Effects of education on adult mortality: a global systematic review and meta-analysis
Publicado: 24 enero 2024 a las 2:00 pm
Categorías: Artículos
[responsivevoice_button buttontext=”Dale clic para escuchar” voice=”UK English Female”]
By HME-CHAIN Collaborators DOI:https://doi.org/10.1016/S2468-2667(23)00306-7
Summary
Background
Methods
Findings
Interpretation
Funding
Introduction
This positive trend is attributable to improvements in a wide range of health determinants, such as health-care access and quality of care, technological advancements, reductions in poverty, access to water and sanitation, labour rights, and, crucially, access to education.
which advocated reducing disparities in mortality by addressing the social factors leading to ill health and mortality.
The main pathways through which education can improve health are believed to include social and psychosocial, economic, and cognitive benefits.
As such, education has been recognised as a key determinant to achieve socioeconomic development, gender and social empowerment, and social mobility,
which are all necessary prerequisites to survive and to thrive.
This focus on social determinants of population health is reflected in the UN Sustainable Development Goals (SDGs) in their entirety and, especially, in SDGs 4.1 and 4.3, which aim to ensure that children complete primary and secondary education, and adults have equal access to tertiary education, respectively.
and these changes have been associated with effects on mortality.
In particular, parental education has been highlighted for its effect on child mortality rates, where each additional year of maternal education has been shown to reduce under-5 mortality by 3·0% and each year of paternal education has been shown to reduce this risk by 1·6%.
Methods
Search strategy and selection criteria
Discrepancies during screening were resolved by consensus or referred to a third reviewer. We included studies that assessed individual-level data on years of schooling and adult mortality. We excluded studies that relied on case-crossover or ecological study designs to reduce the risk of bias from unlinked data; studies that did not report key measures of interest, such as relative risk (RR) or hazard ratios; and commentaries, editorials, and letters to the editor publication types. Only studies of adults aged 18 years and older were included; however, exceptions were made for eight studies of cohorts containing small proportions of participants (average 2·9%) below that age threshold. When possible, we ran recalculations of raw data to obtain the appropriate effect measures and to maximise the number of papers included in the systematic review. A complete list of study criteria can be found in appendix 1 (pp 4–5) and a list of the ten extraction languages employed in the full-text reading stage, along with additional details, can be found in appendix 1 (pp 4–6).
Data analysis
technical details are provided in appendix 1 (p 7). We used a ratio model that was able to parameterise RR point estimates with different exposure and reference levels of education (described in Zheng and colleagues
). This process entails using exposure ranges as an independent variable in the regression model and can be used with exposure and reference ranges as well as point values. 10% of datapoints were automatically trimmed, consistent with methods employed by other global meta-analyses.
,
Additionally, we deemed fitting for a non-linear dose–response relationship unnecessary after testing (see appendix 1 pp 8, 14–15 for more details).
,
Covariates were included to allow the main effect to vary by the age of the adult. The estimated effect sizes presented here reflect adjustment for these standardised study-level covariates. 95% uncertainty intervals (UIs) are also reported.
Additionally, we investigated differences between males and females in the reduction of mortality risk by running separate regressions on observations from entirely male or female populations. These sex-stratified models controlled for age only, as there is some evidence to indicate that the effects of marriage differ between sexes.
Finally, we investigated whether any changes in the proportional distribution of the population across given levels of education over time or between cohorts had an undue effect on the aggregate relationship between education and mortality using time-and-cohort disaggregated models.
Sensitivity analyses for the selection and inclusion of individual covariates, predictions associated with these different scenarios, and our approach to standardising non-standard data for each of the models are described in appendix 1 (pp 8–21). Analyses were completed with Python (3.10.9) and R (4.1.2). Statistical code used is publicly available online.
applied to the residuals of the model. The model also estimates the degree of between-study heterogeneity (γ). Our predictions do not incorporate between-study heterogeneity in calculations of uncertainty but we do provide uncertainty in the estimation of between-study heterogeneity in line with other meta-analyses.
More details on the estimation of γ and other model parameters are presented by Zheng and colleagues.
