What is the NWTPS?
What information am I going to find in this fact sheet?
Who is responsible for the NWTPS?
Who operates the Nuclear Weapons Test Participants Study?
Why was this study undertaken?
Why was the NRPB selected to carry out the NWTPS research?
How does an epidemiological study work?
What has the NWTPS study shown?
How is information captured for the study?
How were study participants identified?
What is the ‘Blue Book’ data?
How do I know that the research is not biased?
Is there any independent scrutiny of the NWTPS?
Who funds the research?
Apart from MoD, does anyone else provide funds for the researchers?
Who is included in the study?
What is the Control Group?
Why did the study need a control group?
Why did the study rely on MoD records? Surely this could have biased the results?
How do we know that the researchers were provided with all relevant records?
Does it matter if people have been missed from the list of participants?
Why does the study not include all service personnel who were test participants?
Could the 15% of test participants who were not included in the study have experienced very different health outcomes to the 85% who were included?
I have heard that the NRPB studies counted an individual more than once if they attended more than one test. Is this true?
I’ve heard that some people were included in the study even though they were not really involved in the tests – is this true?
Were all test participants exposed to radiation?
How many people were involved in the nuclear weapons test program?
Who is not included in the study?
Why don’t the researchers add me to the study cohort if I can prove that I was at the tests?
Do the researchers know what level of radiation exposure I received?
If radiation dose data were ‘lost’ or not reported to the researchers, would this mean the study will be invalid?
Should I be worried about excessive radiation exposure because I saw the flash through my hands?
I see references in the press which say that most nuclear weapons test participants are now dead. Is this true?
Did the study only look at cancer deaths?
I have been told that the researchers are only interested in participants who have died of cancer since 1971. Is this true?
Why did the researchers not include cancer incidence data from before 1971?
Why don’t the researchers use a different data resource for cancer information then?
What details do the HPA researchers hold?
I’ve heard that the researchers will require access to my body or tissues after I’m dead – is this true?
Are my details held securely?
Who gave permission for my details to be included in the study?
I’ve heard about some research from New Zealand – is this different to the UK study?
What is the HPA and what is the CRCE?
Who or what is the NRPB?
What is the AWE or AWRE?
How can I find out more about the nuclear weapons tests?
How can I find out more about the NWTPS study?
What publications are available?
The Nuclear Weapons Test Participants Study (NWTPS) is a long-term follow up study of the health of UK personnel who were present at areas in Australia and the South Pacific Islands where the UK nuclear atmospheric weapons tests were conducted between 1952 and 1967.
This fact sheet attempts to provide responses to a number of frequently asked questions. In many cases the answers given here summarise information which is published elsewhere. In particular, our reports provide very detailed explanations of the work undertaken in connection with this study including details of data collection and data verification as well as the way that the statistical analyses were conducted, results analysed and conclusions drawn. A reference list that details related publications is shown towards the end of this fact sheet.
This project, a long-term follow up epidemiological study, is the responsibility of researchers at the Health Protection Agency’s Centre for Radiation, Chemical and Environmental Hazards (HPA-CRCE). The work is conducted in a department that was formerly part of the NRPB.
The NWTPS is operated by staff within the Health Protection Agency’s CRCE Epidemiology Department. The team consists of database staff, statisticians and epidemiologists. They also work on other projects.
In the years after the UK atmospheric nuclear weapons tests there were a number of suggestions that the health of participants had suffered and, in
1983, scientists from the National Radiological Protection Board (NRPB) and the Imperial Cancer Research Fund (now Cancer Research UK) were commissioned by the Ministry of Defence (MoD) to carry out an independent epidemiological study of participants in the UK atmospheric nuclear weapons tests, both in the Pacific and in Australia.
A key point about this study is that it included a well defined group of test participants and a control group of men who had not taken part in the tests but who were otherwise as similar as possible to the test participants. This overcame the difficulties faced by those who looked at smaller sets of response-sourced data.
The National Radiological Protection Board (NRPB) had expertise in epidemiological research of radiation-exposed populations and, by its function, an interest in furthering the knowledge of health effects of human exposure to radiation. NRPB was a non-departmental public body, able to conduct research and to publish findings independently. The combination of the database, statistical and epidemiological expertise of the researchers with access to NRPB colleagues such as radiation physicists, chemists and biologists provided expertise that was able to cover the wide range of issues that were relevant to this research.
