Chapter 3




<>Ivanov V.K.1, Tsyb A.F.1, Agapov A.M.2, Panfilov A.P.2, Kaidalov O.V.1, Gorski A.I.1,
<> Maksioutov M.A.1, Suspitsin Y.V.3, Vaizer V.I.4  


1 Medical Radiological Research Centre of Russian Academy of Medical Sciences, Obninsk;

2 Department of Nuclear and Radiation Safety, Federal Agency on Atomic Energy, Moscow;

3 Central Medical and Sanitary unit № 8, Obninsk;

4 Department of Radiation Safety and Environmental Protection, Institute of Physics and Power Engineering, Obninsk



The paper discusses the provision of targeted health care to nuclear workers in Russia based on radiation-epidemiological estimates of cancer risks. Cancer incidence rates are analyzed for the workers of the IPPE (the world’s first nuclear installation) who were subjected to individual dosimetric monitoring from 1950 to 2002. The value of the excess relative risk for solid cancers was found to be ERR/Gy=0.22 (95% CI: -4.22; 7.96). The derived ERR/Gy is about half of that derived for the LSS cohort (Hiroshima and Nagasaki).  The difference can possibly be explained by the protracted character of radiation exposure for nuclear industry workers. It has been shown that 81.8% of the persons covered by individual dosimetric monitoring have a potential attributive risk up to 5%, and the risk is more than 10% for 3.7% of the workers. Among the detected cancer cases, 73.5% of the individuals show the attributive risk up to 5% and the risk is in excess of 10% for 3.9% of the workers. Principles for provision of targeted health care, given voluntary health insurance, are outlined.




The 11th Congress of the IRPA was held at the end of May 2004 in Madrid. There were nearly 1500 participants from more than 90 countries. At the conference special attention was given to the new Memorandum of ICRP that highlighted the need to elaborate the existing recommendations and regulations on radiation protection for the public and nuclear industry workers [1]. In the plenary presentation made by Professor Roger Clarke, Chairman of ICRP and presentations by other participants advanded substantial that there is a need to modify the currently used principles of radiation protection.

Let us first turn to one of the key issues of radiological protection - estimation of individual risks. The importance of focusing on this issue is emphasized in the new ICRP Memorandum which, in particular, points to limitations in application of the collective dose. This was certainly our (USSR) experience when we were dealing with the consequences of the Chernobyl accident. The collective dose received after the Chernobyl accident by the populations of the Bryansk, Kaluga, Tula and Oryol regions of Russia is equal to several tens of thousand person-Sv [2]. If the collective dose model (ICRP Publication 60) is used to predict the total number of radiation-induced cancers, with the risk from ICRP publication 60 (5×10-2 pers.Sv-1) several thousand additional radiation-induced cancers should have been expected in these areas by now.   Yet, the National Radiation and Epidemiological Registry data [3] do not reveal any statistically significant excess in cancer incidence rates as compared to the spontaneous level (except for thyroid cancer among children).   If these radiation iduced cancers exist, they are buried in the background of spontaneous cases.   This leads us to make two major conclusions: first, the collective dose (large groups of individuals with low doses) does not provide an adequate projection of radiological health (carcinogenic) effects; and second - for prediction purposes a potential risk group needs to be identified based on an approach using individual dose.[4].

Being aware of all this, ICRP suggests in the new guidelines (due to be published in 2005) that a dose-time matrix be used instead of the collective dose for decision-making and optimization of radiation protection. Knowing how a person moves over a dose-time matrix does provide a basis for estimating individual risk. The transition to the new concept, like analysis of the dose-time matrix for decision making purposes, is suggested to be entrusted with national commissions on radiological protection and executive authorities. In practice, the concept of potential risk group has been used successfully in the UK [5].

What is the principal challenge in implementation of the proposed principles of individual radiological protection? It certainly has to do with the level of knowledge achieved by radiation epidemiology to date. The large-scale radiation-epidemiological studies in the latest 20-30 years (Hiroshima-Nagasaki, registries of nuclear industry workers, Chernobyl, Semipalatinsk, South Urals etc.) made possible the approximate estimate of the magnitude of individual stochastic radiation effects. At the same time, a number of epidemiological issues remain unresolved. We mention only two of them that seem to be of greater importance. First, the models to estimate individual attributive risk proposed by UNSCEAR are based on Hiroshima and Nagasaki data. These epidemiological data continue to play a key role in developing radiation protection standards. At the same time, it is still an open question whether radiation risks associated with acute irradiation (Hiroshima and Nagasaki) are applicable to protracted exposure.   As of now a dose and dose rate effectiveness factor – DDREF – is used and assumed to be equal to 2, which means that protracted exposure leads to half the radiation risk as compared to the risk for the same acute exposure dose.