Role of the funding source
Results
| Number of observations, n (%) | |
|---|---|
| Age group | |
| 15–49 years | 2131 (20·58%) |
| 50–59 years | 2933 (28·32%) |
| 60–69 years | 3436 (33·18%) |
| ≥70 years | 1855 (17·91%) |
| Year interval midpoint | |
| 1890–99 | 8 (0·08%) |
| 1960–69 | 198 (1·91%) |
| 1970–79 | 215 (2·08%) |
| 1980–89 | 1386 (13·38%) |
| 1990–99 | 4698 (45·37%) |
| 2000–09 | 3258 (31·46%) |
| 2010–22 | 592 (5·72%) |
| Super-region | |
| Central Europe, eastern Europe, and central Asia | 463 (4·47%) |
| High income | 8893 (85·88%) |
| Latin America and Caribbean | 126 (1·22%) |
| South Asia | 98 (0·95%) |
| Southeast Asia, east Asia, and Oceania | 713 (6·89%) |
| Sub-Saharan Africa | 59 (0·57%) |
| Socio-demographic Index level | |
| 0–0·19 | 8 (0·08%) |
| 0·20–0·39 | 82 (0·79%) |
| 0·40–0·59 | 664 (6·41%) |
| 0·60–0·79 | 6916 (66·79%) |
| 0·80–1 | 2685 (25·93%) |
| Study population is representative of geography | |
| Yes | 4844 (46·78%) |
| No | 5511 (53·22%) |
| Study design | |
| Case-control | 20 (0·19%) |
| Cross-sectional | 574 (5·54%) |
| Prospective cohort | 2700 (26·07%) |
| Retrospective cohort | 7061 (68·19%) |
| Length of follow-up for retrospective cohort studies (n=6977) | |
| 0–4 years | 1308 (18·75%) |
| 5–9 years | 2010 (28·81%) |
| 10–14 years | 1401 (20·08%) |
| 15–19 years | 653 (9·36%) |
| 20–24 years | 561 (8·04%) |
| ≥25 years | 1044 (14·96%) |
| Study-level controls | |
| Age | 6985 (67·46%) |
| Sex | 3556 (34·34%) |
| Race or ethnicity | 2993 (28·90%) |
| Marital status | 2417 (23·34%) |
| Income | 1660 (16·03%) |
| Smoking | 1485 (14·34%) |
| BMI or obesity | 1016 (9·81%) |
| Alcohol use | 939 (9·07%) |
| Employment status | 932 (9·00%) |
| Occupation | 790 (7·63%) |
Discussion
For instance, the 34·3% (95% UI 32·5–36·5) reduction in all-cause mortality risk provided by 18 years of education relative to no education is similar to the reduction in risk of ischaemic heart disease associated with optimal vegetable consumption relative to no vegetable consumption (RR ~0·77)
and the reduction in all-cause mortality risk for adults meeting physical activity guidelines compared with adults not meeting guidelines for aerobics and strengthening (RR ~0·60).
The risk of all-cause mortality for an adult with no education compared with 18 years of education is similar to that of lung cancer incidence or mortality for a person who currently smokes (5 pack-years) compared with a person who has never smoked (RR ~1·52)
and all-cause mortality for a high-volume alcohol drinker compared with an occasional drinker (RR ~1·41).
These comparisons suggest that the benefits of increased investment in education on future population health are comparable to more commonly discussed public health threats, underscoring the crucial importance of increased and equitable educational attainment as a global health goal.
Higher education facilitates access to better employment, higher earnings, quality health care, and increased health knowledge.
Moreover, individuals who are more highly educated tend to develop a larger set of social and psychological resources that shape the health and duration of their life.
,
Although some evidence exists that there are diminishing returns of schooling at higher levels, often attributed to the burden of non-communicable disease and some behavioural risk factors in high-income settings,
we did not find evidence of diminishing benefit in these analyses (appendix 1 pp 14–15).
,
As an individual reaches older age, genetic disposition, daily habits, diet, or other socioeconomic predictors of mortality appear to have a greater influence on mortality risk than their level of educational attainment.
However, despite these influences, educational inequalities in mortality are persistent across the entire lifespan, and the pattern remains the same across time periods and cohorts. The differences observed across the age groups captured in our dataset contribute to the heterogeneity of all observed effect measures.
This variability between sexes is also seen in lower-income countries.
