Abstract
INTRODUCTION
We conducted a cross‐national comparison of the association between main lifetime occupational skills and later‐life cognitive function across four economically and socially distinct countries.
METHODS
Data were from population‐based studies of aging and their Harmonized Cognitive Assessment Protocols (HCAPs) in the US, South Africa, India, and Mexico (N = 10,037; Age range: 50 to 105 years; 2016 to 2020). Main lifetime occupational skill was classified according to the International Standard Classification of Occupations. Weighted, adjusted regression models estimated pooled and country‐specific associations between main lifetime occupational skill and later‐life general cognitive function in men and women.
RESULTS
We observed positive gradients between occupational skill and later‐life cognitive function for men and women in the US and Mexico, a positive gradient for women but not men in India, and no association for men or women in South Africa.
DISCUSSION
Main lifetime occupations may be a source of later‐life cognitive reserve, with cross‐national heterogeneity in this association.
Highlights
No studies have examined cross‐national differences in the association of occupational skill with cognition.
We used data from Harmonized Cognitive Assessment Protocols in the US, Mexico, India, and South Africa.
The association of occupational skill with cognitive function varies by country and gender.
Keywords: aging, cognitive function, cross‐national comparison, gender, occupation, risk factors, work environment
1. BACKGROUND
Dementia is a syndrome associated with aging that involves symptoms of cognitive impairment that affect the ability to conduct everyday activities. 1 Employment is a social determinant of health that may have a complex relationship with dementia risk. 2 In addition to providing economic benefits, employment can be a source of cognitive stimulation through the use of skills such as reading, problem solving, communication, social engagement, and task‐specific skill mastery, all of which are associated with improved later‐life cognitive function and reduced dementia risk. 3 , 4 Time spent in work takes up a large proportion of waking hours for much of the global population. 5 When accumulated over a lifetime, the activities undertaken at work may have strong implications for later‐life cognitive outcomes. Improved understanding of the relationship between lifetime occupational skill complexity and later‐life cognitive outcomes may help to elucidate dementia etiology and inform better workplace design and policies to support the health of rapidly aging populations around the world.
The existing evidence on lifetime occupational skill and later‐life cognitive health outcomes is somewhat mixed, with most studies observing protective associations, 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 and others observing null, negative, or mixed associations across different types of skills and cognitive domains. 18 , 19 , 20 , 21 , 22 , 23 , 24 Most of this existing research is from high‐income countries, with the exception of a study in Brazil; much of it includes non‐population‐representative samples, and it does not capture the diversity of job types experienced around the world. Greater occupational skill complexity may not be uniformly associated with better later‐life cognitive outcomes across global populations. In addition to reflecting skills, occupational skill categories capture exposures to workplace hazards that may vary across countries, such as physical strain, psychological stressors, exposure to pollutants, and level of safety regulations. 25 Given that over 75% of global dementia cases are projected to occur in low‐ and middle‐income countries by 2050, 26 the evidence base on lifetime occupational skill complexity and later‐life cognitive outcomes should be expanded to better represent diverse global populations. 27
We aimed to estimate and compare the associations between main lifetime occupational skill level and later‐life cognitive function using harmonized, population‐based data on men and women from the United States, Mexico, India, and South Africa. These countries are at differing levels of economic development, with different predominant job industries, employment policies, and societal norms around gender and work. We aimed to exploit cross‐national variation in the types of occupations that involve similar skills according to international classifications, to gain a more comprehensive understanding of the relationship between lifetime occupational skill and later‐life cognitive health across diverse contexts. Main lifetime occupational skill levels were harmonized across diverse job types and industries in the countries under study using the International Standard Classification of Occupations 2008 (ISCO‐08). All associations were estimated separately for men and women, as men and women may experience different cognitive risk and protective factors within the workplace, even with equivalent job titles, due to the gendered nature of work. 28
2. METHODS
2.1. Study design and populations
The United States (US) Health and Retirement Study (HRS) and its International Partner Studies are population‐based longitudinal studies of aging with harmonized designs and measures. The current study used data from the currently available cross‐sectional Harmonized Cognitive Assessment Protocol (HCAP) studies embedded in the US HRS 29 , 30 and its International Partner Studies in South Africa, 31 , 32 Mexico, 33 , 34 and India, 35 , 36 linked to data on main lifetime occupation and covariates from their respective parent cohorts (Table 1). The HRS and its International Partner Studies are designed to be representative of their general middle‐aged and older country populations, except for the “Health and Ageing in Africa: A Longitudinal Study in South Africa” (HAALSI) cohort, which is representative of the Agincourt subdistrict, a low‐income region of former forced racial segregation during apartheid in northeast South Africa. 31 They collect data on sociodemographic and economic factors, health behaviors and conditions, family structure, and physical and functional health through ongoing in‐person interviews, physical assessments, and biomarker data collections. 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 The HCAP studies aim to provide a flexible yet comparable instrument to facilitate cross‐national comparisons of cognitive health of older adults around the world. 29 Eligible participants for the present analysis were direct respondents of each HCAP study cited above, with complete data on all study measures. All participants provided informed consent. The US Health and Retirement Study and its International Partner Studies have received ethical approval from the appropriate institutions. The present analysis was deemed exempt from regulation by the University of Michigan Health Sciences and Behavioral Sciences Institutional Review Board (HUM00178420). This study adheres to the ethical standards of the Declaration of Helsinki.
TABLE 1.
Study populations and parent international partner studies for Harmonized Cognitive Assessment Protocols (HCAPs) in the United States, South Africa, Mexico, and India.
| Country | International partner study; HCAP substudy | HCAP year(s) of data collection | HCAP age range | HCAP N |
|---|---|---|---|---|
| United States | Health and Retirement Study (HRS); HRS‐HCAP | 2016–2017 | ≥65 | 3,347 |
| South Africa | Health and Ageing in Africa: A Longitudinal Study in South Africa (HAALSI); HAALSI‐HCAP (also referred to as HAALSI Dementia Study) | 2019–2020 | ≥50 | 628 |
| Mexico | Mexican Health and Aging Study (MHAS) Cognitive Aging Ancillary Study; MexCog | 2016 | ≥54 | 2,042 |
| India | Longitudinal Aging Study in India (LASI); Longitudinal Aging Study in India ‐ Diagnostic Assessment of Dementia (LASI‐DAD) | 2017–2020 | ≥60 | 4,096 |
Note: The number of direct respondents in the HCAP age ranges shown above are displayed in this table.
