When is the difference method conservative for assessing mediation?

Impact of Poverty on Stroke Recurrence: A Population-Based Study

BACKGROUND AND OBJECTIVES

Poverty is associated with greater stroke incidence. The relationship between poverty and stroke recurrence is less clear.

METHODS

In this population-based study, incident strokes within the Greater Cincinnati/Northern Kentucky region were ascertained during the 2015 study period and followed up for recurrence until December 31, 2018. The primary exposure was neighborhood socioeconomic status (nSES), defined by the percentage of households below the federal poverty line in each census tract in 4 categories (≤5%, >5%-10%, >10%-25%, >25%). Poisson regression models provided recurrence rate estimates per 100,000 residents using population data from the 2015 5-year American Community Survey, adjusting for age, sex, and race. In a secondary analysis, Cox models allowed for the inclusion of vascular risk factors in the assessment of recurrence risk by nSES among those with incident stroke.

RESULTS

Of 2,125 patients with incident stroke, 245 had a recurrent stroke during the study period. Poorer nSES was associated with increased stroke recurrence, with rates of 12.5, 17.5, 25.4, and 29.9 per 100,000 in census tracts with ≤5%, >5%-10%, >10%-25%, and >25% below the poverty line, respectively (p < 0.01). The relative risk (95% CI) for recurrent stroke among Black vs White individuals was 2.54 (1.91-3.37) before adjusting for nSES, and 2.00 (1.47-2.74) after adjusting for nSES, a 35.1% decrease. In the secondary analysis, poorer nSES (HR 1.74, 95% CI 1.10-2.76 for lowest vs highest category) and Black race (HR 1.31, 95% CI 1.01-1.70) were both independently associated with recurrence risk, though neither retained significance after full adjustment. Age, diabetes, and left ventricular hypertrophy were associated with increased recurrence risk in fully adjusted models.

DISCUSSION

Residents of poorer neighborhoods had a dose-dependent increase in stroke recurrence risk, and neighborhood poverty accounted for approximately one-third of the excess risk among Black individuals. These results highlight the importance of poverty, race, and the intersection of the 2 as potent drivers of stroke recurrence.

Mediators of the Association Between Socioeconomic Status and Survival After Out-of-Hospital Cardiac Arrest: A Systematic Review

Low socioeconomic status (SES) is associated with poor outcomes after out-of-hospital cardiac arrest (OHCA). Patient characteristics, care processes, and other contextual factors may mediate the association between SES and survival after OHCA. Interventions that target these mediating factors may reduce disparities in OHCA outcomes across the socioeconomic spectrum. This systematic review identified and quantified mediators of the SES-survival after OHCA association. Electronic databases (MEDLINE, Embase, PubMed, Web of Science) and grey literature sources were searched from inception to July or August 2023. Observational studies of OHCA patients that conducted mediation analyses to evaluate potential mediators of the association between SES (defined by income, education, occupation, or a composite index) and survival outcomes were included. A total of 10 studies were included in this review. Income (n = 9), education (n = 4), occupation (n = 1), and composite indices (n = 1) were used to define SES. The proportion of OHCA cases that had bystander involvement, presented with an initial shockable rhythm, and survived to hospital discharge or 30 days increased with higher SES. Common mediators of the SES-survival association that were evaluated included initial rhythm (n = 6), emergency medical services response time (n = 5), and bystander cardiopulmonary resuscitation (n = 4). Initial rhythm was the most important mediator of this association, with a median percent excess risk explained of 37.4% (range 28.6%-40.0%; n = 5; 1 study reported no mediation) and mediation proportion of 41.8% (n = 1). To mitigate socioeconomic disparities in outcomes after OHCA, interventions should target potentially modifiable mediators, such as initial rhythm, which may involve improving bystander awareness of OHCA and the need for prompt resuscitation.

Do Sleep Problems Explain the Association Between Work Stress and the Trajectories of Work Ability From Midlife to Pensionable Age?

OBJECTIVE

This study examined whether mid-life work stress, defined as job strain and effort-reward imbalance (ERI), predicts work ability trajectories observed 12 years preceding the individual pensionable age. In addition, the role of sleep problems as a mediator in these associations was examined.

METHODS

Survey data were collected from 2707 Finnish municipal employees.

RESULTS

Identified work ability trajectories were "stable excellent," "stable good," "moderate," and "low decreasing." Baseline job strain and ERI were associated with a greater likelihood of belonging to impaired work ability trajectories when compared with "stable good" trajectory. Baseline sleep problems explained the association of job strain by 38% and of ERI by 54%.

CONCLUSIONS

Mid-life work stress is associated with work ability in the last years preceding pensionable age. Sleep problems might be a potential mediator in these associations.

Cooking skills, living alone, and mortality: JAGES cohort study

BACKGROUND

Living alone without someone to cook meals for them can happen more frequently in aging due to bereavement, divorce, or other family changes. Health risks to older adults due to poor cooking skills may be more pronounced among those living alone. We aimed to examine whether cooking skills are associated with mortality according to cohabitation status in older Japanese people.

METHODS

Participants in the Japan Gerontological Evaluation Study, a population-based cohort of independent older adults, were followed for three years (n = 10,647). Cooking skill was assessed using a scale with good validity and modified for Japanese people in the baseline survey. After stratification by living alone or together, participants with high and low cooking skills were matched on demographic, socioeconomic, health-related factors, and availability of food stores using propensity score matching. All-cause mortality risks were compared between high and low cooking skills using Cox regression models.

RESULTS

During the follow-up, 520 of the 10,647 participants died. One hundred and seventy-one pairs of high and low cooking skills were matched among those living alone, and 2,161 pairs among those living with others were matched as well. The hazard ratio of the low level of cooking skills (vs. high) was 2.50 (95% confidence interval [CI]: 1.10-5.68) among those living alone, while 1.05 (95% CI: 0.82-1.33) among those living with others.

CONCLUSION

Lower cooking skills were associated with a higher risk of mortality only among those living alone. Cooking skills may be important for older adults who live alone to reduce mortality risk.

Smoking, blood DNA methylation sites and lung cancer risk

Altered DNA methylation (DNAm) might be a biological intermediary in the pathway from smoking to lung cancer. In this study, we investigated the contribution of differential blood DNAm to explain the association between smoking and lung cancer incidence. Blood DNAm was measured in 2321 Strong Heart Study (SHS) participants. Incident lung cancer was assessed as time to event diagnoses. We conducted mediation analysis, including validation with DNAm and paired gene expression data from the Framingham Heart Study (FHS). In the SHS, current versus never smoking and pack-years single-mediator models showed, respectively, 29 and 21 differentially methylated positions (DMPs) for lung cancer with statistically significant mediated effects (14 of 20 available, and five of 14 available, positions, replicated, respectively, in FHS). In FHS, replicated DMPs showed gene expression downregulation largely in trans, and were related to biological pathways in cancer. The multimediator model identified that DMPs annotated to the genes AHRR and IER3 jointly explained a substantial proportion of lung cancer. Thus, the association of smoking with lung cancer was partly explained by differences in baseline blood DNAm at few relevant sites. Experimental studies are needed to confirm the biological role of identified eQTMs and to evaluate potential implications for early detection and control of lung cancer.

