Skip to main content
NCBI home page
HHS Author Manuscripts logoLink to HHS Author Manuscripts
. Author manuscript; available in PMC: 2025 Oct 1.
Published in final edited form as: Am J Ind Med. 2024 Jul 31;67(10):901–909. doi: 10.1002/ajim.23642

Association between longest-held occupation and mortality risk

Abay Asfaw 1, Anasua Bhattacharya 2
PMCID: PMC11486569  NIHMSID: NIHMS2014164  PMID: 39086145

Abstract

Background:

Occupation is associated with a large part of daily activities, affecting lifestyle and social status. However, limited research exists on the association between longest-held occupation (LHO) and early mortality. We examine if LHO is associated with mortality risk among US adults 51 years of age and older.

Methods:

Using Health and Retirement Study data from 1992 to 2020, we followed 26,758 respondents 51 years of age and older for up to 29 years. We used competing-risks analysis methodology to estimate the risk of mortality.

Results:

Across the average 20.5 follow-up years, women with LHO in the categories of machine operators (subhazard ratio [SHR]: 1.42), food preparation (SHR: 1.39), handlers and helpers (SHR: 1.35), and sales (SHR: 1.15), were more likely to die earlier than women with the LHO in the professional and technical support occupation, the reference occupation. Men with LHO in the categories of food preparation (SHR: 1.43), machine operators (SHR: 1.36), personal services (SHR: 1.34), handlers and helpers (SHR: 1.32), protective services (SHR: 1.31), clerical (SHR: 1.27), farming and fishing (SHR: 1.26), sales (SHR: 1.23), and precision production (SHR: 1.20) had elevated risks of mortality compared to men whose LHO was in the referent professional and technical support occupation.

Conclusions:

Findings from this study provide comprehensive and current evidence that occupation can be one of the risk factors for adverse health outcomes and ultimately for early mortality.

Keywords: competing-risks analysis, early mortality, HRS data, longest-held occupation

1 ∣. INTRODUCTION

Occupation is associated with a large part of daily activities, affecting lifestyle and social status. Different occupations result in different types of workplace exposures, and these exposures accumulate over time, affecting the health and survival of workers.1 Occupation-specific risk factors may have different health effects, and some of these effects may only be recognized long after worker exposure. Not much is known about the association between longest-held occupation (LHO) and early mortality. Some evidence of the association between occupation and mortality is evident in US workers in manual jobs, for example, workers in manual jobs have shorter life expectancy than workers in managerial, professional, and executive occupations.2 Workers in Britain who are in low-grade occupations, such as messengers and other unskilled manual laborers, also have poorer self-reported health, higher incidence of cardiovascular and respiratory disease, and worse health behaviors than workers in high-grade occupations such as in the executive, administrative, and professional fields.3

The ongoing study of the connections between occupation and health spans many specific workplace factors and health effects. Researchers have identified association between occupation and diseases such as pneumoconiosis, cardiovascular disease, asthma, cognitive impairment, and chronic renal disease.4-6 Low wages and lack of access to fringe benefits, which include access to employer-sponsored health insurance, are considered to be occupation-specific risk factors that can affect health.7 Some occupations have also been associated with health behaviors that may adversely affect health, such as smoking,8,9 being physically inactive,8 drinking alcohol,8 or eating unhealthy foods.8 These behaviors and their effects may persist even after retirement.10 Occupation can also affect health and consequently, mortality through exposure to specific biomechanical,11,12 chemical,13 and physical12,13 factors. Work-related psychosocial factors also play a role in health outcomes; these factors include prestige and job satisfaction14 and job demand and control.11 According to the International Labor Organization, occupation can also expose workers to hazardous substances that cause fatal diseases (i.e., cancers and poisonings) or fatal injuries (i.e., fires and explosions).15

The evidence continues to build on specific risk factors related to occupation and the complex relationships among these factors and health. However, limited research exists on the association between LHO and early mortality. Using the Health and Retirement Study, Asfaw et al.16 examined the association between LHO and receiving disability insurance benefits in the United States. The results showed that the hazard of receiving disability insurance benefits was highest among workers with LHOs in personal services, protective services, mechanics and repair, and sales. Li et al.17 also examined the association between LHO in a lifetime and the risk of disability in activities of daily living (ADL) among elderly people (65 years and older) in northern Taiwan. Compared with people who were former legislators, government administrators, or business executives and managers, workers in agriculture, animal husbandry, forestry, or fishing and workers in craft and related trades had significantly increased risks of subsequent ADL disability. The major objective of this study was to improve our understanding of this association by using longitudinal data. We used data collected from a nationally representative, US population of 51 years of age and older from 1992 through 2020. We measured occupation by the LHO to capture the cumulative effects of occupation on early mortality.18 The study examined the association of 16 broad LHO categories (excluding military) presented in Supporting Information S1: Appendix 1 with the hazard of early mortality. Studies like this will shed light on the question of whether LHO is a predictor of early mortality risk. To our knowledge, no recent study has examined the association between LHO and early mortality using a nationally representative longitudinal data set in the United States.