RESEARCH IN CONTEXT
Systematic review: We examined literature on occupational title, occupational skill, occupational complexity, and later‐life cognitive outcomes. Prior studies did not incorporate data from low‐ and middle‐income countries, did not compare data across countries, did not examine gender differences, and did not use measures encompassing multiple domains of cognitive function. Relevant work is cited and discussed.
Interpretation: We identified cross‐national heterogeneity in the generally protective association between lifetime occupational skill and later‐life cognitive function across the US, Mexico, India, and a rural region of South Africa. In India, women experienced stronger associations than men. In South Africa, there was no association among men or women, which may reflect an apartheid‐era lack of labor opportunities.
Future directions: Future research should further investigate reasons for cross‐national differences in this relationship. Consideration of cognitive hazards associated with certain jobs should be considered alongside the cognitive benefits of employment.
2.2. Measures
2.2.1. Exposure: Main lifetime occupational skill
Main lifetime occupational titles were assessed during study interviews in each parent study and harmonized according to the ISCO‐08. 36 The ISCO‐08 is intended to capture all types of jobs around the world and was developed by the International Labour Organization based on national and international statistics with consensus from labor experts around the world. 37 The ISCO‐08 classifies jobs into 436 groups at the finest level of detail, with aggregation to 10 major groups, which are mapped onto one of four hierarchical skill levels: skill level 1 (routine physical or manual tasks), skill level 2 (tasks such as operating machinery, maintenance, storage of information), skill level 3 (performance of complex technical and practical tasks), and skill level 4 (complex problem solving, decision‐making, and creativity). Detailed descriptions of each skill level are as follows:
Skill Level 1, involving the performance of simple and routine physical or manual tasks such as cleaning, digging, and lifting and carrying materials by hand. Example occupations include cleaners, freight handlers, garden laborers, and kitchen assistants.
Skill Level 2, involving the performance of tasks such as operating machinery and electronic equipment; driving vehicles; maintenance and repair of electrical and mechanical equipment; and manipulation, ordering, and storage of information. Example occupations include butchers, bus drivers, secretaries, accounting clerks, shop sales assistants, police officers, building electricians, and motor vehicle mechanics.
Skill Level 3, involving the performance of complex technical and practical tasks that require an extensive body of factual, technical, and procedural knowledge in a specialized field. Example occupations include shop managers, medical laboratory technicians, legal secretaries, commercial sales representatives, and computer support technicians.
Skill Level 4, involving the performance of tasks that require complex problem‐solving, decision‐making, and creativity based on an extensive body of theoretical and factual knowledge in a specialized field. Example occupations include sales and marketing managers, civil engineers, secondary school teachers, medical practitioners, musicians, and computer systems analysts.
Our exposure variable also included categories for never having worked for pay (reference), and for “don't know,” “other,” “military,” and “missing.” These latter four categories were combined due to small numbers, although “other” was the predominant category of the four in all cohorts. The crosswalks between the raw variables for main lifetime occupational titles and the ISCO‐08 categories are available in Tables S1–S6.
2.2.2. Outcome: General cognitive function
Factor scores representing general cognitive function (GCF) were derived from the HCAP cognitive measures in each country, representing memory, orientation, attention, executive function, and verbal fluency abilities. We used an item‐banking approach to pre‐statistical and statistical harmonization of the GCF factor scores, described in detail elsewhere. 38 , 39 , 40 This approach enabled us to account for necessary differences in cognitive test administration across countries, such as language translations, cultural adaptations, and low‐literacy and low‐numeracy adaptations. The factor scores were standardized to the HRS‐HCAP distribution, which had a mean of 0 and a standard deviation (SD) of 1, such that all model estimates are expressed in SD units of the HRS‐HCAP sample.
2.2.3. Covariates
Potential confounders of the relationship between main lifetime occupational skills and later‐life cognitive function would have to arise early in life, prior to entry into the labor force. We considered potential confounders as: age (continuous, mean‐centered by sex/gender), sex/gender (male; female), minority group status (minority; nonminority), educational attainment according to the International Standard Classification of Education (none or early childhood education; primary education; lower secondary education; upper secondary education; any college), mother's education (none; any education; missing), and father's education (none; any education; missing). Parental education was classified according to the dichotomy of “none” versus “any education,” as this was the finest coding available for these variables in HAALSI, which had a high prevalence of lack of parental education (72% of HAALSI participants had a father with no formal education, and 93% had a mother with no formal education). Minority group status was classified in different ways for each country, as each has groups that are racialized or otherwise marginalized in contextually specific ways, and who consistently have less access to resources, power, and prestige in their respective societies. We dichotomized minority group status to facilitate common adjustment across countries in pooled models, while recognizing the data limitation this creates as minoritization is a complex process. This variable was classified in the US according to the race/ethnicity groups of non‐Hispanic Black, Hispanic, and Other (“minority”) and non‐Hispanic White (“nonminority”); in Mexico as rural (“minority”) and urban region of residence (“nonminority”); in India according to caste as Scheduled Caste or Scheduled Tribe (“minority”) and Other Backward Class or Other or no caste group (“nonminority”); and in South Africa according to country of birth as Mozambique or other (“minority”) and South Africa (“nonminority”), as the study region is home to a large population of former refugees from Mozambique belonging to the same ethnic group as the South African‐born living in the region, but who experience many forms of marginalization.
2.3. Statistical analysis
Characteristics of the sample were described overall and by country. Mean GCF factor scores according to main lifetime occupational skill levels for men and women were described overall and by country. We specified multivariable‐adjusted linear regression models to estimate the relationships between main lifetime occupational skill and later‐life general GCF scores in models pooled and then stratified by country. Models were run separately for men and women. To evaluate the statistical significance of gender differences in these associations, we tested statistical interaction terms between main lifetime occupational skill and gender in the country‐specific models. All models were weighted using the HCAP sampling weights provided by each cohort to account for sampling and nonresponse to each HCAP study as well as their respective parent cohorts. The sampling weights thus scale estimates to the national populations of older adults in the US, Mexico, and India, and to the regional population of older adults in Agincourt subdistrict, South Africa. In addition, all models were adjusted for the covariates described above, with adjustment for country interacted with the minority group indicator in pooled models. This interaction allowed differential effects of minoritization on later‐life cognitive functions across countries. We conducted a sensitivity analysis with additional adjustment for adult height in meters, as a marker of genetics and early childhood net nutrition. 41 We adjusted for height as a sensitivity analysis rather than the main analysis, to avoid potentially biasing the main analysis due to missing observations on height (550/10,037; 5%). All analyses were conducted using StataSE 17.0 (College Station, TX).