The second problem is the extrapolation of radiation risks to low doses (to 0.2 Sv). The new ICRP recommendations are based on the linear non-threshold “dose-response” model. Yet, no evidence of statistically significant risk for low doses has been obtained from the radiation epidemiological studies so far conducted.   Indeed it evident that society wishes to protect against risks too small to ever be detected directly.

In order to study the above problems, radiation epidemiological studies were intensified in the last 10-15 years for nuclear workers in the leading countries:  US, UK, Japan, France, and Canada.

In Russia, the Medical Radiological Research Center of Russian Academy of Medical Sciences (Director - Academician of RAMS Tsyb A.F.) and associated National Radiation and Epidemiological Registry (Head - Corresponding Member of RAMS Ivanov V.K.) suggested that workers of nuclear power plants be entered into the system of the departmental medical-dosimetric registry. This proposal was approved by the Scientific and Technical Council № 5 of Minatom of Russia (Head - Academician of RAMS Ilyin L.A.) and endorsed by Rumayantsev A.Yu., Minister of Minatom of Russia. Rosenergoatom (General Director - Saraev O.M.) took a decision that the necessary activities should be started in 2005 and the Department of Radiation Safety be responsible for supervision. In this context, the material below can be regarded as results of preparatory efforts for establishing a registry of nuclear workers.

This paper presents proposals how to tackle three goals of optimization of radiation protection in the nuclear industry in a comprehensive manner (using the IPPE staff as an example):

1. Estimation of radiation risks of cancers in the case of protracted exposure and their comparison with the existing international guidelines.

2. Identifying potential risk groups at the individual level.

3. Development of key principles of providing targeted health care.

It should be noted that the conclusions from radiation epidemiological studies relating to assessment of the risk to nuclear workers at normal operation of the facilities are presented in this paper for the first time.


Materials and methods


Estimation of radiation risks of cancer in case of protracted exposure and their comparison with the existing international recommendations


The cohorts of nuclear workers are of special interest for investigating the relationship between exposure to low-level radiation and cancer incidence. Workers in the nuclear industry are subjected to stringent dosimetric and health monitoring. Direct assessments of radiation risks inferred on the basis of monitoring such cohorts can answer the question whether it is valid to extrapolate health effects from high doses to low doses and protracted exposure.

The IPPE was one of the first nuclear industry installations set up in Russia in the early 50s. Over the operational time period from 1950 to 2002 a total of 5,234 members of the staff (4,284 males and 950 females) were covered by individual dosimetric monitoring (IDM). There are data available for 169 cancer cases (141 males and 28 females). For radiation-epidemiological analysis a follow-up cohort was delineated from all the IPPE workers covered by individual monitoring measurements from 1950 to 2002. The cohort was identified using the following criteria:

·      Time at risk is the time under IDM with allowance for the minimal latent period of 10 years.

·      Members of the cohort are the workers for whom the dates of beginning and end of dosimetric monitoring are available.

·      The attained age during the follow-up period ranges from 20 to 70 years.

·      Only male workers are included.

·      Only solid cancers are considered.

·      The analysis covered those workers covered by individual dosimetric monitoring for a longer time than the latent period of 10 years.

·      To allow for the latent period the dose dynamics were shifted 10 years forward (for example, the dose in 1980 was assigned the value of 1970 and the 2002 dose was considered to be equal to the 1992 dose). This actually means that the analysis in 2002 involved the workers subjected to IDM during 1960-1992.

·      For solid cancer cases a dose was determined based on the date of diagnosis minus the latent period.

A total of 2,320 workers were found to satisfy the above criteria, 102 of them were solid cancer cases.

The cohort’s characteristics are the following:

·      The number of person-years at risk is 40,996.

·      The mean age of the cohort members in 2002 is 56.6 years.