2 ∣. STUDY DATA AND METHODS

2.1 ∣. Data and measurement of variables

The data source for this study was the Health and Retirement Study (HRS). The HRS collects longitudinal data that is nationally representative of noninstitutionalized Americans aged 51 years or older. Sonnega et al.19 report the sampling procedure and other detailed information on the data. This data set is publicly available at https://hrs.isr.umich.edu/. Since the launch of the HRS in 1992, the original sample of respondents have been interviewed every 2 years, and new cohorts are added continuously. Currently, the HRS is one of the most comprehensive and nationally representative longitudinal data sets for Americans 51 years of age and older.19 In this study, we used the 2020 RAND HRS longitudinal file that allows comparability across survey waves. RAND provides more information about the 2020 file used in this study.20 The HRS collects detailed demographic, socioeconomic, and health information about respondents and their spouses, and it provides corresponding population weights.

The dependent variable was death. The HRS collects mortality information including month and year of death from the respondent's proxy (surviving spouses, relatives, and other informants) using exit and follow-up interviews. It also matches the death information to the National Death Index database (Data Access—National Death Index [cdc.gov]).

LHO was our independent variable of interest. The HRS collects information on current and LHO using the 1980, 2000, and 2010 Census Occupation Classifications. To measure LHO consistently, we mapped the 2000 and 2010 occupational codes to the 1980 occupational codes using the BLS occupation crosswalks available at Industry and Occupation Code Lists & Crosswalks (census.gov). The 1980 occupational code has 17 major occupations. We considered 16 major occupations (excluding military occupation) presented in Table 1. Professional and technical support occupation (henceforth “professional” for brevity) was used as a reference LHO because it is the largest and a relatively safe category in terms of mortality risk. Supporting Information S1: Appendix 1 lists the detailed occupations included in the 16 major LHOs.

TABLE 1.

Distribution of socioeconomic and demographic variables associated with early mortality among Health and Retirement Study respondents aged 51 years and older (1992–2020) for the entire population and by sex.

Variable Total Women Men
Number of respondents (resp.) 36,445 19,836 16,609
Weighted sample (million) 114.2 57.5 56.7
No. of resp. with missing longest-held occupation 9687 5856 3,831
No. of resp. considered in the main analysis 26,758 13,980 12,778
No. of resp. died 9640 4319 5321
No. of resp. requested to be removed from HRS 5899 3140 2759
Median age in years 63 63 62
Race or ethnicity (column %)
 Non-Hispanic White 17,650 (76.3) 9042 (76.1) 8608 (76.5)
 Non-Hispanic Black 5067 (11.0) 2900 (11.8) 2167 (10.1)
 Hispanic 3127 (8.5) 1576 (8.1) 1551 (8.9)
 Non-Hispanic others 914 (4.2) 462 (3.9) 452 (5.0)
Marital status (column %)
 Married/partner 25,210 (73.9) 11,939 (68.1) 13,271 (79.8)
 Separated/divorced/widowed 9414 (20.6) 6859 (26.4) 2555 (14.6)
 Never married 1821 (5.5) 1038 (5.5) 783 (5.6)
Education (weighted column %)
 No college 16,640 (55.7) 8850 (56.9) 7790 (54.4)
 College (%) 10,118 (44.3) 5130 (43.1) 4988 (45.6)
BMI (column %)
 Normal or underweight 19,257 (71.1) 9767 (69.8) 9490 (72.3)
 Obese (BMI > 30) (%) 7501 (28.9) 4213 (30.2) 3288 (27.7)
Smoking status (column %)
 Never smoked 10,916 (42.2) 6862 (48.7) 4054 (35.6)
 Ever smoked 15,842 (57.8) 7118 (51.3) 8724 (64.4)
Alcohol consumption (column %)
 Never drink alcohol 9958 (32.9) 5943 (37.6) 4015 (28.1)
 Ever drink alcohol 16,800 (67.1) 8037 (62.4) 8763 (71.9)
Longest-held occupation (column %)a
 1. Managerial [003–037] 3245 (13.5) 1335 (10.7) 1910 (16.2)
 2. Professional and technical support [043–235] 4238 (18.0) 2444 (19.8) 1794 (16.1)
 3. Sales [243–285] 2459 (9.5) 1402 (10.1) 1057 (8.8)
 4. Clerical and administrative support [303–389] 4052 (14.9) 3362 (24.4) 690 (5.2)
 5. Private household service [403–407] 796 (2.3) 532 (3.0) 264 (1.6)
 6. Protective service [413–427] 469 (1.7) 128 (0.9) 341 (2.5)
 7. Food preparation service [433–444] 982 (3.3) 777 (5.0) 205 (1.5)
 8. Health services [445–447] 1039 (4.1) 915 (6.8) 124 (1.3)
 9. Personal services [448–469] 1690 (5.3) 1253 (7.7) 437 (2.9)
 10. Farming, forestry, and fishing [473–499] 666 (2.1) 148 (0.9) 518 (3.3)
 11. Mechanics and repair [503–549] 1002 (4.1) 41 (0.3) 961 (8.0)
 12. Construction trade and extractives [553–617] 1205 (4.7) 46 (0.4) 1159 (9.0)
 13. Precision production [633–699] 1376 (5.2) 453 (3.0) 923 (7.4)
 14. Machine operators [703–799] 1550 (4.6) 774 (4.4) 776 (4.7)
 15. Transportation [803–859] 1418 (5.1) 209 (1.4) 1209 (8.8)
 16. Handlers and helpers [863–889] 571 (1.9) 161 (1.1) 410 (2.7)
Region (column %)
 Northeast 4439 (18.1) 2373 (18.8) 2066 (17.4)
 Midwest 6144 (24.8) 3168 (24.5) 2976 (25.0)
 South 10,873 (36.2) 5701 (36.0) 5172 (36.3)
 West 5292 (21.0) 2731 (20.8) 2561 (21.2)
Open in a new tab