Neighborhood Factors, Individual Stressors, and Cardiovascular Health Among Black and White Adults in the US: The Reasons for Geographic and Racial Differences in Stroke (REGARDS) Study

Importance

Chronic stress has been posited to contribute to racial disparities in cardiovascular health. Investigation of whether neighborhood- and individual-level stressors mediate this disparity is needed.

Objective

To examine whether racial differences in ideal cardiovascular health (ICH) are attenuated by experiences with neighborhood- and individual-level stressors within a racially and geographically diverse population sample.

Design, Setting, and Participants

This cross-sectional study examined data from 7720 participants in the Reasons for Geographic and Racial Differences in Stroke (REGARDS) Study who completed the second in-home visit (2013-2016). The REGARDS study is a population-based, longitudinal study of 30 239 non-Hispanic Black and non-Hispanic White adults aged 45 years or older at baseline (2003-2007). Data for the present study were analyzed from June to July 2021 and in March 2022.

Exposures

Neighborhood physical environment (eg, excessive noise, violence; scored from 7-28, with higher scores indicating more problems), neighborhood safety (scored as very safe, safe, or not safe), neighborhood social cohesion (eg, shared values; scored from 5-25, with higher scores indicating higher cohesion), perceived stress (eg, coping; scored from 0-16, with higher scores indicating greater perceived stress), and the experience of discrimination (yes or no).

Main Outcomes and Measures

Ideal cardiovascular health (ICH), measured as a composite of 4 health behaviors (cigarette smoking, diet, physical activity, body mass index) and 3 health factors (blood pressure, cholesterol, and glucose levels).

Results

The sample included 7720 participants (mean [SD] age, 71.9 [8.3] years; 4390 women [56.9%]; 2074 Black participants [26.9%]; and 5646 White participants [73.1%]). Black participants compared with White participants reported higher perceived stress (mean [SD] score, 3.2 [2.8] vs 2.8 [2.7]) and more often reported discrimination (77.0% vs 24.0%). Black participants also reported poorer neighborhood physical environment (mean [SD] score, 11.2 [3.8] vs 9.8 [2.9]) and social cohesion (mean [SD] score, 15.5 [2.0] vs 15.7 [1.9]) and more often reported their neighborhoods were unsafe (54.7% vs 24.3%). The odds of having a high total ICH score (ie, closer to ideal) were lower for Black adults compared with White adults, both overall (adjusted odds ratio [AOR], 0.53; 95% CI, 0.45-0.61) and by gender (men: AOR, 0.73 [95% CI, 0.57-0.93]; women: AOR, 0.45 [95% CI, 0.37-0.54]). In mediation analyses, the racial disparity in total ICH score was attenuated by neighborhood physical environment (5.14%), neighborhood safety (6.27%), neighborhood social cohesion (1.41%), and discrimination (11.01%). In stratified analyses, the factors that most attenuated the racial disparity in total ICH scores were neighborhood safety among men (12.32%) and discrimination among women (14.37%). Perceived stress did not attenuate the racial disparity in total ICH scores.

Conclusions and Relevance

In this cross-sectional study of Black and White US adults aged 45 years and older, neighborhood-level factors, including safety and physical and social environments, and individual-level factors, including discrimination, attenuated racial disparities in cardiovascular health. Interventional approaches to improve ICH that separately target neighborhood context and discrimination by gender and race are warranted.

Mediation analysis in the presence of continuous exposure measurement error

The difference method is used in mediation analysis to quantify the extent to which a mediator explains the mechanisms underlying the pathway between an exposure and an outcome. In many health science studies, the exposures are almost never measured without error, which can result in biased effect estimates. This article investigates methods for mediation analysis when a continuous exposure is mismeasured. Under a linear exposure measurement error model, we prove that the bias of indirect effect and mediation proportion can go in either direction but the mediation proportion is usually be less biased when the associations between the exposure and its error-prone counterpart are similar with and without adjustment for the mediator. We further propose methods to adjust for exposure measurement error with continuous and binary outcomes. The proposed approaches require a main study/validation study design where in the validation study, data are available for characterizing the relationship between the true exposure and its error-prone counterpart. The proposed approaches are then applied to the Health Professional Follow-up Study, 1986-2016, to investigate the impact of body mass index (BMI) as a mediator for mediating the effect of physical activity on the risk of cardiovascular diseases. Our results reveal that physical activity is significantly associated with a lower risk of cardiovascular disease incidence, and approximately half of the total effect of physical activity is mediated by BMI after accounting for exposure measurement error. Extensive simulation studies are conducted to demonstrate the validity and efficiency of the proposed approaches in finite samples.

Traumatic brain injury, stroke, and epilepsy: A mediation study in a Danish nationwide cohort

OBJECTIVE

Traumatic brain injury (TBI) and stroke are well-known causes of acquired epilepsy. TBI is also a risk factor for stroke, and injury-induced stroke may indirectly convey a proportion of the epilepsy risk following TBI. We studied the extent to which the effect of TBI on epilepsy operated through intermediary stroke.

METHODS

We analyzed a nationwide, matched, register-based cohort of adults ≥ 40 years of age whose first TBI at Danish hospitals was recorded between 2004 and 2016. A matched reference population was sampled for comparison. During follow-up, we recorded all acute strokes. Cox proportional hazard models and the difference method were used to estimate the total and controlled direct effect hazard ratios (HRs) of TBI on epilepsy and the indirect effect HRs of TBI on epilepsy operating through stroke, and to calculate the proportion eliminated. Analyses were stratified by severity of, age at, and time since TBI.

RESULTS

We followed 57 900 persons with TBI (48.6% males) from median age 61 years (interquartile range = 51-75), and 561 977 age- and sex-matched references. The total effect of TBI on epilepsy was higher for persons aged 40-59 years (HR = 5.15, 95% confidence interval [CI] = 4.65-5.72) than for persons aged ≥ 60 years (HR = 4.55, 95% CI = 4.19-4.95). In contrast, the indirect effect of TBI mediated by stroke was lower for persons aged 40-59 years (HR = 1.02, 95% CI = 1.02-1.03) than for persons aged ≥ 60 years (HR = 1.05, 95% CI = 1.04-1.06). We estimated 2.3% and 5.6% of the risk of epilepsy after TBI to operate through stroke for these age groups, respectively.

SIGNIFICANCE

Less than 6% of the risk of epilepsy following TBI operated through intermediary stroke. However, this mechanism seems to play an increasing role with age and for late onset epilepsies. This warrants further investigation.