·      The mean cumulative dose for the cohort as a whole was 71.7 mSv and 73.9 mSv for cancer cases.

The mean dose was determined weighted by the time at risk using the formula:


The summation is done over the number of cohort members and time at risk.

The mean time at risk is 17.4 years. For averaging we used the formula:


The structure of the cancer incidence in the cohort under study is shown in Table 1. As can be seen from table 1, the dominant cancers are those of the digestive system, skin, respiratory and urinary systems.


Table 1

Structure of the cancer incidence in the IPPE workers



ICD-10 code

% of total number

Lip, oral cavity and pharynx



Digestive system



Respiratory system



Melanoma and other skin



Mesothelium and soft tissue



Male genital organs



Urinary system



Eye, brain, nervous system



Thyroid and endocrine system





This epidemiological analysis is based on comparison of two groups of subjects:  those  exposed and those unexposed to radiation.    To avoid a systematic bias in the risk estimates the groups are homogenized, to the extent possible, and modified by other confounding factors which may influence the estimate. The confounding factors are age (because cancer incidence tends to increase with age), calendar time (because spontaneous cancer incidence varies with time), sex, social factors and others. The groups are homogenized based on data stratification: groups are divided into subgroups with similar characteristics.

The methodology for estimating the “dose-cancer incidence” relationship in case of protracted exposure differs in important ways  from the approaches applied to one-time acute irradiation. Chronic exposure for nuclear workers increases with time and depends on both time and attained age, and if this is neglected, effects of exposure can be significantly overestimated (risk of cancer increases with age and radiation dose). Therefore, when radiation risks of cancer are assessed,  data should be stratified by time and especially by age (1-2 years intervals). The approach often used for evaluating effects of acute exposure, when groups with high and low doses are compared, is not applicable to chronic exposure, because for chronic risks the low dose group consists primarily of young workers for whom cancer risk is not high, whereas the group with high doses is formed by older personnel with a considerable cancer risk. Disregard of this fact would lead to distortion of the dose-response relationship and an overestimation of risk.

The observed cancer incidence in the study cohort is composed of two parts: the spontaneous incidence in the  unexposed population and those cancers assumed to be radiation-induced. These parts are described by the linear non-threshold relative risk model.  The relative risk is the ratio of the incidence rates in the exposed and the unexposed groups. According to the ICRP and UNSCEAR experts, the relative risk model is preferable for solid tumors.

The model s written as:


where i, j are the strata indices by calendar time and age, k is the dose group index;  is the spontaneous cancer rate in the cohort; ERR1Sv is the excess relative risk per unit dose 1 Sv (the angular factor of the dependence of relative risk on radiation dose); di,j,k is the cumulative dose at time moment i, at age j and in dose group k.

After removing the parentheses in the model the first summand represents the spontaneous incidence and the second represents the radiogenic cancers.

For determination of the contributions of these processes to the observed incidence the statistics package EPICURE (AMFIT module) was used [6]. The stratification by calendar time and attained age is done with an interval of 1 year. All data were divided into three dose groups: 0, 40, 100, 100+ mSv. The model parameters and consequently the spontaneous and radiogenic components of the incidence were inferred by minimizing the deviations of the model values from the actual ones (observed incidence).

The main characteristics of dose groups derived by the AMFIT program are shown in Table 2.


Table 2

Main characteristics of dose groups


Dose group

Mean cumulative dose (mSv)

Observed cases


Expected cases (spontaneous)a






















a The values have been adjusted for age and calendar time.



As can be seen from Table 2, the number of radiogenic cancers among the registered cases (the difference of the observed and expected number of cases) is two (102-100=2).  The data of Table 2 provide a basis for estimating the relative risk and the excess relative risk per unit dose. The relative risk is 102/100=1.02, i.e. 2% and the excess (more than one) relative risk per unit dose 1 Sv is (1.02-1)/0.083 = 0.24 per 1 Sv. This value of excess relative risk is about half the risk associated with acute exposure (the value recommended by UNSCEAR for males is 0.43), which is in conformity (but with a huge uncertainty of the measurement) with the DDERF recommended by UNSCEAR and ICRP for transition from acute exposure to chronic exposure.