The epidemiological researchers at HPA-CRCE (as NRPB has become) are also involved in other research including the largest study of UK radiation workers (the NRRW, the National Registry for Radiation Workers) and European and international studies of radiation workers.
An epidemiological study such as this involves linking details about a person’s involvement in an event or events (such as attendance at the nuclear test programme) with outcome data such as cause of death or cancer incidence. Statistical analyses then consider what conclusions may be drawn concerning risks of mortality and cancer incidence from comparisons between the study cohort and other comparable groups.
For this study, comparisons were made against the national population and against a control group.
The main finding was that test participants had, in general, a better life expectancy than members of the general UK population. This ‘Healthy Worker Effect’ (or ‘Healthy Soldier Effect’) reflects the fact that a group of people who have been employed will (overall) be healthier than the general population because the general population will include the chronically sick who are unable to work.
When compared with the control group, the test participant group had similar overall levels of age-related mortality and cancer incidence.
There was also, however, an indication that test participation may have caused a small increase in the risk of leukaemia (other than Chronic Lymphatic Leukaemia). The evidence for any such increased risk appears to have been greatest in the early years after the tests, but a small risk may have persisted in subsequent years. There was not, however, enough evidence to confirm that this is a fact and there is some evidence to indicate that the finding might be a statistical artefact (a ‘chance finding’).
The NWTPS is a records based study, which means that details for each individual are drawn from existing records and do not require individuals to be contacted for data or to complete a questionnaire. This approach both reduces the impact on the individual as well as reducing any risk of bias in response from study subjects. This does help to assure that any findings are not affected by that particular bias.
The published NWTPS reports include more detail about data collection.
The NRPB researchers were faced with a huge task as there had never been a complete list of all those that had attended the nuclear weapons tests. The study team had to rely on searches of a vast number of documents dating from the time of the weapons tests to identify the test participants.
Many names appeared on more than one document but it became clear that some people were not listed on any of the main documentation and further searches, including searches of service records, were undertaken.
The published NRPB reports do describe, in some detail, the work that was undertaken to identify the study participants and the work that was also undertaken to verify the information collected and its completeness.
The ‘Blue Books’ were compiled between 1981 and 1982 by AWE (the Atomic Weapons Establishment) using information from documents that had been written during the period that the nuclear weapons test programme was undertaken. It is not a complete list of people who attended the test sites. No complete list was ever compiled.
In 1982, the list that had been compiled by AWE was sorted (by name within service) and printed off in a series of documents which were bound in blue covers and became known as ‘the Blue Books’. The ‘Blue Book’ is dated 1982. The foreword to the listing states that it is a “summary provisional listing” and that “the listing is not claimed to be free from errors, incorrect assertions or omissions”.
The ‘Blue Book’ data was used by NRPB as one of the sources of information for identifying the NWTPS study participants. The NRPB researchers had access to the original documents that AWE had used to compile the ‘Blue Books’ as well as additional information resources that were identified as the NRPB work progressed. The researchers were careful to ensure that all the documents used dated from the period of the test programme (‘contemporary documents’) to ensure that no bias was introduced into the study cohort. The sources are listed in the NRPB reports and include health physics documents, planning documents and trials reports.
The HPA researchers are independent. They do not work for the MoD but are employed by the independent HPA and were formerly employed by the independent NRPB. All of the researchers’ reports have been published, in full, by the NRPB and additional papers have been published in respected peer-reviewed scientific journals. The reports and papers are written by the researchers and are not subject to MoD approval. In order to further demonstrate independence, the most recent analysis was overseen by an advisory group led by Professor Nicholas Wald, FRS.
The HPA is an independent organisation and research such as the NWTPS study is carefully conducted. The researchers do, however, welcome independent peer-review and the paragraphs below indicate some of the ways that this is provided.
An independent advisory group was set up under the Chairmanship of Professor Nicholas Wald to provide independent oversight of the third analysis. Details of the advisory group can be found in the analysis report (NRPB-W27, references below).