Mediation of lifestyle-associated variables on the association between occupation and incident cardiovascular disease

The main aim was to examine the association between occupational groups and incident cardiovascular disease (CVD), and to which extent associations are mediated by lifestyle-associated variables (cardiorespiratory fitness, smoking, BMI, exercise, and diet). A total of 304,702 participants (mean age 42.5 yrs., 47% women), who performed a health profile assessment in Sweden between 1982 and 2019, were included in the analyses. CVD incidence was obtained from national registers. All participants were free from CVD prior to the health profile assessment. Occupational group was defined using the Swedish Standard Classification of Occupations and analyzed separately (13 different occupational groups) as well as after aggregation into four occupational groups (white-collar high-skilled, white-collar low-skilled, blue-collar high-skilled and blue-collar low-skilled). Cardiorespiratory fitness, BMI, exercise, smoking, and diet were included as mediators and analyzed separately in single models and simultaneously in one multiple mediation model. All mediation analyses were adjusted for sex, age, length of education and calendar time. White-collar high-skilled was set as reference in all analyses. Blue-collar and low-skilled occupation had a higher risk of incident CVD compared to the reference. Cardiorespiratory fitness, BMI, exercise, smoking, and diet mediated 48% to 54% of the associations between reference and the other aggregated occupational groups. In the single model, the strongest mediators were cardiorespiratory fitness, smoking and BMI. In conclusion, blue-collar and low-skilled occupations had a significantly higher risk for incident CVD compared to white-collar high-skilled workers, with the association mediated to a large extent by variation in lifestyle-associated variables.

Stigma towards patients with mental illness: An online survey of United States nurses

People with mental illness experience significant health disparities, including morbidity and premature mortality. Evidence suggests that stigma is a contributing factor to these observed inequities. The tripartite conceptualization of stigma proposes that three problems underlie stigma: problems of knowledge (ignorance), attitudes (prejudice) and behaviour (discrimination). There is limited prior research concerning stigma towards mental illness among nurses in the United States (US). The aims of this study were to assess stigma among US nurses towards patients with mental illness, compare the stigma expressed by nurses working in medical/surgical settings with mental health settings, and identify factors associated with stigma. Participants were recruited online from national professional nursing organizations in the US. We collected demographic data and administered measures of mental health stigma and stigma-related mental health knowledge. Mental health nurses demonstrated comparatively lower levels of stigma and higher levels of knowledge than the medical/surgical nurses. Nursing speciality and personal contact with mental illness were the most significant predictors of stigma and knowledge. Knowledge was found to partially mediate the relationship between nursing speciality and stigma. We found support for the 'contact hypothesis', that is, having a personal experience of mental illness or a friend or family member who has a mental illness is associated with lower stigma towards mental illness. These findings support the development of contact-based and educational anti-stigma interventions for nurses in order to reduce stigma towards mental illness.

Is the Product Method More Efficient Than the Difference Method for Assessing Mediation?

Mediation analysis is widely used in biomedical research to quantify the extent to which the effect from an exposure on a health outcome is through a mediator and the extent to which the effect is direct. A traditional approach for quantifying mediation is through the difference method. The other popular approach uses a counterfactual framework from which the product method arises. However, there is little prior work to articulate which method is more efficient for estimating 2 key quantities in mediation analysis, the natural indirect effect and mediation proportion. To fill in this gap, we investigated the asymptotic relative efficiency for mediation measure estimators given by the product method and the difference method. We considered 4 data types characterized by continuous and binary mediators and outcomes. Under certain conditions, we show analytically that the product method is equally efficient to the difference method, or more efficient. However, our numerical studies demonstrate that the difference method is usually at least 90% as efficient as the product method under realistic scenarios in epidemiologic research, especially for estimating the mediation proportion. We demonstrate the efficiency results by analyzing the MaxART study (Eswatini, 2014-2017), which aimed to evaluate the effectiveness of the early access to antiretroviral therapy among human immunodeficiency virus-positive patients.

An examination of mediation by DNA methylation on birthweight differences induced by assisted reproductive technologies

BACKGROUND

Children born after assisted reproductive technologies (ART) differ in birthweight from those naturally conceived. It has been hypothesized that this might be explained by epigenetic mechanisms. We examined whether cord blood DNA methylation mediated the birthweight difference between 890 newborns conceived by ART (764 by fresh embryo transfer and 126 frozen thawed embryo transfer) and 983 naturally conceived newborns from the Norwegian Mother, Father, and Child Cohort Study (MoBa). DNA methylation was measured by the Illumina Infinium MethylationEPIC array. We conducted mediation analyses to assess whether differentially methylated CpGs mediated the differences in birthweight observed between: (1) fresh embryo transfer and natural conception and (2) frozen and fresh embryo transfer.

RESULTS

We observed a difference in birthweight between fresh embryo transfer and naturally conceived offspring of - 120 g. 44% (95% confidence interval [CI] 26% to 81%) of this difference in birthweight between fresh embryo transfer and naturally conceived offspring was explained by differences in methylation levels at four CpGs near LOXL1, CDH20, and DRC1. DNA methylation differences at two CpGs near PTGS1 and RASGRP4 jointly mediated 22% (95% CI 8.1% to 50.3%) of the birthweight differences between fresh and frozen embryo transfer.

CONCLUSION

Our findings suggest that DNA methylation is an important mechanism in explaining birthweight differences according to the mode of conception. Further research should examine how gene regulation at these loci influences fetal growth.

Longitudinal associations of mid-life employment status with impaired physical function in the Study of Women's Health Across the Nation

PURPOSE

This study examined whether employment status during mid-life and older adulthood is associated with physical function impairment.

METHODS

Participants were 2700 women in the multiracial/multiethnic Study of Women's Health Across the Nation. Time-varying, lagged, and cumulative exposure analyses modeled associations between self-reported employment status and the likelihood of severe physical function impairment across 19 years of follow-up.

RESULTS

Independent of demographic variables, women who were not working (OR = 1.58, 95% CI = 1.22, 2.04) or employed part-time (OR = 1.29, 95% CI = 1.04, 1.61) were more likely to report severe physical function impairments than women employed full-time. This same pattern was seen in lagged analyses predicting risk of physical function impairment from employment status at the prior assessment (not working vs. full-time: OR = 1.53, 95% CI = 1.08, 2.18; part-time vs. full-time: OR = 1.53, 95% CI = 1.17, 2.00). The likelihood of severe physical function impairment increased by 20% for every additional 10% of follow-up spent not working (OR = 1.02, 95% CI: 1.01, 1.03). Associations were robust to adjustment for health-related variables, body mass index, and physical activity.

CONCLUSIONS

Women with lower levels of employment from mid-life to older adulthood were more likely to experience severe impairment in physical function. However, the underlying mechanisms, and the timescales over which associations unfold, require further study.