For more precise interval estimates of risk the methods outlined in [7] were used. These techniques are based on statistical analysis of the difference between the observed and expected (spontaneous) number of cases. The verification of the zero hypothesis that the relative risk is equal to 1 gave the value p=0.52. Then the relative risk is 1.02 (0.65, 1.66 95% confidence intervals) and the value of excess relative risk per dose 1 Sv is 0.24 (-4.22, 7.96 95% confidence intervals).

It should be stressed that it is the first time that the value of radiation risk of cancers is inferred for routine operations of the nuclear facility in Russia (IPPE). The risk is not statistically significant and half of that associated with acute exposure. Considering the confidence limits of the derived risk estimates, the presently used UNSCEAR methodology for estimating individual attributive risk [8] should be declared well-grounded.


Creating potential risk groups at individual level


For identifying potential risk groups the multiplication model UNSCEAR-94 [8] was used. According to this model, the excess relative risk ERR of cancer incidence depends on age at exposure e and radiation dose d as follows:

,                                                           (1)

with parameters a and b allowing for the sites, as shown in Table 3.




Table 3

Parameters of excess relative risk ERR for solid cancers of different sites



Parameter а, Sv-1

Parameter b, year-1



Respiratory systema




















Mammary gland








All solid cancers





a tracheas, bronchus, lung.



As follows from formula (1), in the case of one-time exposure the excess relative risk is a function of radiation dose and age at exposure. With protracted exposure, the risks from annual exposure are summed up with allowance for dose and age at exposure [9]. Then changes in the individual excess relative risk with age are described by a regular differential equation with the lagging parameter:

,                                                               (2)

where ERR(u) is the total excess relative risk at age u at a given time moment. As the lagging parameter in the equation we use the latent period TL assumed to be 10 years for solid cancers. After replacement of the variable u-TL=e, equation (2) is integrated as follows:

,                                                               (3)

where e0 is the age at which the worker is first exposed to radiation. Thus, integral (3) provides the value of excess relative risk of the worker in the current year. As radiation doses are available up till the present time, formula (3) can be used to project risk for TL years in future when the worker’s age is u + TL years:

.                                                           (4)

Since the mode of exposure or dependence d(e) is specific to each worker, equations (3) and (4) can be integrated only numerically. With introduction of a discrete step of 1 year by age, integrals (3) and (4) are written as follows:

,                                           (5)

,                               (6)

where e0, e, u are the discrete values with 1 year step, De is the radiation dose received at age e.

Hence, the calculation and projection of the individual excess relative risk ERR requires knowing at what age e the worker received dose De and his present-day age u. Given such data are available, the total risk ERR for solid cancers at the current time moment can be determined by formulas (5) and (6).

After the value ERR(u) is found the individual attributive risk AR(u) is calculated by the formula:


with allowance for sex and disease site (see Table 3).

To calculate the individual attributive risk and identify a potential risk group from the IPPE personnel database 1160 male employees working in 2003 were selected, their age varying from 20 to 81 years and the time under IDM - from 1 year to 54 years. The mean accumulated dose for these workers is 82 mSv.

Figure 1 shows the distribution of the IPPE workers subjected to IDM for more than 10 years by the attributive risk intervals (5% intervals) for solid cancers. The calculations allow for the latent period of 10 years. The attributable risk is found to be less than 5% for 618 persons and 28 persons have the risk of 10% or more. Table 4 contains personal data about the studied workers. As can be seen from the table, individual monitoring measurements for these workers were started in the period from 1951 to 1975 and they were aged from 47 to 74 years in 2003. The accumulated dose for them is estimated to be 273 to 1,653 mSv. In the table below 4 workers having risk more than 20% are highlighted by gray.



Fig. 1. Distribution of workers by the attributable risk in 2003
(755 persons, subjected to IDM for more than 10 years, solid cancers).



Since exposure is only associated with elevated risk after the latent period, the radiation doses accumulated by the worker from 1994 to 2003 will not have a full effect until 2013. Thus the attributable risk for currently working personnel can be projected for 10 years. Figure 2 shows the distribution by the attributable risk in 2003 and 2013 for the IPPE workers who were 60 years old and younger in 2003. As follows from the figure, the group with the attributable risk 5% or more is extended from 30 to 52 persons in the course of 10 years.