Note: Number of observations are unweighted while percentage values are weighted.

Abbreviation: BMI, body mass index.

a

Numbers in [] are 1980 Occupation Codes.

We included the following covariates in the study that can affect early mortality: race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, and non-Hispanic Other); marital status (married/had a partner, separated, and never married); college education; obesity (body mass index [BMI] ≥ 30); smoked ever; drank alcohol ever; and region (Northeast, Midwest, South, West). Except for race/ethnicity, these covariates changed values through time, and we used the relevant values of the variables over time (see Statistical analysis section below). Age is recommended to be a time scale (time from birth to until death, until requested by the respondent to be permanently removed from the HRS sample, or until the end of the survey year [2020]) instead of a covariate for survival analysis of the elderly population where baseline age is the time origin.21,22 This is particularly important when age is highly related with the outcome variable.22 This study used age as a time scale. Income is not included as one of the covariates because it is one of the major factors in the causal pathway between occupation and mortality, and including it as a covariate would take out part of the influence of occupation on mortality.

The outcome variable was early mortality. The follow-up period was from the age of entry into the HRS survey until death, until requested by the respondent to be permanently removed from the HRS sample, or until the end of the survey year (2020), whichever came first. No Institutional Review Board approval was necessary because the data were publicly available and individual respondents were not identifiable with the data.

2.2 ∣. Statistical analysis

In our sample, 22% of the respondents requested to be permanently removed from the HRS sample. This means that these respondents were no longer at risk of death from the date they asked to be removed up to the end of the follow-up period, 2020. Using a standard survival model would assume any competing risk such as request to be permanently removed from a survey as censored for mortality, and this might produce biased hazard ratios, especially if the competing event is frequent or occurs early.23 An alternative approach is to use a competing-risks model suggested by Fine and Gray24 that estimates the risk of early mortality accounting for the permanent removal from the HRS sample. There was also a significant difference in the association between early mortality rate and LHO by sex. Therefore, we estimated the model for women and men separately. Because marital status, alcohol consumption, obesity, and region change through time, we considered these variables as time-varying in our estimation. We estimated subhazard ratios (SHRs) and 95% confidence intervals for each LHO and other covariates. In all our estimations, we used the corresponding population weights.

Around 27% of respondents in our sample had missing values for LHO. First, we estimated the model by excluding these respondents. Second, we tested if excluding these respondents might lead to a potential bias in our subhazard ratio estimates using a multiple imputation25 method. This method replaces missing values for LHO with reasonable multiple values based on a selected model.26 We imputed LHO using chained equations and a multinomial logistic model. Sex, age, race/ethnicity, marital status, college education, obesity, alcohol consumption, smoking status, annual family income, and region were used to impute five unique values for each missed LHO. We also included the outcome variable, mortality, and the Nelson-Aalen estimate of the baseline cumulative hazard in our imputation model.26,27 The validity of the imputation outcomes was evaluated by comparing their distributions with actual and combined (actual and imputed values combined) values. We used Stata® 17 for all estimations.

3 ∣. RESULTS

Overall, the cohort included 36,445 respondents aged 51 years or older with work experience. These respondents were followed for an average of 20.5 years (minimum 1 and maximum 29). Out of these respondents, 9687 (27%) with missing LHOs were not considered in the initial analysis. Overall, 26,758 respondents with a total of 189,941 observations were considered. These respondents represented 114.2 million people aged 51 years or older in the country. See Table 1 for the details.

Table 1 presents the descriptive statistics of socioeconomic and demographic variables associated with early mortality for the entire population and by sex. The median age of respondents was 63 years. Seventy-six percent of them were non-Hispanic White, 11% were non-Hispanic Black, and 9% were Hispanic. Nearly 74% of the respondents were married, 21% were separated/divorced/widowed, and 6% were never married. Nearly 29% were obese; 42% never smoked; 33% never consumed alcohol, and 44% had college education. Most of the respondents were from the South (36%), followed by the Midwest (25%) and West (21%).