2.4. Diversity, equity, and inclusion
Diversity, equity, and inclusion were essential to the concept of this analysis, which is concerned with inequalities in later‐life cognitive function according to main lifetime occupational skills and whether this association is modified across countries with varying levels of economic development. We stratified all analyses a priori by gender as we considered that societal gender roles and norms would result in differential occupational opportunities for men and women, to differing degrees across countries. We created a harmonized minority status variable as best possible while maintaining data harmonization, as described above, aiming to reflect the process of minoritization as it occurs across the four countries under study. We adjusted for this variable in modeling while allowing it to have differential associations with the outcome across countries. We incorporate diversity, equity, and inclusion into the interpretation of our results by considering the structural social and economic life course influences on occupational opportunities for men and women of the birth cohorts represented in the countries under study.
3. RESULTS
A total of 10,037 participants (99% of eligible) had complete data and were included (Figure S1). Table 2 describes characteristics of the sample, overall and by country. The mean age (SD) of participants ranged from 68.1 (9.0) years (Mexico) to 76.6 (7.5) years (US). Lifetime occupational skill levels varied by country, with the proportion never having worked ranging from 4% (US) to 32% (India), and the proportion with a main lifetime occupation at skill levels 3 or 4 ranging from 6% (India) to 31% (US). Across all countries except for the US, women were more likely to have never worked than men (Table S7). The most common occupational titles and their corresponding ISCO‐08 skill levels by gender and country are shown in Table S8.
TABLE 2.
Characteristics of the sample, overall and by country, Harmonized Cognitive Assessment Protocols.
| Characteristic | Overall, N = 10,037 | HRS‐HCAP (United States), N = 3344 | MexCog (Mexico), N = 2024 | LASI‐DAD (India), N = 4075 | HAALSI‐HCAP (South Africa), N = 594 |
|---|---|---|---|---|---|
| N (%) | N (%) | N (%) | N (%) | N (%) | |
| Age in years (mean, SD) | 71.7 (8.9) | 76.6 (7.5) | 68.1 (9.0) | 69.7 (7.6) | 69.2 (11.5) |
| Age range, years | 50–105 | 64–102 | 54–104 | 60–105 | 50–99 |
| Sex/gender | |||||
| Male | 4262 (42%) | 1325 (40%) | 825 (41%) | 1882 (46%) | 230 (39%) |
| Female | 5775 (58%) | 2019 (60%) | 1199 (59%) | 2193 (54%) | 364 (61%) |
| Ethnic minority status | |||||
| Minority | 2999 (30%) | 962 (29%) | 877 (43%) | 956 (23%) | 204 (34%) |
| Non‐minority | 7038 (70%) | 2382 (71%) | 1147 (57%) | 3119 (77%) | 390 (66%) |
| Education | |||||
| None or preschool | 3915 (39%) | 22 (1%) | 1023 (51%) | 2543 (62%) | 327 (55%) |
| Primary education | 1309 (13%) | 131 (4%) | 452 (22%) | 527 (13%) | 199 (34%) |
| Lower secondary | 1107 (11%) | 453 (14%) | 317 (16%) | 312 (8%) | 25 (4%) |
| Upper secondary | 2366 (24%) | 1773 (53%) | 60 (3%) | 502 (12%) | 31 (5%) |
| Any college | 1340 (13%) | 965 (29%) | 172 (8%) | 191 (5%) | 12 (2%) |
| Father's education | |||||
| None | 4065 (41%) | 113 (3%) | 737 (36%) | 2785 (68%) | 430 (72%) |
| Any education | 4885 (49%) | 2778 (83%) | 998 (49%) | 1046 (26%) | 63 (11%) |
| Missing | 1087 (11%) | 453 (14%) | 289 (14%) | 244 (6%) | 101 (17%) |
| Mother's education | |||||
| None | 5043 (50%) | 110 (3%) | 906 (45%) | 167 (4%) | 553 (93%) |
| Any education | 4305 (43%) | 2965 (89%) | 873 (43%) | 3474 (85%) | 33 (6%) |
| Missing | 689 (7%) | 269 (8%) | 245 (12%) | 434 (11%) | 8 (1%) |
| Main lifetime occupational skill | |||||
| Never worked | 1971 (20%) | 137 (4%) | 373 (18%) | 1305 (32%) | 156 (26%) |
| Level 1 | 1810 (18%) | 185 (6%) | 748 (37%) | 588 (14%) | 289 (49%) |
| Level 2 | 4063 (40%) | 1701 (51%) | 669 (33%) | 1611 (40%) | 82 (14%) |
| Levels 3/4 | 1528 (15%) | 1053 (31%) | 224 (11%) | 227 (6%) | 24 (4%) |
| Don't know, other, military, or missing | 665 (7%) | 268 (8%) | 10 (< 1%) | 344 (8%) | 43 (7%) |
Abbreviations: HAALSI, Health and Ageing in Africa: A Longitudinal Study in South Africa; HCAP, Harmonized Cognitive Assessment Protocol; HRS, Health and Retirement Study; LASI‐DAD, Longitudinal Aging Study in India ‐ Diagnostic Assessment of Dementia; MexCog, Mexican Health and Aging Study Cognitive Aging Ancillary Study.
Table 3 shows mean (SD) GCF scores across categories of main lifetime occupational skill level among men and women, overall and by country. There were positive gradients in mean GCF scores with increasing main lifetime occupational skill level for men and women in all countries, except for South Africa. These same trends were apparent in multivariable‐adjusted, sampling‐weighted regression models (Table 4). In all countries pooled together, the estimated mean GCF scores for individuals who never worked for pay and were in the reference category of covariates (the model intercepts) were −1.21 SD units (95% confidence interval [CI]: −1.36, −1.06) for men and −1.11 SD units (95% CI: −1.22, −1.01) for women (Table 4). The mean difference in GCF score for those with a main lifetime occupational skill level of 3/4 was 0.25 SD units (95% CI: 0.12, 0.39) for men and 0.20 SD units (95% CI: 0.11, 0.28) for women (Table 4). To contextualize the magnitudes of these effect estimates, the estimates for 1 year of age in our models were −0.036 SD units for men and −0.040 SD units for women (not shown). Thus, these estimates for main lifetime occupational skill are comparable to an approximate 6.9‐year lower age for men (0.25 SD units/−0.036 SD units) and an approximate 5.0‐year lower age for women (0.20 SD units/−0.040 SD units).