According to the UNSCEAR-94 risk model the excess relative risk and consequently the attributable risk is dependent on disease site. Let us consider as an example the distribution of workers by the attributable cancer risk for the respiratory system shown in figure 3. While the excess relative risk decreases with age at exposure (see table 3, parameter b is negative), for diseases of the respiratory system the excess relative risk increases with age (parameter b is positive).With exposure at the age of 25 years, the risk is 70% of the risk associated with exposure at 50 years. As can be seen from figure 3, the group with the attributable risk more than 15% includes 20 persons, whereas the same group for all solid cancers consists of 7 persons.



Table 4

Personalized data about the IPPE workers having attributive risk 10% or more


Dosimeter number

Birth year

IDM start year

Accumulated dose, mSv

Attributive risk of solid cancers, %

















































































































































Fig. 2. Distribution by the attributive risk in 2003 (left) and 2013 (right)
for workers aged 60 and younger in 2003 (375 persons).






Fig. 3. Distribution by the attributive risk in 2003 (755 persons,
subjected to IDM more than 10 years, the respiratory system).



In the previous section of the paper 102 solid cancer cases among the IPPE workers with individual dosimetry were discussed. Figure 4 shows the distribution of workers who developed cancer by the attributive risk at time of diagnosis. As can be seen from the figure, 75 persons show the risk less than 5% (75% of all workers with cancer) and 4 persons have risk more than 10% (about 4% of all individuals with the disease). For two persons the attributive risk is more than 40%, their doses being 1.96 Sv (born in 1930 and covered by IDM since 1954, diagnosed cancer of respiratory organs in 1986) and 1.57 Sv (born in 1926 and covered by IDM since 1956, diagnosed skin cancer in 1982).



Fig. 4. Distribution of the workers diagnosed cancer by the attributive risk
at time of diagnosis (102 cases, solid cancers).

When radiation-epidemiological justification is put together, for identifying a potential risk group the threshold value of individual attributive risk needs to be defined: if the threshold is exceeded for a person he is assigned to the group. For example, in the UK compensations for individuals can be considered when the attributive risk is exceeded by 20% [5].


Fig. 5. Cancer incidence rates among males in the central regions of Russia
closest to the IPPE location in 1998 (per 100,000).



In setting a threshold for the potential risk group it is essential to take into account changes in spontaneous cancer rates at the regional level. For example (figure 5), the attributive risk due to variations in the spontaneous incidence rate for 6 regions of Russia around the IPPE location is about 11%. This must be kept in mind when the threshold value of individual attributive risk (for radiation) is established as a basis for making up potential risk groups.


Development of key principles of providing targeted
health care to nuclear workers


The nuclear power industry is an integral part of the energy sector in Russia, and its successful development is now more often linked with improving the effectiveness of occupational health protection for nuclear workers. This mission has to be accomplished in the circumstances when financial and other resources available to the existing health care system are fairly limited. This being so, the occupational health protection in the nuclear industry should be targeted in character and rely on objective criteria for identifying potential risk groups.

Development of the radiation-epidemiological principles and associated recommendations for providing targeted health care, in turn, will be geared towards enhancing the effectiveness of radiological protection from the standpoint of individual risk management.

In order to achieve the set goals it is necessary to improve the epidemiological monitoring, methods and means of individual radiation monitoring and to refine estimation and projection of health effects of radiation (especially possible increased cancer rates).

We are proposing a scheme for interaction between a nuclear industry installation, a hospital providing health monitoring of nuclear workers and an insurance company with the aim to optimize allocation of insurance premium for treatment schemes and in-depth medical checks among members of potential risk groups which could be organized on the base of the National Registry.

Figure 6 shows a flow diagram of the proposed scheme of providing targeted health care to nuclear workers. As can be seen from the figure, the information framework of these activities is formed by data of individual dosimetric monitoring and the unified protocol of cancer data exchange between hospitals conducting health monitoring and the National Registry.



Text Box: ICRP


Fig. 6. Flow diagram of the proposed scheme for provision of  targeted health care to nuclear workers.




Specialists of the National Registry are currently developing a methodology for estimating a probability of causation (death causes) and a methodology for identifying potential radiation risk groups with respect to different diseases. These methodologies will provide a basis for analysis of medical and dosimetric information supplied by nuclear facilities and health monitoring organizations to generate two lists: a) persons with a high probability of causation for a given disease and b) persons subject to in-depth medical checks.