The LHO categories also appear in Table 1. The LHO category containing the highest percentage of respondents was professional (18%), followed by clerical and administrative support (15%), managerial (14%), and sales (10%). The table also shows significant differences in most of the variables relating to LHO between women and men. While the share of men with LHO in construction, mechanics and repair, transportation, and managerial occupations was higher than women, the share of women with LHOs in clerical and administrative support, health services, personal services, and food preparation was higher than men.

Of the 26,758 respondents in our sample, 9640 (36%) died during the 29-year follow-up period, 5899 (22%) requested to be permanently removed from the HRS, and 11,219 (42%) were alive. Supporting Information S1: Appendix 2 presents the unadjusted survival curve for women and men. Overall, after the age of 55, the probability of survival was lower for men than women.

Table 2 presents the competing-risks regression results by sex, separately for women and men. After adjusting for time-invariant and time-variant variables, the subhazard ratio of early mortality was 42%, 39%, and 35% higher for women whose LHO was in the categories of machine operator, food preparation, and handlers and helpers, respectively than for women whose LHO was in the reference occupation. The subhazard ratio of early mortality was also 16% and 15% higher for women with the LHO in personal services and sales, respectively, than women with LHO in the reference occupation.

TABLE 2.

Competing-risks regression results by sex.

Subhazard ratio
Variable Women Men
Longest-held occupation
 Professional and technical support (Ref.)
 Managerial 1.09 1.10
 Sales 1.15* 1.23**
 Clerical 1.04 1.27**
 Private household service 0.85 0.67
 Protective service 1.14 1.31**
 Food preparation 1.39** 1.43**
 Health services 1.11 0.74
 Personal services 1.16* 1.34**
 Farming and fishing 0.89 1.26**
 Mechanics and repair 0.96 1.02
 Construction and extractives 0.83 1.09
 Precision production 0.96 1.20**
 Machine operators 1.42** 1.36**
 Transportation 0.71 1.13
 Handlers and helpers 1.35* 1.32**
Race: Non-Hispanic White (Ref.)
 Hispanic 0.58** 0.68**
 Non-Hispanic Black 1.06 0.97
 Non-Hispanic Others 0.66** 0.77**
Time-varying covariates
Education: No college degree (Ref)
 College 1.00** 0.99**
Body mass index: normal or underweight (Ref)
 Obese 0.99** 0.99**
Alcohol consumption: no alcohol ever (Ref)
Drink alcohol ever 0.98** 0.99**
Marital status: Married (Ref.)
 Separated/divorced/widowed 1.00 1.00**
 Never married 1.00** 1.01
Smoking status: never smoked (Ref.)
 Smoked ever 1.01** 1.00**
Region: South (Ref.)
 Northeast 1.00 1.00*
 Midwest 1.00 1.00***
 West 1.00 1.00**
Number of obs. 103,488 86,453
Number of subjects 13,980 12,778
Number faileda 4319 5321
Number competingb 3140 2759
Number censoredc 6521 4698
Open in a new tab
a

Number of respondents died.

b

Number of respondents requested to be removed from Health and Retirement Study.

c

Number of respondents who were alive at the end of the follow-up period.

*

p < 0.10

**

p < 0.01

***

p < 0.05.

Controlling for covariates, the subhazard ratio of early mortality for men with LHO in the food preparation, machine operators, personal services, and handlers and helpers categories was 43%, 36%, 34%, and 32%, respectively, higher than men with the LHO in the reference occupation. The subhazard ratio of mortality was also 31%, 27%, 26%, 23%, and 20% higher for men with LHO in protective service, clerical, farming and fishing, sales, and precision production occupations, respectively, than men with LHO in the reference occupation.

The results also showed that there was no significant difference in the risk of early mortality for respondents with LHO in private household service, managerial, health services, and mechanics and repair occupations for both sexes compared with respondents with LHO in the reference occupation. For both women and men, the subhazard ratio of early mortality was lower from Hispanic and non-Hispanic other ethnic/racial respondents compared to White respondents. Most of the time-varying variables such as marital, drinking, and smoking status, and obesity were statistically significant, probably due to large sample size, but the SHR were very close to one. This implies that these variables have a statistically significant association with the outcome variable, but their effects were very small and may be of little clinical or practical significance. The SHR for drinking and smoking might also be related to the variable used to measure alcohol consumption and smoking status as a self-reported dichotomous variable rather than an ordinal scale of reporting alcohol consumption and smoking level.

