The healthy worker effect : does socioeconomic status influence cause-specific mortality among retirees at Department of Energy facilities in Oak Ridge, Tennessee?

Author

Joseph L. Weaver

Date of Award

12-1990

Degree Type

Dissertation

Degree Name

Doctor of Education

Major

Health Promotion and Health Education

Major Professor

Bill Wallace

Committee Members

Bob Kirk, Jack Pursley, Judy Fiene

Abstract

Many epidemiologic studies have shown that mortality rates of occupational groups are usually lower than those of the general population. This favorable health status has been termed the "healthy worker effect" The primary purpose of this retrospective cohort mortality study was to examine the magnitude of the healthy worker effect between white and blue collar populations and to attempt to determine to what extent this phenomenon effectively continued following retirement The mortality experience of 4,444 retirees (4,057 white males and 387 white females) of the Oak Ridge National Laboratory, the Oak Ridge Gaseous Diffusion Plant, and the Y-12 Plant (all located in Oak Ridge, Tennessee) was evaluated for the years 1960 through 1979. Hiis study emphasized the mortality experience of the salaried and hourly retirees who were employed at least 10 years and who lived a minimum of 30 days beyond their retirement dates. Historical follow-up was conducted for the years 1960 to 1984, and ascertainment of vital status was achieved for 99.6% of the male cohort and 99.9% of the female cohort. There were 43,807.78 person-years of observation. Standardized mortality ratios were computed to contrast the subjects' cause-specific mortality experience with that of a standardized population within the study cohort and with that of the general U.S. population. In the total three-plant population there were 1,620 male deaths and 73 female deaths observed for the all causes category. These deaths comprised 106% and 84%, respectively, of the total numbers of deaths expected. Several distinct phases were used to examine various subgroups within the study cohort The initial phase examined the overall death rate among white salaried males across the three Oak Ridge facilities by comparing it to the U.S. average. Overall mortality among salaried male retirees was found to be essentially as expected for all causes at the Oak Ridge Gaseous Diffusion Plant (SMR = 1.01) and at the Y-12 Plant (SMR = 1.00). However, the mortality rate at the Oak Ridge National Laboratory was less than expected (SMR = 0.90). In the next phase the white collar cohort was combined and compared to a similar segment of the U.S. population. For the all causes category the mortality experience remained favorable in this combined white collar group (SMR = 0.97). Death rates from cause-specific mortality among the white coUar population was also found to be generally better than that of the U.S. This finding supported the postulation that the healthy worker effect continues after an individual retires. Socioeconomic status was implicated as a primary contributor in the good health characteristics found in the salaried male cohort The next phase of this study examined the overall mortality experience among white hourly males across the three facilities by comparing it to the general U.S. population. Overall mortality among blue collar retirees was slightly higher than expected in the Oak Ridge National Laboratory (SMR = 1.09) and the Y-12 Plant (SMR = 1.06) populations. The increase in mortality at these two sites was found not to be significant. However, a statistically significant increase in mortality from all causes combined was found in the Oak Ridge Gaseous Diffusion Plant (SMR = 1.17) hourly male population. In the next phase the blue collar population was combined and compared to the U.S. The death rate for all causes in the combined group was found to be significantly elevated. A significantly elevated mortality rate was found in this population for the all cancers category. There was no evidence of carryover in the healthy worker effect in the blue collar cohort. Several causal factors (i.e., socioeconomic status, lifestyle, and occupational hazards) were identified as major contributors for these findings. The last phase of the male investigation directly compared the combined white collar male group to the combined blue collar male group. This comparison showed statistically significant increases in the mortality rates in the all causes category (SMR = 1.24). Two cause-specific categories-all cancers (SMR = 1.45) and circulatory diseases (SMR = 1.23)- were found to be significantly elevated in the blue collar population when directly compared to the white collar population. Relative to the U.S. white female population, both the combined salaried and combined hourly female cohort exhibited few^r deaths in many categories of diseases. Specifically, fewer deaths were recorded in the all causes (salaried SMR = 0.87; hourly SMR = 0.76), all cancers (salaried SMR = 0.91; hourly SMR = 0.93), and the diseases of the circulatory system (salaried SMR = 0.70; hourly SMR = 0.61) categories. These results were indicative of the healthy worker effect, and the low«r-than-expected standard mortality ratios observed in the female cohort were probably due to the selection of healthy workers at the first date of hire. The low mortality ratios could also be attributed to neither the hourly nor the salaried females having socioeconomic dominance over the other (but generally better than the U.S. average for those groups). Further evidence of a healthy worker effect was found in the white and blue collar female population when they were compared to one another. In fact, virtually no differences in overall death rates were found between white and blue collar females when they were compared to each other for all causes (SMR = 1.08), all cancers (SMR = 1.11), and diseases of the circulatory system (SMR = 1.08). Although no evidence was provided to support the postulation, it is suspected that a common link for the congruence in death rates between the two female populations may be attributed to their lifestyle similarities during the study period. In this study mortality appeared to be highly associated with socioeconomic status. In other words, multi-factorial variables (i.e., health-related behavior, diet, and potential exposure to radiation, organic chemicals, metals, or other substances) may have come together over time, to cause adverse mortality outcome in certain subgroups within the study population. The persistently high standard mortality ratios observed in the hourly male retirees suggest that socioeconomic status or related factors had a profound effect on health status over time. Conversely, the persistently low ratios observed throughout the follow-up period in the salaried male population and the entire female population suggest that the healthy worker effect does not diminish in retirement years. This study also demonstrated that males (both white and blue collar) who retired before age 62 were at an elevated risk of mortality from all causes. The implication made relative to this finding is that a large proportion of male retirees may be leaving the workforce because of health reasons. The mortality experiences found in this study population underscore the fact that further follow-up is needed in order to examine the independent and combined effects of potential work-related exposures and lifestyle behaviors.

Recommended Citation

Weaver, Joseph L., "The healthy worker effect : does socioeconomic status influence cause-specific mortality among retirees at Department of Energy facilities in Oak Ridge, Tennessee?. " PhD diss., University of Tennessee, 1990.
https://trace.tennessee.edu/utk_graddiss/11524