TABLE 3.
Unadjusted mean (SD) general cognitive function scores by main lifetime occupational skill, overall and by country, Harmonized Cognitive Assessment Protocols.
| Overall | HRS‐HCAP (United States) | MexCog (Mexico) | LASI‐DAD (India) | HAALSI‐HCAP (South Africa) | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Main lifetime occupational skill | Men mean (SD) | Women mean (SD) | Men mean (SD) | Women mean (SD) | Men mean (SD) | Women mean (SD) | Men mean (SD) | Women mean (SD) | Men mean (SD) | Women mean (SD) |
| Never worked | −1.1 (1.0) | −1.3 (1.0) | −0.7 (0.9) | −0.7 (1.0) | −1.1 (1.0) | −0.9 (1.0) | −1.2 (1.0) | −1.5 (0.9) | −0.9 (0.8) | −0.8 (0.8) |
| Level 1 | −0.9 (0.9) | −1.2 (1.0) | −0.2 (0.9) | −0.6 (1.0) | −1.0 (1.0) | −1.1 (1.0) | −1.1 (0.9) | −1.7 (0.8) | −1.0 (0.7) | −0.9 (0.8) |
| Level 2 | −0.7 (1.0) | −0.5 (1.2) | −0.3 (1.0) | 0.0 (1.0) | −0.4 (1.0) | −0.4 (1.1) | −1.0 (0.9) | −1.6 (0.9) | −0.9 (0.7) | −1.1 (0.9) |
| Levels 3/4 | 0.1 (0.9) | 0.3 (1.0) | 0.3 (0.9) | 0.4 (1.0) | 0.3 (0.8) | 0.4 (0.9) | −0.4 (0.8) | −0.1 (0.9) | −0.9 (0.8) | −1.0 (0.8) |
| Don't know, other, military, missing | −0.8 (1.1) | −1.0 (1.2) | −0.1 (1.1) | −0.4 (1.2) | −0.3 (0.6) | 0.6 (0.7) | −1.3 (1.0) | −1.7 (0.7) | −0.8 (0.8) | −0.9 (0.8) |
Note: The general cognitive function scores in all countries are standardized to the HRS‐HCAP distribution, which has a mean of 0 and a standard deviation of 1.
Abbreviations: HAALSI, Health and Ageing in Africa: A Longitudinal Study of an INDEPTH Community in South Africa; HCAP, Harmonized Cognitive Assessment Protocol; HRS, Health and Retirement Study; LASI‐DAD, Harmonised Diagnostic Assessment of Dementia for the Longitudinal Aging Study in India; MexCog, Mexican Health and Aging Study; SD, standard deviation.
TABLE 4.
Associations between main lifetime occupational skill and general cognitive function (GCF) scores, by gender and country.
| Main lifetime occupational skill | Overall | HRS‐HCAP (United States) | MexCog (Mexico) | LASI‐DAD (India) | HAALSI‐HCAP (South Africa) |
|---|---|---|---|---|---|
| Coef. (95% CI) | Coef. (95% CI) | Coef. (95% CI) | Coef. (95% CI) | Coef. (95% CI) | |
| Men | |||||
| Never worked | – | – | – | – | – |
| Level 1 | 0.08 (−0.05, 0.21) | 0.24 (−0.04, 0.53) | 0.28 (−0.11, 0.67) | 0.07 (−0.09, 0.23) | ‐0.20 (−0.63, 0.22) |
| Level 2 | 0.09 (−0.03, 0.21) | 0.23 (−0.03, 0.49) | 0.42 (0.03, 0.82) | 0.04 (−0.11, 0.19) | ‐0.16 (−0.60, 0.28) |
| Levels 3/4 | 0.25 (0.12, 0.39) | 0.53 (0.26, 0.80) | 0.52 (0.11, 0.94) | 0.09 (−0.10, 0.29) | 0.05 (−0.55, 0.65) |
| Intercept | −1.21 (−1.36, −1.06) | −0.94 (−1.91, 0.03) | −1.50 (−1.89, −1.11) | −1.63 (−1.78, −1.48) | −0.98 (−1.41, −0.55) |
| Women | |||||
| Never worked | – | – | – | – | – |
| Level 1 | −0.09 (−0.16, −0.03) | −0.01 (−0.24, 0.23) | −0.03 (−0.12, 0.07) | −0.03 (−0.12, 0.06) | −0.12 (−0.31, 0.06) |
| Level 2 | 0.06 (0.01, 0.12) | 0.17 (0.02, 0.31) | 0.19 (0.08, 0.31) | 0.02 (−0.05, 0.08) | −0.41 (−0.84, 0.02) |
| Levels 3/4 | 0.20 (0.11, 0.28) | 0.29 (0.13, 0.44) | 0.36 (0.14, 0.57) | 0.23 (0.03, 0.43) | −0.19 (−0.77, 0.38) |
| Intercept | −1.11 (−1.22, −1.01) | −0.49 (−0.79, −0.19) | −1.42 (−1.53, −1.31) | −1.88 (−1.93, −1.84) | −0.87 (−1.04, −0.70) |
| p‐values for gender differences | |||||
| Level 1 | 0.094 | 0.209 | 0.124 | 0.237 | 0.738 |
| Level 2 | 0.209 | 0.748 | 0.336 | 0.765 | 0.417 |
| Levels 3/4 | 0.003 | 0.201 | 0.791 | 0.003 | 0.814 |
Note: All models are adjusted for age (mean‐centered by gender), minority status, educational attainment, and mother's and father's educational attainment. The overall models, with all countries pooled together, are additionally adjusted for country interacted with minority group. The coefficients for GCF scores are expressed in standard deviation units of the HRS‐HCAP sample. Coefficients for the “Don't know, other, military, missing” category are not shown, as they are not interpretable. The model intercepts represent the mean value of the GCF score for someone who never worked for pay and is the reference value of all covariates (mean age, belonging to a nonminority group, no education or early childhood education, no maternal education, and no paternal education). The p‐values for gender differences are from models including both men and women with statistical interaction terms between occupational skill and gender. All models incorporate HCAP sampling weights.