As indicated above, 27% of respondents did not report their LHO. We conducted a sensitivity analysis to see if the results were robust or insensitive to missing LHOs. This analysis imputed missing LHOs and estimated a multiple imputation competing-risk analysis. As indicated in Supporting Information S1: Appendix 3, the multiple imputation procedure gave reliable estimates as indicated by similar distribution among the observed (actual), imputed, and combined (actual plus imputed) values. The subhazard ratio and 95% confidence interval from competing-risks multiple imputation results are presented in Supporting Information S1: Appendix 4 alongside the main results. Most of the results were similar in magnitude and statistical significance. However, for women with LHOs of personal services and handlers and helpers and for males in the LHO precision production, the subhazard ratios became insignificant after the inclusion of the imputed LHO values.

4 ∣. DISCUSSION

According to the World Health Organization and the International Labour Organization (WHO/ILO), work-related diseases and injuries were responsible for the deaths of 1.9 million people in 2016.28 This study examined if LHO is one of the risk factors for early mortality, seeking to provide insight into the long-term consequences of occupation on mortality.

Using data from a nationally representative sample of people 51 years old and older in the United States, this study demonstrated that respondents whose longest employment was in certain occupations faced elevated risks of early mortality and that these risks varied by sex. We regressed the risk of early mortality on LHO, controlling for a wide range of time-varying and time-invariant covariates separately for women and men. Our findings showed a statistically significant association between LHO and the risk of dying at an earlier age for some occupations. Overall, men with LHOs in the categories of food preparation, machine operations, personal services, handlers and helpers, protective services, clerical, farming and fishing, sales, and precision production had a higher risk of dying at an early age than men with an LHO in the reference occupation. Similarly, women with LHOs in the categories of machine operations, food preparation, handlers and helpers, personal services, and sales were more likely to die at an earlier age than women with an LHO in the reference occupation. We did not find a statistically significant elevated risk of early mortality for other occupations compared to the reference occupation.

A 27-year follow-up study of longitudinal data from a Dutch population showed that workers in nonskilled general, technical, and transport occupations had an up to 3.5-year shorter life expectancy than workers in academic professions, accounting for different covariates.29 Lee et al.,30 using data from a Korean national employment insurance program between 1995 and 2000, reported differences in mortality rates across the occupational spectrum. The authors found that the age-adjusted overall mortality rates were highest among males in service and sale work (563 per 100,000) and in elementary occupations (street cleaning, carrying luggage, door-keeping, and monitoring and packing manufactured products by hand) (499 per 100,000). The rates were lowest among women in higher occupational categories such as legislators, senior officials and managers, and professionals (93 per 100,000). Using US data from the 2020 National Vital Statistics System mortality file, Billock et al.31 showed that age-adjusted COVID-19 mortality per 100,000 workers was high for workers in protective services, food preparation and serving, transportation and material moving, and farming, fishing, and forestry compared to other occupations such as architecture and engineering. Our study portrays similar results. Men with LHO in food preparation, machine operation, personal services, handlers and helpers, and protective service and women with LHO in machine operation and food preparation were more than 31% more likely to die early than men and women with LHOs in professional.

Occupation may have positive and negative impacts on the health status of workers, and consequently on early mortality. The literature identifies four occupation-related factors that might affect mortality. First, occupation is identified as one of the factors affecting the incidence of occupational injury and illness.32 For instance, the US Bureau of Labor Statistics showed that the incidence rates for nonfatal occupational injuries and illnesses involving days away from work per 10,000 full-time workers in 2020 were 154, 103, 102, and 98 for workers in farming and fishing, protective services, personal services, and food preparation, compared with 11.8 for workers in professional occupations.33 Several recent studies have shown that workers with lost-time injuries had an excess risk of mortality compared to workers with medical-only injuries.34,35

Second, occupational prestige is strongly related to self-esteem, positive social interaction, job satisfaction, and worthiness, and these factors may affect overall health status and mortality.6,36,37 Using a sample of employed respondents from the National Health Interview Survey (pooled over the years 1986 to 1994), prestige scores from the 1989 General Social Survey (GSS), Christ et al.38 showed that the risk of mortality was lower for respondents in high prestige occupations than for respondents in low prestige occupations, after controlling for income and education. Using the 1989 GSS occupational prestige scores and GSS data in 2002 and 2006, Fujishiro et al.14 also showed an inverse relationship between prestige score and self-reported health status. Our results also support these findings. Except for protective service occupations, the average occupational prestige index for occupations with elevated risk of early mortality was around half of the reference professional occupations. For instance, the average prestige scores for food preparation, machine operation, and personal services were 24%, 33%, and 25%, respectively, compared with 64% for the professional occupation. See Supporting Information S1: Appendix 5 for more details.

Third, occupation has been associated with behaviors that may adversely affect health such as smoking,8,9,39 being physically inactive,8 drinking,8 or eating unhealthy foods8 and that may eventually increase the risk of early mortality. We found that respondents with LHO in high early mortality risk occupations tended to have unhealthy behaviors, except for alcohol consumption. The percentage of current or former smokers with LHOs in mechanics and repair, handlers and helpers, precision production, and protective service were 70%, 69%, 68%, and 67%, respectively, compared with 49% of workers with LHO in the professional occupation. The percentage of obese respondents with LHOs in protective services, food preparation, and mechanics and repair were 40%, 35%, and 34%, respectively, compared with 28% of respondents with professionals as their LHO. See Supporting Information S1: Appendix 6 for the details.