,
,
,
These region-specific or period-specific trends might be obscured in our global analysis. However, it has been shown that education for female individuals has a stronger intergenerational effect than education in male individuals,
and efforts targeting education in primary and secondary age girls, particularly in the regions of the world where education for girls is still behind education for boys, are still needed to reduce existing inequities in educational attainment and improve future population health.
which have accrued largely to countries with a low Socio-demographic Index. We found a similar protective effect of educational attainment at every Socio-demographic Index level present in the available data, emphasising the importance of schooling across all levels of economic development. Several underlying contextual and data-driven reasons might explain this finding: since the Millennium Declaration, variation in Socio-demographic Index level between countries has been decreasing; greater increases in Socio-demographic Index levels in low-income countries than in high-income countries might have been accompanied by increased economic, educational, and gender inequalities; and our analysis is predominantly based on data from high Socio-demographic Index settings given the scarcity of data from lower Socio-demographic Index settings. This relationship should be revisited as the available literature expands and evolves. Alternatively, there might be a universal role of educational attainment in preventing mortality, which is similar between different social contexts.
this limitation highlights the need for additional high-quality research as to the effect of higher education on mortality risk in low-income and middle-income countries. The scarcity of studies from sub-Saharan and north Africa should be addressed in future research and support for this research should be prioritised by relevant funders. Second, we identified variability by region in the quantity and type of study-level controls used across the input studies, which has implications for our analysis and future research (appendix 1 p 12). To minimise any bias resulting from the uneven distribution of controls, we opted to adjust for study-level confounders that we found to have a consistent effect across geographical and economic settings and were present across all regions. This approach is comparable to many other global analyses of the relationship between risk factors and health outcomes. Perhaps most notably, we did not include measures of socioeconomic status, such as wealth or income, in the analysis due to the inconsistency in the variable type and presence across input studies. We cannot control for all sources of confounding in the relationship between education and mortality risk and acknowledge that the effects of education stand behind several essential material, social, and psychosocial resources for health preservation. Thus, controlling for these potential confounders would mean controlling, to some extent, for the effect of education itself. Third, although we limited our review to studies published in 1980 or later, we did not exclude studies that were based on the underlying population years, and, therefore, our data span several decades. This length potentially blurs the influence of medical advances and the associated increasing life expectancy on the effect of educational inequalities on mortality and, although we include sensitivity analyses by time period and cohort, we are unable to make causal claims about the effect of education across the life course. Fourth, although this analysis included studies assessing the effects of education in distinct regions across the world, we are only able to report the average relationship between education and all-cause adult mortality globally and, given the data scarcity, we do not have the power to conduct separate analyses by region. We identified heterogeneity of effect sizes across our large geographically and temporally diverse dataset, which we assume captures factors including between-geography heterogeneity, differences in study population composition, variation in quality and level of adjustment, and other unmeasured differences or non-sampling error across the input studies. A strength of our modelling strategy is that it allows us to estimate a strong signal from these data while incorporating uncertainty from heterogeneity in underlying data. However, studies of more granular effects, such as those specific to region or group,
are needed to increase the efficacy of interventions in education to improve population health.
,
and, although education is not the only possible strategy to improve health and reduce inequalities, it is one key strategy in concerted efforts to healthier and more equitable societies. However, it is important to note that increasing educational disparities in life expectancy have been observed within high-income settings already experiencing increasing educational attainment.
These disparities are seen in our analyses by birth cohort and in existing studies.
Although there are benefits to increasing educational attainment broadly across populations, it is important to apply the proportionate universalism principle
to investments in education to address existing and increasing health inequalities.
) is not sufficient to meet the SDGs.
SDG 4 objectives aim for all children to complete primary and secondary education. Although the world was mostly on track to attain nearly universal primary education before the COVID-19 pandemic,
progress in attainment of higher education is needed to reduce the global health loss attributed to low education.
,
or are heavily focused on infant and child mortality in low-income and middle-income countries.
,
Our research adds to the limited body of research on inequalities in adult all-cause mortality globally and shows that improvements in the social conditions of daily life are necessary for future population health. By increasing years of global schooling, we can help counteract growing disparities in mortality. However, progress will require international commitment and continued investments in educational institutions worldwide. Education cannot remain neglected as a social determinant of health.
Viewing investments in education as investments in health can help address this neglect.
Data sharing
Declaration of interests
Supplementary Materials
-
Supplementary appendix 1
-
Supplementary appendix 2
-
Supplementary appendix 3
References
- 1.