When reports are published in reputable scientific journals, those reports are first examined by independent experts who will provide a view on the reports, the findings and the work undertaken. If they think that the work has not been undertaken properly they will say so. This is called a peer-review process and such reputable journals will only publish papers which are accepted by the independent experts it consults. The three NWTPS analysis projects completed so far have been reported in two highly respected journals – the British Medical Journal, and the Journal of Occupational and Environmental Medicine.
The NWTPS study and database work undertaken by the HPA (and formerly the NRPB) are funded through a contract with the Ministry of Defence (MoD).
The HPA-CRCE Epidemiology department is involved in a range of work and the NWTPS work is only one of its research projects. Other work is funded by a variety of sources including Government departments and agencies such as the Department of Health, the Health and Safety Executive, by the European Commission as well as by charities such as Cancer Research UK.
The NRPB study aimed to include all UK personnel from the RAF, the Army, the Royal Navy, AWE or UKAEA (UK Atomic Energy Authority) who had had potential for exposures to radiation as a result of involvement in the UK atmospheric nuclear weapon test programmes in Australia or the Christmas Island region between 1952 and 1967.
The control group contained roughly the same number of men as the participants’ group and, apart from not participating in the tests, the controls were chosen to have similar characteristics to the participants. For test participants in the armed forces, the controls were selected from service personnel who served in tropical or sub-tropical areas other than the test locations around the time that the tests were taking place. For AWE test participants, the controls were chosen from other men working for AWE at around the same time as the weapons tests. The 22,333 men in the control group were very similar to the participants with respect to the split between services, ranks/social class, year of birth, year of enlistment/employment and year of discharge/end of employment
Participants in the nuclear weapons test programme would have differed in some ways from men of the same age in the general UK population. For example, test participants needed to have been fit enough to be selected for overseas service, and they would have experienced a different lifestyle during their period in a tropical or desert environment. Consequently, as well as comparing mortality and cancer rates among test participants with the corresponding national rates, comparisons were also made with the control group.
There was no alternative to using MoD records for establishing a substantially complete cohort and obtaining estimates of radiation dose. All other sources would have been far less complete and most would have been biased in terms of the health experience of those included. Nevertheless, attempts were made to collect as much data as possible from other sources, both to assess the completeness of coverage of the study population and to examine the possibility of bias.
The researchers were aware from the outset of their work that incomplete records could bias their results. An important aspect of the researchers work was therefore to identify all relevant records and to check that these records were correctly examined. The 1988 report details the work that the researchers undertook to verify this.
The researchers also wanted assurance that the study dataset was not biased in any way and they therefore asked for assistance from veterans’ organisations and a number of other independent sources. This ‘independent responder’ information established that the NWTPS dataset includes approximately 85% of all eligible test participants and that the study coverage is higher than 85% for all of the groups identified as potentially being at higher risk. This is fully documented in the NRPB reports.
It was not necessary to include all test participants in the study, though it was desirable that as many as possible of them were included, as the ability of the study to detect any harmful effect of test participation would be larger the larger the population studied.
It was, however, essential that men were selected for inclusion in the study in an unbiased way. That is, that they were representative of the whole group of test participants in terms of their subsequent health, and were not selected in such a way as to over or under represent those who have subsequently suffered ill-health or died.
The database for the study was created following the examination of an extensive range of documents which had been compiled during the test programme of the 1950s and 1960s. These included planning documents, technical reports, health physics records, Royal Navy ships’ ledgers and RAF operational record books as well as service records and other documents. Unfortunately however, there never had been any complete log of those involved and so it was not possible to check against such a list.
The question of potential bias of the study population derived from MoD records is an important one and the researchers were alert to it from the start. The work that was undertaken to investigate such a possibility is discussed in detail in the published reports.
In particular, in addition to the lists compiled from official sources, the NRPB/HPA researchers were also able to assemble a list of test participants from sources independent of MoD. It was this list which enabled the researchers to estimate that approximately 15% of individuals may not have been included on the list compiled from MoD records.
Looking at the total group of people identified from independent sources (including those that had also been identified from the MoD records), it was apparent that the health experience of the people included on this independent list was worse than that expected from comparisons with the general population and also worse than the participant group identified from the contemporary MoD documents. This bias had been anticipated as people who are not suffering ill-health are less likely to report, than those who are, or have been, suffering.
When the researchers looked within the independently compiled list however, they found that the health experience of those that had also been identified from the official MoD records was similar to the health experience of the 15% who had not been identified from the official MoD records.