Fourth, there is extensive evidence on the relationship among income, health, and mortality at both macro and micro levels.40-42 This could be partly because individuals with higher income can invest more in health and safety,11 or healthy people are more likely to have higher incomes.2 Our data showed that the percentage of respondents who lived below the poverty line was between 13% and 23% for those with LHOs in food preparation, farming and fishing, handlers and helpers, personal services, and machine operation compared with 3% for those with LHO in the professional occupations.

Some limitations of this study should be considered when interpreting the results. First, inclusion of behavioral variables such as moking, obesity, and alcohol consumption as covariates might underestimate the association between LHO and early mortality because occupation is also associated with such behaviors that may adversely affect health and consequently early mortality.8,9,39 When we excluded these variables from the analysis, the subhazard ratios for most of the LHO variables increased, and the subhazard ratio for male respondents with LHO in the transportation occupation became statistically significant (results not shown). Second, in our data, there was significant variation in the duration of LHO. While the average tenure at LHOs was 19 years, nearly 10% of respondents stayed at their LHOs for less than 5 years. Third, as indicated above, the study did not examine why respondents with LHOs in high-mortality-risk occupations die earlier than respondents with LHOs in the professional and other occupations. We did not have information on the overall health status of the cohort; respondents in high early mortality risk occupations might have had overall poor health status compared with respondents in low early mortality risk occupations. Future research might examine the specific causes of death in each LHO category. We also observed statistically significant sex differences in early mortality risk by LHO. Partly this could be due to uneven gender distribution by major occupations. Billock et al.31 also reported sex differences in COVID-19 mortality by occupation. Future studies might examine why men have excess risk of early mortality in protective service and farming and fishing occupations compared to women using more detailed occupational categories. Finally, the association between LHO and early mortality should not be taken as causation because of potential reverse causality between occupation and early mortality as respondents might choose their occupations based on their overall health status.

5 ∣. CONCLUSION

Examining the relationship between LHO and early mortality helps to illuminate the long-term consequences of occupation. This in turn can inform workers, employers, and policy makers as they consider appropriate preventive health measures for occupations with elevated risks of early mortality. The results of this study showed that both men and women in food preparation, machine operation, personal services, sales, and handlers and helpers occupations had an elevated risk of early mortality compared with respondents in the professional occupations, which was the reference occupation. In addition, men in farming and fishing, clerical, and precision production occupations were more likely to die at earlier age than men in the reference occupation. Overall, LHOs can predict the risk of early mortality.

Supplementary Material

Supplementary Material

ACKNOWLEDGMENTS

The authors thank Walter Alarcon, Amel Omari, and Regina Pana-Cryan, all of CDC, for their valuable comments and suggestions on an earlier draft of the paper. The authors also thank Mary Bohman for her excellent editorial assistance.

Footnotes

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

DISCLOSURE BY AJIM EDITOR OF RECORD

John Meyer declares that he has no conflict of interest in the review and publication decision regarding this article.

ETHICS APPROVAL AND INFORMED CONSENT

Written informed consent was provided by all participants of the Health and Retirement Survey (HRS) and the University of Michigan's Institutional Review Board-approved study protocols. Health and Retirement Survey (HRS) data are publicly available and under the federal regulations for human subjects (45 CFR Part 46), research involving publicly available data sets would not require IRB review. We followed the STROBE checklist in preparing the manuscript.

DISCLAIMER

The findings and conclusions in this study are those of the author and do not necessarily represent the views of the Centers for Disease Control and Prevention (CDC) or the National Institute for Occupational Safety and Health (NIOSH).

SUPPORTING INFORMATION

Additional supporting information can be found online in the Supporting Information section at the end of this article.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available in The Health and Retirement Study at https://hrs.isr.umich.edu/. These data were derived from the following resources available in the public domain: HRS-RAND version, https://hrs.isr.umich.edu/news/data-announcements/rand-hrs-public-data-releases-and-updates.