Asthma Knowledge, Self-Efficacy, and Self-Management Among Rural Adolescents with Poorly Controlled Asthma

Rural adolescents with asthma are a disparate group. Self-management is essential to asthma control. We describe asthma knowledge, self-efficacy, and self-management behaviors among 198 rural adolescents with poorly controlled asthma, exploring demographic differences; we also test the application of Social Cognitive Theory to asthma self-management examining if self-efficacy mediates associations between knowledge and self-management. Asthma knowledge and self-management were relatively poor in our sample, particularly among male and White adolescents; greater knowledge was significantly associated with better symptom prevention and management. Self-efficacy partially mediated the association between knowledge and symptom prevention, but not acute symptom management, suggesting that knowledge may not improve symptom prevention behaviors without confidence to implement such behaviors and that factors beyond knowledge and self-efficacy likely play a role in asthma self-management in this population. Addressing asthma knowledge and self-efficacy could improve self-management and, ultimately, enhance asthma control among rural adolescents with poorly controlled asthma.

Challenges in Systematic Reviews and Meta-Analyses of Mediation Analyses

Systematic reviews and meta-analyses of mediation studies are increasingly being implemented in practice. Nonetheless, the methodology for conducting such review and analysis is still in a development phase, with much room for improvement. In this paper, we highlight and discuss challenges that investigators face in systematic reviews and meta-analyses of mediation studies and propose ways of accommodating these in practice.

Estimating the natural indirect effect and the mediation proportion via the product method

Background

The natural indirect effect (NIE) and mediation proportion (MP) are two measures of primary interest in mediation analysis. The standard approach for mediation analysis is through the product method, which involves a model for the outcome conditional on the mediator and exposure and another model describing the exposure–mediator relationship. The purpose of this article is to comprehensively develop and investigate the finite-sample performance of NIE and MP estimators via the product method.

Methods

With four common data types with a continuous/binary outcome and a continuous/binary mediator, we propose closed-form interval estimators for NIE and MP via the theory of multivariate delta method, and evaluate its empirical performance relative to the bootstrap approach. In addition, we have observed that the rare outcome assumption is frequently invoked to approximate the NIE and MP with a binary outcome, although this approximation may lead to non-negligible bias when the outcome is common. We therefore introduce the exact expressions for NIE and MP with a binary outcome without the rare outcome assumption and compare its performance with the approximate estimators.

Results

Simulation studies suggest that the proposed interval estimator provides satisfactory coverage when the sample size ≥500 for the scenarios with a continuous outcome and sample size ≥20,000 and number of cases ≥500 for the scenarios with a binary outcome. In the binary outcome scenarios, the approximate estimators based on the rare outcome assumption worked well when outcome prevalence less than 5% but could lead to substantial bias when the outcome is common; in contrast, the exact estimators always perform well under all outcome prevalences considered.

Conclusions

Under samples sizes commonly encountered in epidemiology and public health research, the proposed interval estimator is valid for constructing confidence interval. For a binary outcome, the exact estimator without the rare outcome assumption is more robust and stable to estimate NIE and MP. An R package mediateP is developed to implement the methods for point and variance estimation discussed in this paper.

Supplementary Information

The online version contains supplementary material available at (10.1186/s12874-021-01425-4).

Mediation analysis methods used in observational research: a scoping review and recommendations

Background

Mediation analysis methodology underwent many advancements throughout the years, with the most recent and important advancement being the development of causal mediation analysis based on the counterfactual framework. However, a previous review showed that for experimental studies the uptake of causal mediation analysis remains low. The aim of this paper is to review the methodological characteristics of mediation analyses performed in observational epidemiologic studies published between 2015 and 2019 and to provide recommendations for the application of mediation analysis in future studies.

Methods

We searched the MEDLINE and EMBASE databases for observational epidemiologic studies published between 2015 and 2019 in which mediation analysis was applied as one of the primary analysis methods. Information was extracted on the characteristics of the mediation model and the applied mediation analysis method.

Results

We included 174 studies, most of which applied traditional mediation analysis methods (n = 123, 70.7%). Causal mediation analysis was not often used to analyze more complicated mediation models, such as multiple mediator models. Most studies adjusted their analyses for measured confounders, but did not perform sensitivity analyses for unmeasured confounders and did not assess the presence of an exposure-mediator interaction.

Conclusions

To ensure a causal interpretation of the effect estimates in the mediation model, we recommend that researchers use causal mediation analysis and assess the plausibility of the causal assumptions. The uptake of causal mediation analysis can be enhanced through tutorial papers that demonstrate the application of causal mediation analysis, and through the development of software packages that facilitate the causal mediation analysis of relatively complicated mediation models.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12874-021-01426-3.

Occupational socioeconomic risk associations for head and neck cancer in Europe and South America: individual participant data analysis of pooled case-control studies within the INHANCE Consortium

BACKGROUND

The association between socioeconomic disadvantage (low education and/or income) and head and neck cancer is well established, with smoking and alcohol consumption explaining up to three-quarters of the risk. We aimed to investigate the nature of and explanations for head and neck cancer risk associated with occupational socioeconomic prestige (a perceptual measure of psychosocial status), occupational socioeconomic position and manual-work experience, and to assess the potential explanatory role of occupational exposures.

METHODS

Pooled analysis included 5818 patients with head and neck cancer (and 7326 control participants) from five studies in Europe and South America. Lifetime job histories were coded to: (1) occupational social prestige-Treiman's Standard International Occupational Prestige Scale (SIOPS); (2) occupational socioeconomic position-International Socio-Economic Index (ISEI); and (3) manual/non-manual jobs.

RESULTS

For the longest held job, adjusting for smoking, alcohol and nature of occupation, increased head and neck cancer risk estimates were observed for low SIOPS OR=1.88 (95% CI: 1.64 to 2.17), low ISEI OR=1.74 (95% CI: 1.51 to 1.99) and manual occupations OR=1.49 (95% CI: 1.35 to 1.64). Following mutual adjustment by socioeconomic exposures, risk associated with low SIOPS remained OR=1.59 (95% CI: 1.30 to 1.94).

CONCLUSIONS

These findings indicate that low occupational socioeconomic prestige, position and manual work are associated with head and neck cancer, and such risks are only partly explained by smoking, alcohol and occupational exposures. Perceptual occupational psychosocial status (SIOPS) appears to be the strongest socioeconomic factor, relative to socioeconomic position and manual/non-manual work.

Metabolic profiles of socio-economic position: a multi-cohort analysis

Background

Low socio-economic position (SEP) is a risk factor for multiple health outcomes, but its molecular imprints in the body remain unclear.

Methods

We examined SEP as a determinant of serum nuclear magnetic resonance metabolic profiles in ∼30 000 adults and 4000 children across 10 UK and Finnish cohort studies.