This provides powerful evidence against the idea that individuals were significantly less likely to be included in the study population if their health had suffered.
No, this is not the case. The researchers were aware that many individuals attended more than one test and their reports indicate this clearly. The most recent (2003) report indicates that the study cohort included 21,357 men for whom 27,505 attendances are recorded. This includes attendances at the Christmas Island tests as well as the attendances at tests in Australia.
There is a dilemma in defining the population to be studied. On the one hand, if it included only those clearly likely to be exposed to radiation (eg aircrew sampling the radioactive plumes from the explosions), then it would exclude other groups who should also be regarded as bona fide test participants. On the other hand, including as many as possible of those involved in the tests would inevitably include some groups who had little real chance of having been exposed and might dilute away any excess of disease in those most at risk.
To try to deal with this issue, the investigators took a wide definition of test participation but also carried out special analyses of those sub-groups which might have been at increased risk. A fuller description is available in the published reports.
All of the world’s population is exposed to radiation to some extent as naturally occurring radiation affects us all, whether at a nuclear test site or not.
Some test participants were exposed to higher levels of radiation as a result of their work within the test programme and, as detailed in the published reports, these groups of people were looked at in specific analysis sub-groups.
There is no complete list of those who visited test sites during the atmospheric nuclear test programme. The 2003 report details the study cohort of 21,357 participants and completeness checks undertaken by the NRPB/HPA researchers indicate that this represents 85% of the total eligible study population.
Although employees of many other organisations were involved in the test programme (the Meteorological Office and the Merchant Navy for example), the records available to the researchers in 1983 were not sufficiently complete to allow identification of the relevant individuals and so these groups were not included in the NRPB study group. Health follow-up for non-UK personnel could not be ascertained in the same way as for UK personnel and therefore foreign personnel were also not included in the NWTPS study. Additionally, it was also decided to exclude the very few women who had taken part in the tests as the number was too small for useful analysis.
The exclusion of records from the study does not mean that the individuals are necessarily different to those included in the study, rather that the data was not sufficiently complete to add value to the epidemiological work.
The researchers have established a large study cohort based on information from contemporary documents. They have also used information from test participants to identify how compete their study database is. This ‘independent responder’ information has helped to establish that the NWTPS includes approximately 85% of all eligible test participants and that the study coverage is higher than 85% for all of the groups identified as potentially being at higher risk. This is fully documented in the NRPB reports.
The researchers will not add additional records to the study cohort because they do not want to introduce ‘bias’ to the study. For example, if they added details of those men who confirmed that they had been present this could distort the findings because those that had already died could not contact the researchers.
Health physics records were made available to the NRPB/HPA researchers. These provided lists of individuals who had been issued with film badges to record their possible exposure to external radiation and details of any doses above the threshold of detection which had been incurred.
The published reports provide more detail about what is known of the monitoring regime at the tests but the researchers were content that the data available to them was sufficient and they were also happy that their study design was such that significant unreported exposures would also have been apparent to them.
The study was designed in such a way as to pick up increased risks of cancer incidence or mortality (from cancer or other causes) to participants from unreported radiation exposure – or from any other hazards - if they had been large enough.
It has not always been appreciated that the contemporary records indicate that relatively few test participants received any measurable radiation dose as a result of the tests.
Some participants have thought that the flash of intense light that accompanied some explosions and could be detected through their fingers with their hands held over their eyes was a dose of x-rays. In fact, x-rays are not visible and it was just intense light that was visible through their hands.
At the time of the last analysis, only 4,902 (23%) of the 21,357 study participants were known to have died. Some had emigrated but the majority, 14,560 (68%) were still alive.
As nuclear weapons test participants were adults in the 1950s and 1960s, even the youngest of them would, by the year 2000, have been aged over 50. Those aged in their later twenties or older at the time of the earliest test, in 1952, would have been at least in their seventies by the year 2000.
The life expectancy of a boy born in 1930 was less than 60 years so it is not surprising that an increasing number of former test participants are no longer alive.
It should be noted, however, that some of the figures quoted in the press are not based on coverage of all of those that attended a test site. Some only relate to those who attended a particular test site and many are based on other subsets of the group studied by the NRPB/HPA.