REFERENCES

  • 1.Johnson NJ, Sorlie PD, Backlund E. The impact of specific occupation on mortality in the US National Longitudinal Mortality Study. Demography. 1999;36:355–367. [PubMed] [Google Scholar]
  • 2.Cutler DM, Lleras-Muney A, Vogl T. Socioeconomic Status and Health: Dimensions and Mechanisms. National Bureau of Economic Research; 2008. [Google Scholar]
  • 3.Marmot MG, Shipley MJ, Rose G. Inequalities in death—specific explanations of a general pattern? Lancet. 1984;323:1003–1006. doi: 10.1016/s0140-6736(84)92337-7 [DOI] [PubMed] [Google Scholar]
  • 4.Leigh JP. Occupational injury and illness in the United States: estimates of costs, morbidity, and mortality. Arch Intern Med. 1997;157:1557–1568. [PubMed] [Google Scholar]
  • 5.Mullan RJ, Murthy LI. Occupational sentinel health events: an updated list for physician recognition and public health surveillance. Am J Ind Med. 1991;19:775–799. doi: 10.1002/ajim.4700190610 [DOI] [PubMed] [Google Scholar]
  • 6.Kim H-J, Min J-Y, Min K-B. The association between longest-held lifetime occupation and late-life cognitive impairment: Korean longitudinal study of aging (2006–2016). Int J Environ Res Public Health. 2020;17:6270. doi: 10.3390/ijerph17176270 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Krieger N, Barbeau EM, Soobader M-J. Class matters: US versus UK measures of occupational disparities in access to health services and health status in the 2000 US National Health Interview Survey. Int J Health Serv. 2005;35:213–236. doi: 10.2190/JKRE-AH92-EDV8-VHYC [DOI] [PubMed] [Google Scholar]
  • 8.LaCroix AZ, Guralnik JM, Berkman LF, Wallace RB, Satterfield S. Maintaining mobility in late life. II. Smoking, alcohol consumption, physical activity, and body mass index. Am J Epidemiol. 1993;137:858–869. doi: 10.1093/oxfordjournals.aje.a116747 [DOI] [PubMed] [Google Scholar]
  • 9.Lee DJ, Fleming LE, Arheart KL, et al. Smoking rate trends in US occupational groups: the 1987 to 2004 National Health Interview Survey. J Occup Environ Med. 2007;49:75–81. doi: 10.1097/JOM.0b013e31802ec68c [DOI] [PubMed] [Google Scholar]
  • 10.Chung S, Domino ME, Stearns SC, Popkin BM. Retirement and physical activity. Am J Prev Med. 2009;36:422–428. doi: 10.1016/j.amepre.2009.01.026 [DOI] [PubMed] [Google Scholar]
  • 11.Murcia M, Chastang J-F, Cohidon C, Niedhammer I. Contribution of occupational factors to social inequalities in self-reported health among French employees. Int Arch Occup Environ Health. 2013;86:541–552. doi: 10.1007/s00420-012-0784-2 [DOI] [PubMed] [Google Scholar]
  • 12.Stattin M, Järvholm B. Occupation, work environment, and disability pension: a prospective study of construction workers. Scand J Public Health. 2005;33:84–90. doi: 10.1080/14034940410019208 [DOI] [PubMed] [Google Scholar]
  • 13.Schrijvers C. Socioeconomic inequalities in health in the working population: the contribution of working conditions. Int J Epidemiol. 1998;27:1011–1018. doi: 10.1093/ije/27.6.1011 [DOI] [PubMed] [Google Scholar]
  • 14.Fujishiro K, Xu J, Gong F. What does “occupation” represent as an indicator of socioeconomic status?: exploring occupational prestige and health. Soc Sci Med. 2010;71:2100–2107. doi: 10.1016/j.socscimed.2010.09.026 [DOI] [PubMed] [Google Scholar]
  • 15.International Labour Organization (ILO). Exposure to Hazardous Chemicals at Work and Resulting Health Impacts: A Global Review. International Labour Office; 2021. [Google Scholar]
  • 16.Asfaw A, Pana-Cryan R, Quay B. Association between longest-held occupation and social security disability insurance benefits receipt. Am J Ind Med. 2020;63:676–684. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Li C-Y, Wu SC, Wen SW. Longest held occupation in a lifetime and risk of disability in activities of daily living. Occup Environ Med. 2000;57:550–554. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Gueorguieva R, Sindelar JL, Falba TA, et al. The impact of occupation on self-rated health: cross-sectional and longitudinal evidence from the health and retirement survey. J Gerontol B Psychol Sci Soc Sci. 2009;64:118–124. doi: 10.1093/geronb/gbn006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Sonnega A, Faul JD, Ofstedal MB, Langa KM, Phillips JW, Weir DR. Cohort profile: the health and retirement study (HRS). Int J Epidemiol. 2014;43:576–585. doi: 10.1093/ije/dyu067 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.RAND. Health and Retirement Study (HRS). RAND HRS Longitudinal File 2020. RAND; 2023. [Google Scholar]
  • 21.Lamarca R, Alonso J, Gomez G, Munoz A. Left-truncated data with age as time scale: an alternative for survival analysis in the elderly population. J Gerontol A Biol Sci Med Sci. 1998;53:M337–M343. doi: 10.1093/gerona/53a.5.m337 [DOI] [PubMed] [Google Scholar]
  • 22.Cheung YB, Gao F, Khoo KS. Age at diagnosis and the choice of survival analysis methods in cancer epidemiology. J Clin Epidemiol. 2003;56:38–43. doi: 10.1016/s0895-4356(02)00536-x [DOI] [PubMed] [Google Scholar]