Results

In risk-factor-adjusted analysis of 233 metabolic measures, low educational attainment was associated with 37 measures including higher levels of triglycerides in small high-density lipoproteins (HDL) and lower levels of docosahexaenoic acid (DHA), omega-3 fatty acids, apolipoprotein A1, large and very large HDL particles (including levels of their respective lipid constituents) and cholesterol measures across different density lipoproteins. Among adults whose father worked in manual occupations, associations with apolipoprotein A1, large and very large HDL particles and HDL-2 cholesterol remained after adjustment for SEP in later life. Among manual workers, levels of glutamine were higher compared with non-manual workers. All three indicators of low SEP were associated with lower DHA, omega-3 fatty acids and HDL diameter. At all ages, children of manual workers had lower levels of DHA as a proportion of total fatty acids.

Conclusions

Our work indicates that social and economic factors have a measurable impact on human physiology. Lower SEP was independently associated with a generally unfavourable metabolic profile, consistent across ages and cohorts. The metabolites we found to be associated with SEP, including DHA, are known to predict cardiovascular disease and cognitive decline in later life and may contribute to health inequalities.

Social inequality in type 2 diabetes mellitus in the Faroe Islands: a cross-sectional study

Aims: The Faroe Islands is considered a homogeneous society and has a low Gini coefficient, but the knowledge about the social distribution of health and disease is sparse. In a large population-based sample we investigated: (a) the association between socioeconomic position defined by level of education and the prevalence of type 2 diabetes mellitus by self-report in the Faroe Islands; and (b) to what degree lifestyle factors mediate the association. Methods: We used cross-sectional data from the population-based Public Health Survey Faroes 2015 (n=1095). We present odds ratios for type 2 diabetes mellitus by socioeconomic position from logistic regression models. In our main model we adjusted for potential confounders and in a secondary model we additionally adjusted for potential mediating lifestyle factors. Results: Individuals with middle and low levels of education display higher odds ratios of type 2 diabetes mellitus of 2.80 (95% confidence interval 1.32-5.92) and 4.65 (95% confidence interval 1.93-11.17) in adjusted analysis, respectively, compared to their counterparts with high education. After adjustment for potentially mediating lifestyle factors the estimates were attenuated slightly, but a significant statistical association remained, with lifestyle-related mediating factors in total explaining 21% for middle education and 34% for low education participants. Conclusions: Our results demonstrate that there may be a social gradient in the distribution of type 2 diabetes mellitus in the Faroe Islands, and that the association is partly mediated by lifestyle factors.

Introduction to Mediation Analysis and Examples of Its Application to Real-world Data

Traditional epidemiological assessments, which mainly focused on evaluating the statistical association between two major components-the exposure and outcome-have recently evolved to ascertain the in-between process, which can explain the underlying causal pathway. Mediation analysis has emerged as a compelling method to disentangle the complex nature of these pathways. The statistical method of mediation analysis has evolved from simple regression analysis to causal mediation analysis, and each amendment refined the underlying mathematical theory and required assumptions. This short guide will introduce the basic statistical framework and assumptions of both traditional and modern mediation analyses, providing examples conducted with real-world data.

Targeting continuity of care and polypharmacy to reduce drug-drug interaction

Drug–drug interaction (DDI) is common among the elderly, and it can have detrimental effects on patients. However, how DDI can be targeted has been under-researched. This study investigates whether DDI can be reduced by targeting continuity of care (COC) through reducing polypharmacy. Population claims data of Taiwan National Health Insurance were used to conduct a 7-year-long longitudinal study on patients aged ≥ 65 years (n = 2,318,766). Mediation analysis with counterfactual method and a 4-way decomposition of the effect of COC on DDI was conducted. Mediation effect through excessive polypharmacy differed from that through lower-level polypharmacy. Compared with the low COC group, the high COC group demonstrated reduced excess relative risk of DDI by 26% (excess relative risk =  − 0.263; 95% Confidence Interval (CI) =  − 0.263 to − 0.259) to 30% (excess relative risk =  − 0.297; 95% CI =  − 0.300 to − 0.295) with excessive polypharmacy as the mediator. The risk only reduced by 8% (excess relative risk =  − 0.079; 95% CI, − 0.08 to − 0.078) to 10% (excess relative risk =  − 0.096; 95% CI, − 0.097 to − 0.095) when the mediator was changed to lower-level polypharmacy. The effect of COC on DDI was mediated by polypharmacy, and the mediation effect was higher with excessive polypharmacy. Therefore, to reduce DDI in the elderly population, different policy interventions should be designed by considering polypharmacy levels to maximize the positive effect of COC on DDI.

Invited Commentary: Frontiers of Power Assessment in Mediation Analysis

The development of tools for power and sample-size calculations for mediation analysis has lagged far behind the development of methods. The accompanying paper by Rudolph et al. (Am J Epidemiol. 2020;189(12):1559-1567) is a helpful contribution in using simulations as a tool for power calculations for more complex methods and settings. Much work remains to be done in the development of easy-to-use packages and simple online websites for carrying out power and sample-size calculations for mediation analysis. Much remains to be learned with respect to the relative power of different methods in different settings. There will likely be feedback between these 2 important frontiers of the tools we have available and of our understanding of power when conducting mediation analysis.

Association of asthma severity and educational attainment at age 6-7 years in a birth cohort: population-based record-linkage study

Background

There is conflicting research about the association between asthma and poor educational attainment that may be due to asthma definitions. Our study creates seven categories of current chronic and acute asthma to investigate if there is an association for poorer educational attainment at age 6–7 years, and the role of respiratory infections and school absence.

Methods

This study used a population-based electronic cross-sectional birth cohort 1998–2005, in Wales, UK, using health and education administrative datasets. Current asthma or wheeze categories were developed using clinical management guidelines in general practice (GP) data, acute asthma was inpatient hospital admissions and respiratory infections were the count of GP visits, from birth to age 6–7 years. We used multilevel logistic regression grouped by schools to ascertain if asthma or wheeze was associated with not attaining the expected level in teacher assessment at Key Stage 1 (KS1) adjusting for sociodemographics, perinatal, other respiratory illness and school characteristics. We tested if absence from school was a mediator in this relationship using the difference method.

Results

There were 85 906 children in this population representative cohort with 7-year follow-up. In adjusted multilevel logistic regression, only asthma inpatient hospital admission was associated with increased risk for not attaining the expected level at KS1 (adjusted OR 1.14 95% CI (1.02 to 1.27)). Lower respiratory tract infection (LRTI) GP contacts remained an independent predictor for not attaining the expected level of education. Absence from school was a potential mediator of the association between hospital admission and educational attainment.

Conclusions

Clinicians and educators need to be aware that children who have inpatient hospital admissions for asthma or wheeze, or repeated LRTI, may require additional educational support for their educational outcomes.