Forecasting life expectancy, years of life lost, and all-cause and cause-specific mortality for 250 causes of death: reference and alternative scenarios for 2016–40 for 195 countries and territories.
Lancet. 2018; 392: 2052-2090
- 2.
Fertility, mortality, migration, and population scenarios for 195 countries and territories from 2017 to 2100: a forecasting analysis for the Global Burden of Disease Study.
Lancet. 2020; 396: 1285-1306
- 3.
Mortality due to low-quality health systems in the universal health coverage era: a systematic analysis of amenable deaths in 137 countries.
Lancet. 2018; 392: 2203-2212
- 4.
Why do some countries do better or worse in life expectancy relative to income? An analysis of Brazil, Ethiopia, and the United States of America.
Int J Equity Health. 2020; 19: 202
- 5.
Health inequalities among British civil servants: the Whitehall II study.
Lancet. 1991; 337: 1387-1393
- 6.
Global strategy for health for all by the year 2000.https://www.who.int/publications-detail-redirect/9241800038Date accessed: April 14, 2023
- 7.
Closing the gap in a generation: health equity through action on the social determinants of health.
World Health Organization, Geneva2008
- 8.
Education and nonmarket outcomes.
US National Bureau of Economic Research, August, 2005
https://www.nber.org/papers/w11582Date: April 16, 2023 - 9.
Why education is the key to socioeconomic differentials in health.
in: Bird CE Conrad P Fremont AM Timmermans S Handbook of medical sociology. 6th edn. Vanderbilt University Press, Nashville, TN2010: 33-51
- 10.
Social conditions as fundamental causes of health inequalities: theory, evidence, and policy implications.
J Health Soc Behav. 2010; 51: S28-S40
- 11.
Stress and health: major findings and policy implications.
J Health Soc Behav. 2010; 51: S41-S53
- 12.
The population education transition curve: education gradients across population exposure to new health risks.
Demography. 2017; 54: 1873-1895
- 13.
Improving health with programmatic, legal, and policy approaches to reduce gender inequality and change restrictive gender norms.
Lancet. 2019; 393: 2522-2534
- 14.
Parental education and inequalities in child mortality: a global systematic review and meta-analysis.
Lancet. 2021; 398: 608-620
- 15.
Transforming our world: the 2030 Agenda for Sustainable Development.https://sustainabledevelopment.un.org/post2015/transformingourworld/publicationDate: 2015Date accessed: April 26, 2023
- 16.
Measuring and forecasting progress towards the education-related SDG targets.
Nature. 2020; 580: 636-639
- 17.
Mapping local variation in educational attainment across Africa.
Nature. 2018; 555: 48-53
- 18.
The impact of increasing education levels on rising life expectancy: a decomposition analysis for Italy, Denmark, and the USA.
Genus. 2019; 75: 11
- 19.
The effect of changes in educational composition on adult female mortality in Brazil.
Res Aging. 2016; 38: 283-298
- 20.
Rayyan—a web and mobile app for systematic reviews.
Syst Rev. 2016; 5: 210
- 21.
Trimmed constrained mixed effects models: formulations and algorithms.
J Comput Graph Stat. 2021; 30: 544-556
- 22.
The Burden of Proof studies: assessing the evidence of risk.
Nat Med. 2022; 28: 2038-2044
- 23.
The effects of family position and status on health.
Soc Sci Med. 1992; 35: 453-464
- 24.
Marital status, living arrangements, and mortality in middle and older age in Europe.
Int J Public Health. 2020; 65: 627-636
- 25.
Global age-sex-specific fertility, mortality, healthy life expectancy (HALE), and population estimates in 204 countries and territories, 1950–2019: a comprehensive demographic analysis for the Global Burden of Disease Study 2019.
Lancet. 2020; 396: 1160-1203
- 26.
The increase in marital status differences in mortality up to the oldest age in seven European countries, 1990–99.
Popul Stud (Camb). 2007; 61: 287-298
- 27.
Inequality in nature and society.
Proc Natl Acad Sci USA. 2017; 114: 13154-13157
- 28.
Bias in meta-analysis detected by a simple, graphical test.
BMJ. 1997; 315: 629-634
- 29.
Uncertainty in heterogeneity estimates in meta-analyses.
BMJ. 2007; 335: 914-916
- 30.
Education: a neglected social determinant of health.