  • 23.Scherzer R. A tutorial on accounting for competing risks in survival analysis. Clinical Research Statistical Methods Seminar, Kidney Health Research Collaborative, San Francisco VA Medical Center. University of California at San Francisco. 2017. [Google Scholar]
  • 24.Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94:496–509. doi: 10.1080/01621459.1999.10474144 [DOI] [Google Scholar]
  • 25.Rubin DB. Multiple Imputation for Nonresponse in Surveys. John Wiley & Sons; 2004. [Google Scholar]
  • 26.Cleves M, Gould W, Gould WW. An Introduction to Survival Analysis Using Stata. Stata Press; 2008. [Google Scholar]
  • 27.White IR, Royston P, Wood AM. Multiple imputation using chained equations: issues and guidance for practice. Stat Med. 2011;30:377–399. doi: 10.1002/sim.4067 [DOI] [PubMed] [Google Scholar]
  • 28.World Health Organization. WHO/ILO:Almost 2 Million People Die from Work-Related Causes Each Year. World Health Organization; 2021. [Google Scholar]
  • 29.Deeg DJH, De Tavernier W, de Breij S. Occupation-based life expectancy: actuarial fairness in determining statutory retirement age. Front Sociol. 2021;6:675618. doi: 10.3389/fsoc.2021.675618 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Lee H-E, Kim H-R, Chung YK, Kang S-K, Kim E-A. Mortality rates by occupation in Korea: a nationwide, 13-year follow-up study. Occup Environ Med. 2016;73:329–335. doi: 10.1136/oemed-2015-103192 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Billock RM, Steege AL, Miniño A. COVID-19 mortality by usual occupation and industry: 46 states and New York City, United States, 2020. NCHS National Health Statistics Reports. 2022. [PubMed] [Google Scholar]
  • 32.Dembe AE, Erickson JB, Delbos R. Predictors of work-related injuries and illnesses: national survey findings. J Occup Environ Hyg. 2004;1:542–550. doi: 10.1080/15459620490478376 [DOI] [PubMed] [Google Scholar]
  • 33.U.S. Bureau of Labor Statistics (BLS). Survey of Occupational Injuries and Illnesses Data Injuries, Illnesses, and Fatalities. U.S. Bureau of Labor Statistics; 2020. [Google Scholar]
  • 34.Applebaum KM, Asfaw A, O'Leary PK, Busey A, Tripodis Y, Boden LI. Suicide and drug-related mortality following occupational injury. Am J Ind Med. 2019;62:733–741. doi: 10.1002/ajim.23021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Boden LI, Asfaw A, Busey A, et al. Increased all-cause mortality following occupational injury: a comparison of two states. Occup Environ Med. 2022;79:816–823. doi: 10.1136/oemed-2022-108481 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Faunce WA. Occupational status-assignment systems: the effect of status on self esteem. Am J Sociol. 1989;95:378–400. https://www.jstor.org/stable/2780904 [Google Scholar]
  • 37.Matthews KA, Räikkönen K, Everson SA, et al. Do the daily experiences of healthy men and women vary according to occupational prestige and work strain? Psychosom Med. 2000;62:346–353. doi: 10.1097/00006842-200005000-00008 [DOI] [PubMed] [Google Scholar]
  • 38.Christ SL, Fleming LE, Lee DJ, Muntaner C, Muennig PA, Caban-Martinez AJ. The effects of a psychosocial dimension of socioeconomic position on survival: occupational prestige and mortality among US working adults. Sociol Health Illn. 2012;34:1103–1117. doi: 10.1111/j.1467-9566.2012.01456.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Barbeau EM, Krieger N, Soobader M-J. Working class matters: socioeconomic disadvantage, race/ethnicity, gender, and smoking in NHIS 2000. Am J Public Health. 2004;94:269–278. doi: 10.2105/ajph.94.2.269 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Kaplan GA, Pamuk ER, Lynch JW, Cohen RD, Balfour JL. Inequality in income and mortality in the United States: analysis of mortality and potential pathways. BMJ. 1996;312:999–1003. doi: 10.1136/bmj.312.7037.999 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Gornick ME, Eggers PW, Reilly TW, et al. Effects of race and income on mortality and use of services among Medicare beneficiaries. N Engl J Med. 1996;335:791–799. doi: 10.1056/NEJM199609123351106 [DOI] [PubMed] [Google Scholar]
  • 42.Hadley J, Osei A. Does income affect mortality? An analysis of the effects of different types of income on age/sex/race-specific mortality rates in the United States. Med Care. 1982;20:901–914. doi: 10.1097/00005650-198209000-00004 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material
NIHMS2014164-supplement-Supplementary_Material.docx (197.2KB, docx)

Data Availability Statement

The data that support the findings of this study are available in The Health and Retirement Study at https://hrs.isr.umich.edu/. These data were derived from the following resources available in the public domain: HRS-RAND version, https://hrs.isr.umich.edu/news/data-announcements/rand-hrs-public-data-releases-and-updates.

RESOURCES