Direct effect and indirect effect on an outcome under nonlinear modeling

Exact formulae relating parameters in conditional and reduced generalized linear models are introduced where the reduced model omits a continuous mediator from the conditional model. For certain link functions including logit, the natural direct effect and the natural indirect effect of the counterfactual method are smaller in magnitude than, respectively, the direct effect used by the difference method and the indirect effect by the product method. Contrary to what is implicitly assumed in Jiang and VanderWeele [11] for logit link, the total effect of the counterfactual method and the total effect used for the difference method are generally not the same. They are equal to each other only under special situations. For accelerated failure time models the difference method and the product method are equivalent regardless of censoring or not, a result stated in VanderWeele [6] in the absence of censorship but proved in a misleading manner. For proportional hazards models, maximum likelihood analysis indicates that these two methods can be equivalent in the absence of censorship. In the case of logit link, one can focus on the treatment effect on the marginalized odds instead of the odds of the marginalized event so that the product method would be equivalent to the difference method. Similarly, for the proportional hazards model, one can focus on the treatment effect on the marginalized hazards instead of the hazards for the reduced model.

Severity of Hypertension Mediates the Association of Hyperuricemia With Stroke in the REGARDS Case Cohort Study

Previous studies do not widely support hyperuricemia as a risk factor for stroke and other cardiovascular diseases. We assessed the relationship between hyperuricemia and ischemic stroke (≈900 cases) using a large data set from the REGARDS study (Reasons for Geographic and Racial Differences in Stroke). We employed a case-cohort design (incident stroke cases and randomly selected cohort participants) and weighted Cox-proportional hazard models to estimate the association of serum urate level ≥6.8 mg/dL (ie, hyperuricemia) and 6.0 to <6.8 mg/dL versus <6.0 mg/dL (reference) with incident stroke. Analyses were stratified by race, gender, and age. Mediation of cardiovascular disease comorbidities on the serum urate-stroke association was tested. Hyperuricemia was associated with stroke (hazard ratio, 1.40 [95% CI, 1.10-1.78]) after adjustment for demographic variables and systolic and diastolic blood pressure. This association was substantially attenuated (hazard ratio, 1.17 [95% CI, 0.90-1.51]) by additional covariate adjustment. In particular, apparent treatment-resistant hypertension (systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg on 3 antihypertensive medications or use of ≥4 antihypertensive medications) and the count of antihypertensive medication classes significantly reduced the effect of hyperuricemia on ischemic stroke. Specifically, apparent treatment-resistant hypertension and number of antihypertensive, respectively, mediate 45% and 43% of the association. There was no effect modification in the association between hyperuricemia and stroke by age, race, or gender. We conclude that hyperuricemia may be a risk factor for stroke. The substantial attenuation of this association by apparent treatment-resistant hypertension and number of antihypertensive suggests that severe hypertension may be a mediator.

Intention to Retire in Employees over 50 Years. What is the Role of Work Ability and Work Life Satisfaction?

Background: We investigated work ability and trajectories of work life satisfaction (WLS) as predictors of intention to retire (ITR) before the statutory age. Methods: Participants were Finnish postal service employees, who responded to surveys in 2016 and 2018 (n = 1466). Survey measures included ITR, work ability and WLS. Mixture modelling was used to identify trajectories of WLS. A generalized linear model was used to determine the measures of association (Risk Ratios, RR; 95% Confidence Intervals, CI) between exposures (work ability and WLS) and ITR. Results: Approximately 40% of respondents indicated ITR. Four distinct trajectories of WLS were identified: high (33%), moderate (35%), decreasing (23%) and low (9%). Participants with poor work ability (RR 1.79, 95% CI 1.40–2.29) and decreasing WLS (1.29, 1.13–1.46) were more likely to indicate an ITR early compared to the participants with excellent/good work ability and high WLS. Job control mediated the relationship between ITR and work ability (9.3%) and WLS (14.7%). Job support also played a similar role (14% and 20.6%). Conclusions: Work ability and WLS are important contributors to the retirement intentions of employees. Ensuring workers have appropriate support and control over their work are mechanisms through which organisations may encourage employees to remain at work for longer.

Does smoking during pregnancy mediate educational disparities in preterm delivery? Findings from three large birth cohorts

BACKGROUND

Socio-economic disparities in preterm delivery have often been attributed to socially patterned smoking habits. However, most existing studies have used methods that potentially give biased estimates of the mediating effect of smoking. We used a contemporary mediation approach to study to which extent smoking during pregnancy mediates educational disparities in preterm delivery.

METHODS

We performed a comparative analysis of data from three large birth cohort studies: the Danish National Birth Cohort (DNBC), the Dutch Generation R Study, and the Norwegian Mother and Child Cohort Study (MoBa). Risk of preterm delivery by maternal education is reported as risk differences and decomposed into a part explained by smoking and a part explained by other pathways.

RESULTS

Proportions of preterm singleton deliveries were 4.8%-4.9% in all three cohorts. Total effects of maternal education were 2.0 (95% confidence interval [CI] 1.4, 2.5), 3.2 (95% CI 0.8, 5.2) and 2.0 (95% CI 0.9, 3.0) excess preterm deliveries per 100 singleton deliveries in DNBC, Generation R and MoBa when comparing primary/lower secondary education to an academic degree or equivalent. Smoking mediated, respectively, 22%, 10% and 19% of the excess risk in the DNBC, Generation R and MoBa cohorts. Adjustment for potential misclassification of smoking only increased mediated proportions slightly.

CONCLUSIONS

Smoking during pregnancy explains part of educational disparities in preterm delivery, but the mediated proportion depends on the educational gradient in smoking, emphasising that educational disparities in preterm birth may be mediated by different risk factors in different countries.

The prospective relationship between social cohesion and depressive symptoms among older adults from Central and Eastern Europe

Background

Social cohesion has a potential protective effect against depression, but evidence for Central and Eastern Europe is lacking. We investigated the prospective association between social cohesion and elevated depressive symptoms in the Czech Republic, Russia and Poland, and assessed whether alcohol drinking and smoking mediated this association.

Methods

Cohort data from 15 438 older urban participants from the Health, Alcohol and Psychosocial factors In Eastern Europe project were analysed. Baseline social cohesion was measured by five questions, and depressive symptoms were measured 3 years later by the 10-item Center for Epidemiological Depression (CES-D) Scale. Nested logistic regression models estimated ORs of elevated depressive symptoms (CES-D 10 score ≥4) by z-scores and tertiles of social cohesion.

Results

Per 1 SD decrease in social cohesion score, adjusted ORs of elevated depressive symptoms were 1.13 (95% CI 1.05 to 1.23) and 1.05 (95% CI 0.99 to 1.13) in men and women, respectively. Further adjustment for smoking and drinking did not attenuate these associations in either men (OR=1.13, 95% CI 1.05 to 1.22) or women (OR=1.05, 95% CI 0.99 to 1.13). Similarly, the fully adjusted ORs comparing the lowest versus highest social cohesion tertile were 1.33 (95% CI 1.10 to 1.62) in men and 1.18 (95% CI 1.01 to 1.39) in women.