Abbreviations: CI, confidence interval; HAALSI, Health and Ageing in Africa: A Longitudinal Study of an INDEPTH Community in South Africa; HCAP, Harmonized Cognitive Assessment Protocols; HRS, Health and Retirement Study; LASI‐DAD, Harmonised Diagnostic Assessment of Dementia for the Longitudinal Aging Study in India; MexCog, Mexican Health and Aging Study; SD, standard deviation.
In the United States and Mexico, the association between main lifetime occupational skill and later‐life cognitive function appeared stronger among men than women, although the gender differences were not statistically significant (Table 4). In contrast, in India, the association was stronger among women than men (0.23; 95% CI: 0.03, 0.43 [women] and 0.09; 95% CI: −0.10, 0.29 [men] for skill level 3/4 vs never worked for pay; p‐value for gender difference = 0.003; Table 4). In South Africa, there was no gradient in later‐life cognitive function according to main lifetime occupational skill, although estimates were imprecise (Table 4). The predicted later‐life GCF scores across main lifetime occupational skill levels from the models shown in Table 4 are shown in Figure 1. Results negligibly differed when measured adult height was included in the models (Table S9).
FIGURE 1.
Model‐estimated mean values of general cognitive function according to main lifetime occupational skill level, overall and by country, among men and women. Models incorporate Harmonized Cognitive Assessment Protocol (HCAP) sampling weights and are adjusted for age, minority status, educational attainment, and mother's and father's educational attainment. The overall model additionally adjusts for country. The model‐estimated mean values are scaled to the distribution of the Health and Retirement Study (HRS)‐HCAP sample (eg, −1 indicates −1 SD from the mean of the HRS‐HCAP sample distribution). HAALSI, Health and Ageing in Africa: A Longitudinal Study in South Africa; LASI‐DAD, Longitudinal Aging Study in India ‐ Diagnostic Assessment of Dementia; MexCog, Mexican Health and Aging Study Cognitive Aging Ancillary Study.
4. DISCUSSION
In this large, population‐based, cross‐nationally harmonized study, we observed positive gradients between increasing main lifetime occupational skill level and later‐life general cognitive function in the US, Mexico, and India, but not in South Africa. These associations were strong in magnitude, equivalent to an approximate 6.9‐year reduction in age for men and a 5.0‐year reduction in age for women, for persons having the most highly skilled occupation relative to those never having worked for pay. In the US and Mexico, men may have experienced greater later‐life cognitive returns from highly skilled occupations than women. Conversely, in India, women appeared to experience greater later‐life cognitive returns from highly skilled occupations than men. This study identifies important heterogeneity in the association between lifetime occupational skill and later‐life cognitive function in men and women living in four diverse high‐ and middle‐income countries.
Our results are comparable with those observed in previous studies, mostly conducted in high‐income countries such as the US, England, Sweden, Finland, Australia, Canada, and Italy, all of which indicate that greater lifetime occupational skill complexity is associated with better cognitive outcomes in later life. 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 Despite their largely consistent findings, these previous studies have used varying measures of both occupational skill and cognitive outcomes and in varying types of study samples, including Alzheimer's disease registry cohorts, convenience samples, and regionally or nationally representative cohort studies. Our results advance knowledge on this topic as they expand the existing evidence base to include population‐representative samples from Mexico, India, and South Africa; they use cross‐nationally harmonized cognitive function data to facilitate direct comparison of the cognitive function outcome across countries; and they provide new evidence on gender differences in the association between main lifetime occupational skill and later‐life cognitive function association across countries.
A potential mechanism explaining our findings is that occupations may be a source of cognitive reserve, which is a construct invoked to explain resilience against aging‐related brain pathology. 42 According to cognitive reserve theory, greater occupational skill complexity across the lifetime should help to protect later‐life cognitive function from accumulating aging‐related brain pathology or injury. 42 However, we observed heterogeneity in associations across our countries under study, and some prior studies have observed null associations, mixed associations across different types of occupational skill or cognitive function domains, 18 , 22 , 23 , 24 or even faster rates of cognitive decline over time with greater occupational complexity. 21 Variations in the measurement and definition of occupational skill across studies may further contribute to inconsistencies in associations. 43 In our study, the cross‐national heterogeneity that we observed may be, at least in part, due to differences in the specific job activities, responsibilities, pay, and benefits involved at each skill level across countries, as well as differences in workplace hazards that could be detrimental to cognitive health. Hazards such as hard manual labor, psychological stress, lack of physical safety, lack of workplace protections, and exposure to pollutants may vary across countries despite equivalent job titles, as these hazards would be influenced by country‐specific labor policies. 25
In this study, there was no apparent association between lifetime occupational skill and later‐life cognitive function among men in India or among men and women in South Africa. This finding in India is consistent with studies in South Korea and France, which found occupational class to be associated with later‐life cognitive impairment and dementia with parkinsonism, respectively, among women, but not among men. 19 , 20 In India, occupational opportunities have historically been and remain limited due to social mobility restrictions imposed by India's caste system. Occupational opportunities for non‐White South Africans of the birth cohorts included here were severely restricted due to apartheid. It is likely that the highly skilled jobs that were observed in this sample, namely “small business owner,” did not involve activities that would support later‐life cognitive health due to their often informal nature and lack of associated formal training or benefits. 44 Our results for South African men and women were also imprecise due to limited sample size. Future research should investigate implications of labor conditions for cognitive aging of older non‐White South Africans in larger samples, such as the full HAALSI cohort or national samples such as the National Income Dynamics Study.
Our results may also be explained by different mechanisms of selection into the labor market and into higher‐skilled jobs for men and women in different countries. For example, in countries with less egalitarian labor markets and gender norms, such as India, women of the birth cohorts under study here with highly skilled occupations may be a highly selected population subgroup. 45 Our observation of gender differences in the occupational skill‐cognition association in India, if interpreted causally, indicates that women who achieve highly skilled occupations in India may experience particularly strong later‐life cognitive health benefits. This interpretation is consistent with resource substitution theory, which posits that health‐promoting socioeconomic resources, such as highly skilled jobs, have greater influence for individuals with fewer alternative socioeconomic resources than they do for more advantaged individuals who have a greater number of alternative resources to draw upon. 46 While we adjusted for early‐life factors as best possible, we cannot rule out differential selection into the labor force and more highly skilled jobs for women with higher early‐life cognitive function.