Lancet Public Health. 2020; 5: e361
- 31.
Health effects associated with vegetable consumption: a Burden of Proof study.
Nat Med. 2022; 28: 2066-2074
- 32.
Recommended physical activity and all cause and cause specific mortality in US adults: prospective cohort study.
BMJ. 2020; 370m2031
- 33.
Health effects associated with smoking: a Burden of Proof study.
Nat Med. 2022; 28: 2045-2055
- 34.
Association between daily alcohol intake and risk of all-cause mortality: a systematic review and meta-analyses.
JAMA Netw Open. 2023; 6e236185
- 35.
Educational inequalities in cause-specific mortality in middle-aged and older men and women in eight western European populations.
Lancet. 2005; 365: 493-500
- 36.
The relationship between education and health: reducing disparities through a contextual approach.
Annu Rev Public Health. 2018; 39: 273-289
- 37.
Educational attainment and adult mortality.
in: Rogers RG Crimmins EM International handbook of adult mortality. Springer, Dordrecht, Netherlands2011: 241-261
- 38.
Different indicators of socioeconomic status and their relative importance as determinants of health in old age.
Int J Equity Health. 2017; 16: 173
- 39.
Education and mortality among older adults in China.
Soc Sci Med. 2015; 127: 134-142
- 40.
Mortality inequalities and trends in low- and middle-income countries, 1990–2015.
Population. 2017; 72: 221-296
- 41.
Poverty and prime-age mortality in eastern and southern Africa: evidence from Zambia and Kenya.
World Dev. 2012; 40: 1839-1853
- 42.
Trends and socioeconomic gradients in adult mortality around the developing world.
Popul Dev Rev. 2013; 39: 1-29
- 43.
Five insights from the Global Burden of Disease Study 2019.
Lancet. 2020; 396: 1135-1159
- 44.
Education and adult mortality in middle-income countries: surprising gradients in six nationally-representative longitudinal surveys.
SSM Popul Health. 2020; 12100649
- 45.
Variations in the relation between education and cause-specific mortality in 19 European populations: a test of the “fundamental causes” theory of social inequalities in health.
Soc Sci Med. 2015; 127: 51-62
- 46.
Political declaration of the high-level meeting of the General Assembly on the Prevention and Control of Non-communicable Diseases.https://documents-dds-ny.un.org/doc/UNDOC/GEN/N11/458/94/PDF/N1145894.pdf?OpenElementDate: 2012Date accessed: April 26, 2023
- 47.
The implications of changing education distributions for life expectancy gradients.
Soc Sci Med. 2021; 272113712
- 48.
Widening educational inequalities in mortality in more recent birth cohorts: a study of 14 European countries.
J Epidemiol Community Health. 2023; 77: 400-408
- 49.
Fair society, healthy lives.
Public Health. 2012; 126: S4-10
- 50.
Education and mortality in three Eastern European populations: findings from the PrivMort retrospective cohort study.
Eur J Public Health. 2019; 29: 549-554
- 51.
The Sustainable Development Goals report 2021.https://digitallibrary.un.org/record/3932350Date accessed: May 9, 2023
- 52.
Changes in socioeconomic differentials in old age life expectancy in four Nordic countries: the impact of educational expansion and education-specific mortality.
Eur J Ageing. 2022; 19: 161-173
- 53.
The 1991–2004 evolution in life expectancy by educational level in Belgium based on linked census and population register data.
Eur J Popul. 2009; 25: 175-196
- 54.
Increased educational attainment and its effect on child mortality in 175 countries between 1970 and 2009: a systematic analysis.
Lancet. 2010; 376: 959-974
Article info
Publication history
Identification
Copyright
User license
00306-7&id=gr1.jpg){kind=link}
00306-7&id=gr2.jpg){kind=link}
00306-7&id=gr3.jpg){kind=link}
00306-7&id=gr4.jpg){kind=link}
00306-7&id=gr5.jpg){kind=link}
ScienceDirect
Figures
-
Figure 1PRISMA diagram
-
Figure 2Included studies by location
-
Figure 3Adult all-cause mortality by education level
-
Figure 4Relationship between education and adult mortality by age group
-
Figure 5Dose–response relationship
Tables
Source
https://www.thelancet.com/journals/lanpub/article/PIIS2468-2667(23)00306-7/fulltext
Deja un comentario