No, the NWTPS examines all causes of death, whether from cancer or any other disease or condition. The NRPB reports show details.
No, this is not true. The researchers included all 21,357 study participants in the 2003 analysis. For the 4,902 men that had died before 1st January 1999, all cause of death information was considered, whether the cause was cancer or another disease or condition. Additionally, the researchers have been able to examine the cancer registration data for men who contracted cancer after 1971, whether or not they subsequently died of the disease.
Before 1971 there was no UK-wide system for systematically recording details relating to patients who were diagnosed with cancer. From 1971, as cancer treatment improved and a greater percentage of patients recovered from cancer, regional cancer registries began to work together to collate a national registry and from that point more complete and accurate data is available, mainly for the purposes of judging the benefits of different treatment programmes, etc, but also for epidemiological research studies such as the NWTPS.
Nonetheless, some information about cancer incidence before 1971 is available from information on death certificates and this information was included in the analyses.
The UK national registries are internationally recognised as an invaluable resource in providing excellent and unbiased information about deaths and cancers in the UK population. There is no better resource available for this type of study.
The NWTPS database holds information to identify the test participants (names, service numbers, dates of birth, etc); information about their test participation (name of test, test sites, etc); and, where these events have occurred, follow-up data (such as date and causes of death, and date and details of cancer incidence information).
The information on the database was collated from a variety of sources but the key sources were the contemporary documents detailing test participation, service or employment records and the NHS Central Registers (NHSCRs).
The data collected by the NRPB/HPA researchers from 1983 was collated for the purpose of epidemiological research and the study is registered, with the NHS Information Centre, as a Medical Research study. The HPA database is also registered under the Data Protection Act.
No, this is not true. The epidemiological study conducted by the NRPB/HPA researchers is an approved ‘medical research’ study but there are many types of medical research studies and they do not all require specific examination of body tissues. This epidemiological study is a records-based long-term follow-up study. This means that the details about an individual’s cause of death or their cancer incidence are provided through information sources that already exist. These sources are death certificates and cancer registrations, records which are routinely collated for all of the UK population.
The NWTPS database is registered under the Data Protection Act and information security is taken very seriously by the study team. The database is held on a secure system and is accessible only to authorised users who require various credentials and authorisations to be able to access the database. Personal identifiable information collected for the NWTPS is not made available for any other purpose.
The study was set up in 1983. If it had been necessary to have permission from each individual who had been involved in the tests programme, it would have meant that the researchers needed to trace and contact over 20,000 people. As some of those men were already dead, they would not have been able to reply. Additionally, it is known that if only those that agreed were included in the study, there would have been a higher response rate from those that had been ill than those that had not. Both of these issues would have made it far more difficult to analyse the results of the study as it would have been necessary to take account of the impact of those issues.
The nuclear test veterans’ organisations were involved in discussions with the researchers as the study was developed and the progress and the results of the work have been openly published and widely publicised.
The study was considered, and approved, by the British Medical Association Central Ethical Committee in 1983. As with many other studies of a similar nature, the NWTPS has been reviewed in recent years under the PIAG (Patient Information Advisory Group) and NIGB (National Information Governance Board) reviews of Medical Research studies using NHSCR data without explicit consent. The study continues to retain support for continued data processing under section 251 of the NHS Act 2006.
A team from Massey University in New Zealand did conduct a study of chromosome aberrations in blood samples taken from a group of fifty New Zealand test veterans. This particular study took advantage of developments in biological research techniques to carry out a project which was entirely different in approach to the epidemiological studies that have been undertaken by HPA.
The HPA was interested in the New Zealand research and in 2007, when the results were first presented, the HPA’s cytogenetic specialists reviewed that research.
In summary, the HPA concurred with the New Zealand authors that the results of their work did indicate a statistically significant threefold increase in stable chromosomal translocations for veterans (compared to their control group) and also agreed that it would be possible to ascribe this increase to radiation exposure. However, the presence of unstable aberrations left open the question of whether all of the dose was delivered around the time of the nuclear weapons tests.
The Health Protection Agency (HPA) is an organisation set up in 2003 in order to provide an integrated approach to protecting UK public health through the provision of support and advice. The HPA is a non-departmental public body. The website (at www.hpa.org.uk) provides more detail.
