1987 Fiscal Year Final Research Report Summary
Statistical Model for assessing the life-prolonging effect of the health screening service for the A-bomb survivors
| Project/Area Number |
60480469
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
医学一般
|
|---|
| Research Institution | Nagasaki Univcersity. |
|---|
Principal Investigator |
MINE Mariko scientitic Data Center of A-bomb Disasters, Nagasaki University, 医学部, 助手 (00108292)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
NOSE Yoshiaki School of Medicine, Kyusyu University, 医学部, 教授 (20038920)
KONDO Hisayoshi Scientific Data Center of A-bomb Disasters, Nagasaki University, 医学部, 助手 (00170431)
MORI Hiroyuki Scientific Data Center of A-bomb Disasters, Nagasaki University, 医学部, 助手 (30128227)
|
|---|
| Project Period (FY) |
1985 – 1987
|
| Keywords | Health screenings / Life prolonging effect / 比例ハザードモデル |
|---|
| Research Abstract |
In the 1985 fiscal year questionnaires were carried out to find out any underwent the health screening service and those who did not. Analysis of the questionnaires indicated that the percentage of the persons who think they are healthy is higher in those who did not take the health screening service than those who did. In most of the answers for the questions, however, the difference batween the two groups were not significant. In the next fiscal year, we collected medical records and administrative data to complete our database, then performed statistical analysis as described in the followings. First, the average number of health screening taken per year was obtained for each individual. Secondly, the piecewise linear regression method devised by the authors was applied to find relationships between the average number and the years before death. The results revealed that, although the average number tends to decrease as one approaches the year of death, it does not decrease up to tw
… More
o years before diath. Based on the fact, some statistical models with dependent variable being the time of death and independent variables the average number and age were applied to data. Statistical test for goodness of fit selected a time-independent proportional hazard model as one of the best ones to describe the observed data. In the last fiscal year, the life prolonging effect of the health screening service was evaluated assuming the proportional hazard model. The relative risk of the average number 1 to that of 0 (the ratio of their estimated hazared) is about 0.5, regardless of age-categories (40-80) and sex. The relative risk is not statistically significant when cause of death is restricted to cancer, however that for cerebro-vascular disease is highly significant. These results may involve the so-called "self-selection bias", that is, the persons who are more concerned with their.health may undergo health screenings more often. though randomized cohort study seems desirable to resolve the bias, it is practically impossible. Thus the proportional hazard model will play an important role to find out relative risk estimate adjusting for the bias with adequate covariates representing the level of consciousness for health. We are under study to find those covariates among the answers for the questionnaires. Less
|
-
-
-
-
-
[Publications] Mine, M., nakamura, T., Okajima, S., Toyoda, S., Morikawa, A., Oobo, T.: "Study of the consciousness for health of Abomb survivors in Nagasaki" Hiroshima Igaku. 39. 488-490 (1986)
Description
「研究成果報告書概要(欧文)」より
-
-
-