The lists will be forwarded to nuclear industry installations and insurance companies. Findings of the first list with estimates of individual excess relative risk for each detected cancer case will concurrently be passed to interdepartmental expert councils for elucidating the causality. For each member of the second list scientifically justified recommendations are to be prepared to facilitate in-depth health examination and subsequent treatment.

The proposed scheme of providing targeted health care to nuclear workers is devised as the core of the concept of optimization of radiation protection from the standpoint of individual risks management and providing targeted health care.




1. The radiation-epidemiological studies undertaken among the personnel of the IPEE (the follow-up period 1950-2002) has shown that the risk of cancers for this cohort is not statistically significant, and about half the risk calculated directly from the Japanese cohort of Hiroshima and Nagasaki. We thereby confirmed the validity of using the DDREF (DDREF=2) for estimating carcinogenic effects of protracted occupational exposure as compared to acute irradiation.

2. Individual attributable risk was estimated for the IPPE workers covered by individual dosimetric monitoring. It has been found that 81.8% of workers have the individual attributive risk of cancers up to 5% and the risk is more than 10% for 3.7% of the personnel. Among the detected cancer cases 73.5% of persons had the attributive risk up to 5% and for 3.9% the risk was more than 10%.

3. The analysis of cancer incidence among males living in 6 regions of Russia neighboring the IPEE site in 1998 shows that the attributable risk due to variations in spontaneous incidence at the regional level can be as high as 11%. This estimate may play an important role in defining the threshold value of individual risk associated with radiation when identifying the potential risk group.

4. Principles and procedures are proposed for radiation-epidemiological justification of providing targeted health care to nuclear workers occupationally exposed to radiation.

        5. The performed studies provide a basis for setting up a sectorial medical-dosimetric registry of personnel of nuclear power plants. Putting the developed radiation-epidemiological principles to practice will ensure that the nuclear industry be better prepared for implementation of the new ICRP recommendations on optimization of the radiation protection system. 



[1]      Clark R. Memorandum. Evolution of the radiation protection system: justification of the need to develop new ICRP recommendations  Medical Radiology and Radiation Safety. - 2003. - V. 4b. - P. 26-38.

[2]      UNSCEAR 2000 report. Appendix J.  United Nations  2000 .

[3]      Ivanov V.K., Tsyb A.F., Maksioutov M.A., Gorski A.I., Vlasov O.K., Biryukov A.P., Kaidalov O.V., Matveenko E.G., Khait S.E., Kruglova Z.G., Kochergina E.V. Radiological Health Effects of the Chernobyl Accident for the Population of Russia: Estimation of Radiation Risks. - Moscow: Medicine, 2002.

[4]      Ivanov V.K., Tsyb A.F., Agapov A.M., Ivanov S.I., Panfilov A.P., Kaidalov O.V., Gorski A.I., Maksioutov M.A., Suspitsin Yu.V., Vaiser V.I. Problems of objective ascertainment of occupational cancer incidence in the industry. Bulletin of Nuclear Energy. - 2003. - No 5. - P. 37-44.

[5]      Wakeford R., Antell B., Leigh W. A review of probability of causation and its use in a compensation scheme for nuclear industry workers in the United Kingdom.  Health Phys. - 1998. - V. 74, No 1. - P. 1-9.

[6]      Preston D.L., Lubin J.H., Pierce D.A., McConney M.E. EPICURE. - Seattle: Hirosoft International Corporation, 1993.

[7]      Breslow N.E., Day N.E. Statistical Methods in Cancer Research. IARC Scientific Publication No. 82. - Lyon: IARC, 1987. - P. 91-96.

[8]      United Nations Scientific Committee on the Effects of Atomic Radiation. Sources, Effects and Risks of Ionizing Radiation. UNSCEAR 1994 report to the General Assembly. - New York: United Nations, 1994.

[9]      IAEA, Methods for Estimating the Probability of Cancer from Occupational Radiation Exposure. IAEA-TECDOC-870. - Vienna: IAEA, 1996. - P. 55.

Chapter 1

Chapter 2

Radiation and Risk Main page
Richard Wilson Main Page