The objectives involve several aspects of statistical analysis of excess cancer mortality among the atomic bomb survivors, continuing the work of the investigator for several years in this area. Data on the follow-up of a large population of these survivors will be provided by the Radiation Effects Research Foundation, Hiroshima; these are the primary source of quantitative information regarding radiogenic cancer risks in humans. Interpretation of these data is remarkably complex and challenging, and currently open are a number of fundamental questions which must be clarified as the follow-up approaches near completion in the next few years. This requires both use of the most up-to-date methods of epidemiological cohort analysis, and familiarity with current radiobiological knowledge in the area. Results of a nature to appear in the radiation effects literature, rather than those on statistical methods, are the primary goal.
The specific aims i nclude analysis of: (i) variation in excess cancer risks with age and time-since-exposure, (ii) possible biases due to the selection of the study population by survival of the bombing, (iii) biases due to imprecision in estimation of individual radiation exposures, (iv) uncertainties in projecting beyond the current follow-up to estimate lifetime risks, and (v) combining information from these data and from experimental radiobiology, regarding the shape of the dose-response curve, in order to estimate the risks from lower exposures. Several statistical models, with different causal interpretations, can be used to fit adequately the joint effects of age-at-exposure, gender, time- since exposure, and age-at-risk. Comparison of these will be made in terms of their interpretation and differential effects on projection of lifetime risks. Statistical models for the effects of cohort selection will be formulated, in order to obtain a quantitative understanding of resulting biases in estimating risks for a more general population at low exposures. Alternative methods for projection beyond the current follow-up will be analyzed, with the aim of better understanding the uncertainties involved in this extrapolation. The extent of information available from these data regarding the shape of the dose-response will be quantified, for comparison to and combination with other radiobiological information.
| Lubin, J H; Boice Jr, J D; Edling, C et al. (1995) Lung cancer in radon-exposed miners and estimation of risk from indoor exposure. J Natl Cancer Inst 87:817-27 |
