of the possible increased incidence of leukaemia in young people near the Dounreay nuclear establishment, Caithness, Scotland. London: HMSO, 1988.
9. Committee on Medical Aspects of Radiation in the Environment (COMARE). Third report. Report on the incidence of childhood cancer in the West Berkshire and North Hampshire area in which are sited the Atomic Weapons Research Establishment, Aldermaston and the Royal Ordnance Factory, Burghfield. London: HMSO, 1989.
10. Gardner M, Snee M, Hall A, Powell C, Downes S, Terrell J. Results of case-control study of leukaemia and lymphoma among young people near Sellafield nuclear plant in West Cumbria. BMJ 1990;300:423-9.
11. Urquhart JD, Black RJ, Muirhead MJ, Sharp L, Maxwell M, Eden OB, et al. Case-control study of leukaemia and non-Hodgkin’s lymphoma in children in Caithness near the Dounreay nuclear installation. BMJ 1991;302:687-92.
12. Roman E, Watson A, Beral V, Buckle S, Bull D, Baker K, et al. Case-control study of leukaemia and non-Hodgkin’s lymphoma among children aged 0-4 years living in West Berkshire and North Hampshire health districts. BMJ 1993;306:615-21.
13. Draper GJ, Little MP, Sorahan T, Kinlen LJ, Bunch KJ, Conquest AJ, et al. Cancer in the offspring of radiation workers:a record linkage study. BMJ 1997;315:1181-8.
14. Beral V, Inskip H, Fraser P, Booth M, Coleman D, Rose G. Mortality of employees of the United Kingdom Atomic Energy Authority, 1946-79. BMJ 1985;291:440-7.
15. Smith PG, Douglas AJ. Mortality of workers at the Sellafield plant of British Nuclear Fuels. BMJ 1986;293:845-54.
16. Beral V, Fraser P, Carpenter L, Booth M, Brown A, Rose G. Mortality of employees of the Atomic Weapons Establishment 1951-82. BMJ 1988;297:757-70.
17. Kendall GM, Muirhead CR, MacGibbon BH, O’Hagan JA, Conquest AJ, Goodill AA, et al. Mortality and occupational exposure to radiation: first analysis of the National Registry for Radiation Workers. BMJ 1992;304:220-5.
18. Fraser P, Carpenter L, Maconochie N, Higgins C, Booth M, Beral V. Cancer mortality and morbidity in employees of the United Kingdom Atomic Energy Authority, 1946-86. Br J Cancer 1993;67:615-24.
19. Douglas AJ, Omar RZ, Smith PG. Cancer mortality and morbidity among workers of the Sellafield plant of British Nuclear Fuels. Br J Cancer 1994;70:1232-43.
20. Carpenter L, Higgins C, Douglas A, Fraser P, Beral V, Smith P. Combined analysis of mortality in three United Kingdom nuclear industry workforces, 1946-1988. Radiat Res 1994;138:224-38.
21. Carpenter LM, Higgins CD, Douglas AJ, Maconochie NES, Omar RZ, Fraser P, et al. Cancer mortality in relation to monitoring for radionuclide exposure in three UK nuclear industry workforces. Br J Cancer 1998;78:1224-32.
22. Committee on the Biological Effects of Ionizing Radiation (BIER V). Health Effects of exposure to low levels of ionizing radiation. Washington, DC: National Academy Press, 1990.
23. Roman E, Doyle P, Ansell P, Bull D, Beral V. Health of children born to medical radiographers. Occup Environ Med 1996;53:73-9.
24. Oracle Corporation (www.oracle.com).
25. StataCorp. Stata Statistical Software, Release 5.0. College Station, TX: Stata Corporation, 1996.
26. World Health Organisation. Manual of the international statistical classification of diseases and related health problems, 10th revision. Geneva: World Health Organisation, 1992.
27. World Health Organisation. Manual of the international classification of diseases, 9th revision. Geneva: World Health Organisation, 1977.
28. Smith JW, Inskip H. Estimation of below measurement threshold doses following the remeasurement of a sample of old films. J Soc Radiol Prot 1985;5:159-64.
29. Taylor NA. Estimation of dose received when dosemeter results are recorded below a threshold level. J Radiol Prot 1991;11:191-8.