Conclusions

Lower levels of social cohesion was associated with heightened depressive symptoms after a 3-year follow-up among older Czech, Russian and Polish adults. These effects appeared stronger in men, and alcohol and smoking played no appreciable role in this association.

Religiously or Spiritually-Motivated Forgiveness and Subsequent Health and Well-Being among Young Adults: An Outcome-Wide Analysis

This study performs an outcome-wide analysis to prospectively examine the associations of forgiveness (including forgiveness of others, self-forgiveness and divine forgiveness) with a range of psychosocial, mental, behavioral and physical health outcomes. Data from the Nurses' Health Study II and the Growing Up Today Study (Ns ranged from 5,246 to 6,994, depending on forgiveness type and outcome) with 3 or 6 years of follow-up were analyzed using generalized estimating equations. Bonferroni correction was used to correct for multiple testing. All models controlled for sociodemographic characteristics, prior religious service attendance, prior maternal attachment and prior values of the outcome variables. All forgiveness measures were positively associated with all psychosocial well-being outcomes, and inversely associated with depressive or anxiety symptoms. There was little association between forgiveness and behavioral or physical health outcomes. Forgiveness may be understood as a good itself, and may also lead to better subsequent mental health and psychosocial well-being.

Impact of declining exposure to secondhand tobacco smoke in public places to decreasing smoking-related cancer mortality in the US population

BACKGROUND

The major decrease in exposure to secondhand smoke (SHS) in public places in recent decades could have contributed to the decline in smoking-related cancer mortality observed in the US population.

METHODS

Prospective study among 11,856 non-smoking adults aged ≥40 years who participated in NHANES 1988-1994 or 1999-2004 and were followed for mortality through 2006. We estimated the amount of change in cancer mortality over time attributed to the intermediate pathway of changes in SHS exposure in public places, after adjustment for risk factors and SHS exposure at home.

RESULTS

The adjusted smoking-related cancer mortality rate ratios (95% CI) for a two-fold increase in serum cotinine and a 1-hour increase in occupational SHS exposure time were 1.10 (1.03, 1.17) and 1.14 (1.06, 1.24) for all-cancer, and 1.13 (1.03, 1.24) and 1.14 (1.02, 1.26) for smoking-related cancer, respectively. The absolute reduction in mortality comparing 1999-2004 to 1988-1994 was 75.8 (-25.5, 177.0) and 77.0 (2.6, 151.4) deaths/100,000 person-years, for all-cancer and smoking-related cancer, respectively. Among these avoided all-cancer deaths, 45.8 (2.8, 89.5) and 18.1 (-1.2, 39.6)/100,000 person-year were attributable to changes in serum cotinine concentrations and occupational SHS exposure time, respectively. The corresponding numbers of smoking-related cancer avoided deaths were 36.4 (0.7, 72.8) and 9.9 (-3.8, 24.9)/100,000 person-year.

CONCLUSIONS

Declines in SHS exposure were associated with reductions in all-cancer and smoking-related cancer mortality, supporting that smoking bans in public places may have reduced cancer mortality among non-smoking adults.

Unequal Exposure or Unequal Vulnerability? Contributions of Neighborhood Conditions and Cardiovascular Risk Factors to Socioeconomic Inequality in Incident Cardiovascular Disease in the Multi-Ethnic Study of Atherosclerosis

Risk factors can drive socioeconomic inequalities in cardiovascular disease (CVD) through differential exposure and differential vulnerability. In this paper, we show how econometric decomposition directly enables simultaneous, policy-oriented assessment of these 2 mechanisms. We specifically estimate contributions of neighborhood environment and proximal risk factors to socioeconomic inequality in CVD incidence via these mechanisms. We followed 5,608 participants in the Multi-Ethnic Study of Atherosclerosis (2000-2012) to their first CVD event (median length of follow-up, 12.2 years). We used a summary measure of baseline socioeconomic position (SEP). Covariates included baseline demographics, neighborhood characteristics, and psychosocial, behavioral, and biomedical risk factors. Using Poisson models, we decomposed the difference (inequality) in incidence rates between low- and high-SEP groups into contributions of 1) differences in covariate means (differential exposure) and 2) differences in CVD risk associated with covariates (differential vulnerability). Notwithstanding large uncertainty in neighborhood estimates, our analysis suggested that differential exposure to poorer neighborhood socioeconomic conditions, adverse social environment, diabetes, and hypertension accounted for most of the inequality. Psychosocial and behavioral contributions were negligible. Further, neighborhood SEP, female sex, and white race were more strongly associated with CVD among low-SEP (vs. high-SEP) participants. These differentials in vulnerability also accounted for nontrivial portions of the inequality and could have important implications for intervention.

Declining exposures to lead and cadmium contribute to explaining the reduction of cardiovascular mortality in the US population, 1988-2004

Background

Lead and cadmium exposures have markedly declined in the USA following the implementation of large-scale public health policies and could have contributed to the unexplained decline in cardiovascular mortality in US adults. We evaluated the potential contribution of lead and cadmium exposure reductions to explain decreasing cardiovascular mortality trends occurring in the USA from 1988-94 to 1999-2004.

Methods

Prospective study in 15 421 adults ≥40 years old who had participated in the National Health and Nutrition Examination Survey 1988-94 or 1999-2004. We estimated the amount of change in cardiovascular mortality over time that can be independently attributed to the intermediate pathway of changes in blood lead and urine cadmium concentrations.

Results

There was a 42.0% decrease in blood lead and a 31.0% decrease in urine cadmium concentrations. The cardiovascular mortality rate ratio [95% confidence intervals (CIs)] associated with a doubling of metal levels was 1.19 (1.07, 1.31) for blood lead and 1.20 (1.09, 1.32) for urine cadmium. The absolute reduction in cardiovascular deaths comparing 1999-2004 to 1988-94 was 230.7 deaths/100 000 person-years, in models adjusted for traditional cardiovascular risk factors. Among these avoided deaths, 52.0 (95% CI 8.4, 96.7) and 19.4 (4.3, 36.4) deaths/100 000 person-years were attributable to changes in lead and cadmium, respectively.

Conclusions

Environmental declines in lead and cadmium exposures were associated with reductions in cardiovascular mortality in US adults. Given the fact that lead and cadmium remain associated with cardiovascular disease at relatively low levels of exposure, prevention strategies that further minimize exposure to lead and cadmium may be needed.