This study has limitations. The harmonized cognitive function outcome was measured at a single point in time, and we could not assess rates of cognitive decline over time. We also did not examine specific domains of cognitive function, which is an important area for future inquiry. The raw variables for main lifetime job titles varied across countries in terms of their level of detail, types of job titles assessed, and durations, all of which may lead to measurement error in the main lifetime occupational skill variable. We expect this measurement error would be nondifferential with respect to cognitive function, and, if present, would most likely bias the results towards the null. We did not have data on early‐life health conditions that would limit the ability to work, so residual confounding by this factor is possible. Our results may be subject to selective survival bias if survival to the time of HCAP assessment is conditional on main lifetime occupation and cognitive function earlier in life. 47 The potential degree of selective survival bias present in our results may be stronger for countries with lower life expectancies, as they would have higher rates of all‐cause mortality at younger ages. Future studies should investigate the potential impact of differential selective survival bias in comparisons of aging outcomes across countries with varying life expectancies.
This study has important strengths. This is one of the first analyses of a potentially key life course contributor to later‐life cognitive health using a truly harmonized cognitive function assessment across diverse country settings. The HCAP comprehensively assesses cognitive function in domains of memory, orientation, attention, executive function, and verbal fluency, with appropriate adaptions to suit the social, cultural, and educational contexts of the countries under study. Our use of statistically harmonized factor scores allowed us to take these adaptations in cognitive test items into account through latent variable modeling. Our large sample permitted the investigation of gender differences, which is important for gender equity as the nature of employment and nonemployment varies for men and women across the countries studied here. For example, the reference category of “never worked” may largely represent lack of employment opportunities or long‐term illness or disability for men, while it may largely represent homemaking or informal domestic labor for women. We adjusted for respondent's education and their mother's and father's education, which are key determinants of occupational opportunities, and included HCAP sampling weights to ensure that estimates were population‐representative. Our results were also robust to adjustment for adult height as a marker of genetic factors and early‐life net nutrition.
In summary, this is one of the first population‐representative, cross‐nationally harmonized studies of main lifetime occupational skills and later‐life cognitive function. We observed protective associations comparable in magnitude to a 6.9‐year lower age for men and a 5.0‐year lower age for women, for the most highly skilled occupations relative to never working for pay, across all countries combined. However, we observed notable heterogeneity in this association across countries, suggesting that employment is not always protective of later‐life cognitive health, and that occupational opportunities may vary across countries. This study opens several future research directions about the nature of employment across the life course in relation to later‐life cognitive health for men and women, and the reasons why associations may vary across countries. Overall, the results suggest that enriching occupational opportunities in low‐resource global settings may help to promote population cognitive health.
CONFLICT OF INTEREST STATEMENT
L.C.K. reports receiving grant funding from the National Institute on Aging of the National Institutes of Health and a speaker honorarium from Johns Hopkins University. R.N.J. reports receiving grant funding from the National Institute on Aging of the National Institutes of Health. K.M.L. reports receiving grant funding from the National Institute on Aging of the National Institutes of Health and the United States Alzheimer's Association; consulting fees on National Institutes of Health‐funded projects at Harvard University, University of Pennsylvania, University of Minnesota, University of Colorado, Dartmouth University, and the University of Southern California; payment for expert testimony; and, participation on a Data Monitoring and Safety Board for a clinical trial at Indiana University. D.W. reports receiving grant funding from the National Institute on Aging of the National Institutes of Health and participation on advisory boards for the English Longitudinal Study of Ageing, the Canadian Longitudinal Study of Aging, the Longitudinal Aging Study in India, and the China Health and Retirement Study. J.L. reports receiving grant funding from the National Institute on Aging of the National Institutes of Health and the BrightFocus Bold Ideas Initiative; consulting fees from the RAND Corporation; honoraria from the University of California, Berkely, and Southern Illinois University School of Medicine; support for attending meetings from University College London, Venice International University, the United States Alzheimer's Association, the World Health Organization, and the University of California, Berkeley; and participation on Data Safety Monitoring or Advisory Boards for the Asian Development Bank, the Egyptian Health and Retirement Study, the Malawi Longitudinal Study of Families and Health, the Lausanne Cohort 65+ Study, the Berkeley Initiative for Transparency in the Social Sciences, the Longitudinal Study of Health and Aging in Kenya, the Malaysia Ageing and Retirement Study, the English Longitudinal Study of Ageing, the Japanese Study of Aging and Retirement, the China Health and Retirement Longitudinal Study, the WHO Consortium of Metrics and Evidence for Healthy Aging, and the Cognitive Level Enhancement through Vision Exams and Refraction. R.W. reports receiving grant funding from the National Institute on Aging of the National Institutes of Health. A.L.G. reports receiving grant funding from the National Institute on Aging of the National Institutes of Health and a speaker honorarium from the University of Michigan. B.Q.O. and C.W. have no disclosures to report. Author disclosures are available in the supporting information.
Supporting information
Supporting Information
Supporting Information
ACKNOWLEDGMENTS
This work was supported by the National Institute on Aging of the National Institutes of Health (grant numbers R01 AG070953 to L.C.K. and A.L.G., R01 AG030153 to J.L. and A.L.G., R01 AG042778 and R01 AG051125 to J.L., U24 AG065182 to D.W. and K.M.L., R01 AG051158 and R01 AG018016 to R.W.).
Kobayashi LC, O'Shea BQ, Wixom C, et al. Lifetime occupational skill and later‐life cognitive function among older adults in the United States, Mexico, India, and South Africa. Alzheimer's Dement. 2024;20:1933–1943. 10.1002/alz.13665
DATA AVAILABILITY STATEMENT
The Gateway to Global Aging Data website provides documentation and code for the harmonized datasets used here, with links to the parent cohort data: https://g2aging.org/downloads. The HRS and HRS‐HCAP data are available at: https://hrs.isr.umich.edu/data‐products. The MHAS and MexCog data are available at: https://www.mhasweb.org/DataProducts/Home.aspx. The LASI data are available at: https://iipsindia.ac.in/content/LASI‐data. The LASI‐DAD data are available at: https://lasi‐dad.org/access‐data. The LASI and LASI‐DAD data are also available at: https://g2aging.org/downloads. The HAALSI and HAALSI‐HCAP data are available at: https://haalsi.org/data.