The Centre for Radiation, Chemical and Environmental Hazards (CRCE), which has its headquarters at Chilton in Oxfordshire, provides the focus for research and advice on radiological protection as well as chemical and environmental hazards.
The National Radiological Protection Board (NRPB) was the focus for radiological protection matters in the UK. The NRPB became part of the HPA in 2005 and now forms part of the HPA-CRCE.
The study of the nuclear weapons test participants was started under the NRPB and most of the current research team worked on the study as employees of the NRPB.
The Atomic Weapons Establishment, AWE, was formerly known as AWRE, the Atomic Weapons Research Establishment. This organisation, based at Aldermaston in Berkshire, led the UK atmospheric nuclear weapons test programme.
The NRPB reports and papers give some background detail and provide references to other documents that might help.
Further information is also available through the MoD website (go to www.mod.uk and enter British Nuclear Test Programme in the search box).
The NRPB reports and papers are very detailed. The most recent full report (NRPB-W27, 2003) is available from the HPA website and two review papers, published in 2004, are also still available. A list of related NRPB publications if given elsewhere in this fact sheet.
The NRPB (National Radiological Protection Board) published a number of reports/papers in relation to epidemiological analyses of UK participants in the UK Atmospheric Nuclear Weapons Tests and Experimental Programmes. Details of the key publications are shown below.
The most recent publications were completed in 2004 when 2 review papers describing the NRPB epidemiological studies and the results were published in the scientific journal, the Journal of Radiological Protection.
Since 1983, the NRPB (now HPA-CRCE) researchers have completed three full epidemiological analyses of the UK Participants in the UK Atmospheric Nuclear Weapons Tests and Experimental Programmes. The first epidemiological analysis was completed in 1988 and described the findings based on follow-up to 1983. A subsequent analysis, reported in 1993, described findings based on follow-up to 1990 and a third analysis, which considered follow-up to the end of 1998, was published in 2003. Each of these analyses was reported in a full report published by NRPB and also in a shorter paper which was published in a scientific journal (and as such was reviewed by independent scientists before publication).
1st analysis (1988): A summary of mortality and incidence of cancer in men from the United Kingdom who participated in the United Kingdom's atmospheric nuclear weapon tests and experimental programmes. British Medical Journal, 296, 332-338. Darby S C, Kendall G M, Fell T P, O'Hagan J A, Muirhead C R, Ennis J R, Ball A M, Dennis J A and Doll R.
2nd analysis (1993): Further follow-up of mortality and incidence of cancer in men from the United Kingdom who participated in the United Kingdom's atmospheric nuclear weapon tests and experimental programmes. British Medical Journal, 307, 1530-1535. Darby S C, Kendall G M, Fell T P, Doll R, Goodill A A, Conquest A J, Jackson D A and Haylock R G E.
3rd analysis (2003): Follow up of mortality and incidence of cancer 1952-1998 in men from the UK who participated in the UK's atmospheric nuclear weapons tests and experimental programmes. Occupational and Environmental Medicine, 60, 165-172. Muirhead C R, Bingham D, Haylock R G E, O'Hagan J A, Goodill A A, Berridge G L C, English M A, Hunter N and Kendall G M.
1st analysis (1988): Mortality and cancer incidence in UK participants in UK atmospheric nuclear weapon tests and experimental programmes. Chilton, NRPB-R214 (London, HMSO). Darby S C, Kendall G M, Fell T P, O'Hagan J A, Muirhead C R, Ennis J R, Ball A M, Dennis J A and Doll R.
2nd analysis (1993): Mortality and cancer incidence 1952-1990 in UK participants in the UK atmospheric nuclear weapon tests and experimental programmes.Chilton, NRPB-R266 (London, HMSO). Darby S C, Kendall G M, Fell T P, Doll R, Goodill A A, Conquest A J, Jackson D A and Haylock R G E. (1993b)
3rd analysis: Mortality and Cancer Incidence 1952-1998 in UK Participants in the UK Atmospheric Nuclear Weapons Tests and Experimental Programmes. Chilton, NRPB-W27. Muirhead C R, Bingham D, Haylock R G E, O'Hagan J A, Goodill A A, Berridge G L C, English M A, Hunter N and Kendall G M.
Last reviewed: 10 February 2011