REFERENCES
- 1. Livingston G, Huntley J, Sommerlad A, et al. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet. 2020;396(10248):413‐446. doi: 10.1016/S0140-6736(20)30367-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Benach J, Muntaner C, Solar O, Santana V, Quinlan M. Introduction to the WHO commission on social determinants of health employment conditions network (Emconet) study, with a glossary on employment relations. Int J Health Serv. 2010;40(2):195‐207. doi: 10.2190/HS.40.2.a [DOI] [PubMed] [Google Scholar]
- 3. Wilson RS, Mendes De Leon CF, Barnes LL, et al. Participation in cognitively stimulating activities and risk of incident Alzheimer disease. JAMA. 2002;287(6):742‐748. doi: 10.1001/jama.287.6.742 [DOI] [PubMed] [Google Scholar]
- 4. Duffner LA, Deckers K, Cadar D, Steptoe A, de Vugt M, Köhler S. The role of cognitive and social leisure activities in dementia risk: assessing longitudinal associations of modifiable and on‐modifiable risk factors. Epidemiol Psychiatr Sci. 2022;31:e5. doi: 10.1017/S204579602100069X [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Organisation for Economic Co‐operation and Development . Employment—Hours worked—OECD Data [Internet]. OECD Data. [cited 2023 Feb 24]. Available from: http://data.oecd.org/emp/hours‐worked.htm [Google Scholar]
- 6. Kivimäki M, Walker KA, Pentti J, et al. Cognitive stimulation in the workplace, plasma proteins, and risk of dementia: three analyses of population cohort studies. BMJ. 2021;374:n1804. doi: 10.1136/bmj.n1804 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Karp A, Andel R, Parker MG, Wang HX, Winblad B, Fratiglioni L. Mentally stimulating activities at work during midlife and dementia risk after age 75: follow‐up study from the Kungsholmen Project. Am J Geriatr Psychiatry. 2009;17(3):227‐236. doi: 10.1097/JGP.0b013e318190b691 [DOI] [PubMed] [Google Scholar]
- 8. Pool LR, Weuve J, Wilson RS, Bültmann U, Evans DA, Mendes de Leon CF. Occupational cognitive requirements and late‐life cognitive aging. Neurology. 2016;86(15):1386‐1392. doi: 10.1212/WNL.0000000000002569 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Rydström A, Darin‐Mattsson A, Kåreholt I, et al. Occupational complexity and cognition in the FINGER multidomain intervention trial. Alzheimers Dement. 2022;18(12):2438‐2447. doi: 10.1002/alz.12561 [DOI] [PubMed] [Google Scholar]
- 10. Staudinger UM, Yu YL, Cheng B. Novel information processing at work across time is associated with cognitive change in later life: a 14‐year longitudinal study. Psychol Aging. 2020;35(6):793‐805. doi: 10.1037/pag0000468 [DOI] [PubMed] [Google Scholar]
- 11. Boots EA, Schultz SA, Almeida RP, et al. Occupational complexity and cognitive reserve in a middle‐aged cohort at risk for Alzheimer's disease. Arch Clin Neuropsychol. 2015;30(7):634‐642. doi: 10.1093/arclin/acv041 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Hyun J, Hall CB, Katz MJ, et al. Education, occupational complexity, and incident dementia: a COSMIC collaborative cohort study. J Alzheimers Dis. 2022;85(1):179‐196. doi: 10.3233/JAD-210627 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Andel R, Dávila‐Roman AL, Grotz C, Small BJ, Markides KS, Crowe M. Complexity of work and incident cognitive impairment in Puerto Rican older adults. J Gerontol B Psychol Sci Soc Sci. 2019;74(5):785‐795. doi: 10.1093/geronb/gbx127 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Kröger E, Andel R, Lindsay J, Benounissa Z, Verreault R, Laurin D. Is complexity of work associated with risk of dementia? The Canadian Study of Health And Aging. Am J Epidemiol. 2008;167(7):820‐830. doi: 10.1093/aje/kwm382 [DOI] [PubMed] [Google Scholar]
- 15. Correa Ribeiro PC, Lopes CS, Lourenço RA. Complexity of lifetime occupation and cognitive performance in old age. Occup Med (Lond). 2013;63(8):556‐562. doi: 10.1093/occmed/kqt115 [DOI] [PubMed] [Google Scholar]
- 16. Andel R, Kåreholt I, Parker MG, Thorslund M, Gatz M. Complexity of primary lifetime occupation and cognition in advanced old age. J Aging Health. 2007;19(3):397‐415. doi: 10.1177/0898264307300171 [DOI] [PubMed] [Google Scholar]
- 17. Andel R, Crowe M, Pedersen NL, et al. Complexity of work and risk of Alzheimer's disease: a population‐based study of Swedish twins. J Gerontol B Psychol Sci Soc Sci. 2005;60(5):P251‐P258. doi: 10.1093/geronb/60.5.p251 [DOI] [PubMed] [Google Scholar]
- 18. Lane AP, Windsor TD, Andel R, Luszcz MA. Is occupational complexity associated with cognitive performance or decline? Results from the Australian longitudinal study of ageing. Gerontology. 2017;63(6):550‐559. doi: 10.1159/000475559 [DOI] [PubMed] [Google Scholar]
- 19. Kim HJ, Min JY, Min KB. 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(17):6270. doi: 10.3390/ijerph17176270 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Helmer C, Letenneur L, Rouch I, et al. Occupation during life and risk of dementia in French elderly community residents. J Neurol Neurosurg Psychiatry. 2001;71(3):303‐309. doi: 10.1136/jnnp.71.3.303 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Hyun J, Katz MJ, Lipton RB, Sliwinski MJ. Mentally challenging occupations are associated with more rapid cognitive decline at later stages of cognitive aging. J Gerontol B Psychol Sci Soc Sci. 2021;76(4):671‐680. doi: 10.1093/geronb/gbz122 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Finkel D, Andel R, Gatz M, Pedersen NL. The role of occupational complexity in trajectories of cognitive aging before and after retirement. Psychol Aging. 2009;24(3):563‐573. doi: 10.1037/a0015511 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Soh Y, Eng CW, Mayeda ER, et al. Association of primary lifetime occupational cognitive complexity and cognitive decline in a diverse cohort: results from the KHANDLE study. Alzheimers Dement. 2023;19(9):3926‐3935. doi: 10.1002/alz.13038 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Coleman ME, Roessler MEH, Peng S, et al. Social enrichment on the job: complex work with people improves episodic memory, promotes brain reserve, and reduces the risk of dementia. Alzheimers Dement. 2023;19(6):1655‐1665. doi: 10.1002/alz.13035 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. World Health Organization and International Labour Organization . WHO/ILO joint estimates of the work‐related burden of disease and injury, 2000‐2016: global monitoring report. World Health Organization and International Labour Organization; 2021. [Google Scholar]
- 26. Prince MJ, Wimo A, Guerchet MM, Ali GC, Wu Y‐T, Prina M, et al. World Alzheimer Report 2015 ‐ The Global Impact of Dementia: An analysis of prevalence, incidence, cost and trends. London: Alzheimer's Disease International, 2015. 84 p.
- 27. Glymour MM, Whitmer RA. Using cross‐cultural studies to improve evidence on dementia prevention: lessons from the special issue sponsored by the international research network on dementia prevention (IRNDP). J Alzheimers Dis. 2019;70(s1):S5‐10. doi: 10.3233/JAD-190304 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Mielke MM, Aggarwal NT, Vila‐Castelar C, et al, Diversity and Disparity Professional Interest Area Sex and Gender Special Interest Group . Consideration of sex and gender in Alzheimer's disease and related disorders from a global perspective. Alzheimers Dement. 2022;18(12):2707‐2724. doi: 10.1002/alz.12662 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Langa KM, Ryan LH, McCammon RJ, et al. The health and retirement study harmonized cognitive assessment protocol project: study design and methods. Neuroepidemiology. 2020;54(1):64‐74. doi: 10.1159/000503004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Sonnega A, Faul JD, Ofstedal MB, Langa KM, Phillips JWR, Weir DR. Cohort profile: the health and retirement study (HRS). Int J Epidemiol. 2014;43(2):576‐585. doi: 10.1093/ije/dyu067 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Gómez‐Olivé FX, Montana L, Wagner RG, et al. Cohort profile: health and ageing in Africa: a longitudinal study of an INDEPTH community in South Africa (HAALSI). Int J Epidemiol. 2018;47(3):689‐690j. doi: 10.1093/ije/dyx247 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Bassil DT, Farrell MT, Wagner RG, et al. Cohort profile update: cognition and dementia in the health and aging in Africa longitudinal study of an INDEPTH community in South Africa (HAALSI dementia). Int J Epidemiol. 2022;51(4):e217‐e226. doi: 10.1093/ije/dyab250 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Wong R, Michaels‐Obregon A, Palloni A. Cohort profile: the Mexican health and aging study (MHAS). Int J Epidemiol. 2017;46(2):e2. doi: 10.1093/ije/dyu263 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Mejia‐Arango S, Nevarez R, Michaels‐Obregon A, et al. The Mexican cognitive aging ancillary study (Mex‐Cog): study design and methods. Arch Gerontol Geriatr. 2020;91:104210. doi: 10.1016/j.archger.2020.104210 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Perianayagam A, Bloom D, Lee J, et al. Cohort profile: the longitudinal ageing study in India (LASI). Int J Epidemiol. 2022;51(4):e167‐e176. doi: 10.1093/ije/dyab266 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Lee J, Banerjee J, Khobragade PY, Angrisani M, Dey AB. LASI‐DAD study: a protocol for a prospective cohort study of late‐life cognition and dementia in India. BMJ Open. 2019;9(7):e030300. doi: 10.1136/bmjopen-2019-030300 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. International Labour Organization . International Standard Classification of Occupations 2008 (ISCO‐08): Structure, group definitions and correspondence tables. Vol. 1. International Labour Organization; 2012. [Google Scholar]
- 38. Vonk JMJ, Gross AL, Zammit AR, et al. Cross‐national harmonization of cognitive measures across HRS HCAP (USA) and LASI‐DAD (India). PlOS One. 2022;17(2):e0264166. doi: 10.1371/journal.pone.0264166 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Briceño EM, Arce Rentería M, Gross AL, et al. A cultural neuropsychological approach to harmonization of cognitive data across culturally and linguistically diverse older adult populations. Neuropsychology. 2023;37(3):247‐257. doi: 10.1037/neu0000816 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Briceño EM, Gross AL, Giordani BJ, et al. Pre‐statistical considerations for harmonization of cognitive instruments: harmonization of ARIC, CARDIA, CHS, FHS, MESA, and NOMAS. J Alzheimers Dis. 2021;83(4):1803‐1813. doi: 10.3233/JAD-210459 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Perkins JM, Subramanian SV, Davey Smith G, Özaltin E. Adult height, nutrition, and population health. Nutr Rev. 2016;74(3):149‐165. doi: 10.1093/nutrit/nuv105 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Stern Y. Cognitive reserve in ageing and Alzheimer's disease. Lancet Neurol. 2012;11(11):1006‐1012. doi: 10.1016/S1474-4422(12)70191-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Hussenoeder FS, Riedel‐Heller SG, Conrad I, Rodriguez FS. Concepts of mental demands at work that protect against cognitive decline and dementia: a systematic review. Am J Health Promot. 2019;33(8):1200‐1208. doi: 10.1177/0890117119861309 [DOI] [PubMed] [Google Scholar]
- 44. Clercq De F. Apartheid and the organized labor movement. Rev Afr Polit Econ. 1979(14):69‐77. [Google Scholar]
- 45. UN Women . Country Fact Sheet | UN Women Data Hub [Internet]. Feb 27, 2023. Available from: https://data.unwomen.org/country/india
- 46. Ross CE, Mirowsky J. Sex differences in the effect of education on depression: resource multiplication or resource substitution? Soc Sci Med. 2006;63(5):1400‐1413. doi: 10.1016/j.socscimed.2006.03.013 [DOI] [PubMed] [Google Scholar]
- 47. Mayeda ER, Filshtein TJ, Tripodis Y, Glymour MM, Gross AL. Does selective survival before study enrolment attenuate estimated effects of education on rate of cognitive decline in older adults? A simulation approach for quantifying survival bias in life course epidemiology. Int J Epidemiol. 2018;47(5):1507‐1517. doi: 10.1093/ije/dyy124 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supporting Information
Supporting Information
Data Availability Statement
The Gateway to Global Aging Data website provides documentation and code for the harmonized datasets used here, with links to the parent cohort data: https://g2aging.org/downloads. The HRS and HRS‐HCAP data are available at: https://hrs.isr.umich.edu/data‐products. The MHAS and MexCog data are available at: https://www.mhasweb.org/DataProducts/Home.aspx. The LASI data are available at: https://iipsindia.ac.in/content/LASI‐data. The LASI‐DAD data are available at: https://lasi‐dad.org/access‐data. The LASI and LASI‐DAD data are also available at: https://g2aging.org/downloads. The HAALSI and HAALSI‐HCAP data are available at: https://haalsi.org/data.