Estimation and Inference for the Mediation Proportion

In epidemiology, public health and social science, mediation analysis is often undertaken to investigate the extent to which the effect of a risk factor on an outcome of interest is mediated by other covariates. A pivotal quantity of interest in such an analysis is the mediation proportion. A common method for estimating it, termed the "difference method", compares estimates from models with and without the hypothesized mediator. However, rigorous methodology for estimation and statistical inference for this quantity has not previously been available. We formulated the problem for the Cox model and generalized linear models, and utilize a data duplication algorithm together with a generalized estimation equations approach for estimating the mediation proportion and its variance. We further considered the assumption that the same link function hold for the marginal and conditional models, a property which we term "g-linkability". We show that our approach is valid whenever g-linkability holds, exactly or approximately, and present results from an extensive simulation study to explore finite sample properties. The methodology is illustrated by an analysis of pre-menopausal breast cancer incidence in the Nurses' Health Study. User-friendly publicly available software implementing those methods can be downloaded from the last author's website (SAS) or from CRAN (R).

Pathways between Socioeconomic Disadvantage and Childhood Growth in the Scottish Longitudinal Study, 1991-2001

Socioeconomically disadvantaged children are more likely to be of shorter stature and overweight, leading to greater risk of obesity in adulthood. Disentangling the mediatory pathways between socioeconomic disadvantage and childhood size may help in the development of appropriate policies aimed at reducing these health inequalities. We aimed to elucidate the putative mediatory role of birth weight using a representative sample of the Scottish population born 1991-2001 (n = 16,628). Estimated height and overweight/obesity at age 4.5 years were related to three measures of socioeconomic disadvantage (mother's education, Scottish Index of Multiple Deprivation, synthetic weekly income). Mediation was examined using two approaches: a 'traditional' mediation analysis and a counterfactual-based mediation analysis. Both analyses identified a negative effect of each measure of socioeconomic disadvantage on height, mediated to some extent by birth weight, and a positive 'direct effect' of mother's education and Scottish Index of Multiple Deprivation on overweight/obesity, which was partly counterbalanced by a negative 'indirect effect'. The extent of mediation estimated when adopting the traditional approach was greater than when adopting the counterfactual-based approach because of inappropriate handling of intermediate confounding in the former. Our findings suggest that higher birth weight in more disadvantaged groups is associated with reduced social inequalities in height but also with increased inequalities in overweight/obesity.

Relationship between mediation analysis and the structured life course approach

Many questions in life course epidemiology involve mediation and/or interaction because of the long latency period between exposures and outcomes. In this paper, we explore how mediation analysis (based on counterfactual theory and implemented using conventional regression approaches) links with a structured approach to selecting life course hypotheses. Using theory and simulated data, we show how the alternative life course hypotheses assessed in the structured life course approach correspond to different combinations of mediation and interaction parameters. For example, an early life critical period model corresponds to a direct effect of the early life exposure, but no indirect effect via the mediator and no interaction between the early life exposure and the mediator. We also compare these methods using an illustrative real-data example using data on parental occupational social class (early life exposure), own adult occupational social class (mediator) and physical capability (outcome).

Peak expiratory flow mediates the relationship between handgrip strength and timed up and go performance in elderly women, but not men

OBJECTIVE

The aim of the present study was to verify if there is sex difference in the associations among handgrip strength, peak expiratory flow (PEF) and timed up and go (TUG) test results.

METHODS

The sample included 288 consecutive elderly men (n=93) and women (n=195). Functional capacity was measured using the TUG test, and muscle strength was measured based on handgrip. Moreover, as a measure of current health status, PEF was evaluated. Linear regression procedures were performed to analyze the relationships between handgrip and both PEF and TUG test results, with adjustment for confounders, and to identify the possible mediating role of PEF in the association between handgrip strength and TUG test results.

RESULTS

In men, handgrip strength was associated with both PEF and TUG performance (p<0.01). After adjustment for PEF, the relationship between handgrip strength and TUG performance remained significant. In women, handgrip strength was also associated with both PEF and TUG performance (p<0.01). However, after adjustment for PEF, the relationship between handgrip strength and TUG performance was no longer significant.

CONCLUSION

Mobility in the elderly is sex dependent. In particular, PEF mediates the relationship between handgrip strength and TUG performance in women, but not in men.

Quantifying the mediating effect of body mass index on the relation between a Mediterranean diet and development of maternal pregnancy complications: the Australian Longitudinal Study on Women's Health

BACKGROUND

The contribution of body mass index (BMI) to the observed associations between dietary patterns and risk of gestational diabetes mellitus (GDM) and hypertensive disorders of pregnancy (HDP) remains unclear.

OBJECTIVE

The objective of this study was to formally quantify the mediating effect of prepregnancy BMI in these associations.

DESIGN

Women (aged 25-30 y) participating in the Australian Longitudinal Study on Women's Health were not pregnant at baseline in 2003 and reported ≥1 pregnancy up to 2012. GDM and HDP diagnoses were self-reported for each pregnancy and validated in a subset. A Mediterranean diet score was created by use of a baseline-validated food-frequency questionnaire and dichotomized to reflect low adherence (<25th percentile) and higher adherence (≥25th percentile). A causal inference framework for mediation analysis was used to estimate total, natural direct, and natural indirect effects of the prepregnancy Mediterranean diet on incident GDM and HDP and proportions mediated through prepregnancy BMI.

RESULTS

In 3378 women without a history of diabetes, 240 (7.1%) developed GDM. HDP was reported in 273 (8.6%) of 3167 women with no history of hypertension. Low adherence to the Mediterranean diet was associated with higher risk of GDM (OR: 1.35; 95% CI: 1.02, 1.60) and HDP (OR: 1.41; 95% CI: 1.18, 1.56), after adjustment for education, parity, polycystic ovary syndrome, energy intake, and physical activity. Proportions mediated through prepregnancy BMI (per 1-kg/m(2) increase) were 32% and 22% for GDM and HDP, respectively.

CONCLUSION

These findings suggest that prepregnancy BMI as a single mediator contributes substantially to the total effects of the prepregnancy Mediterranean diet on GDM and HDP risk.

Mediation Analysis for Health Disparities Research

Social epidemiologists often seek to determine the mechanisms that underlie health disparities. This work is typically based on mediation procedures that may not be justified with exposures of common interest in social epidemiology. In this analysis, we explored the consequences of using standard approaches, referred to as the difference and generalized product methods, when mediator-outcome confounders are associated with the exposure. We compared these with inverse probability-weighted marginal structural models, the structural transformation method, doubly robust g-estimation of a structural nested model, and doubly robust targeted minimum loss-based estimation. We used data on 900,726 births from 2003 to 2007 in the Penn Moms study, conducted in Pennsylvania, to assess the extent to which breastfeeding prior to hospital discharge explained the racial disparity in infant mortality. Overall, for every 1,000 births, 3.36 more infant deaths occurred among non-Hispanic black women relative to all other women (95% confidence interval: 2.78, 3.93). Using the difference and generalized product methods to assess the disparity that would remain if everyone breastfed prior to discharge suggested a complete elimination of the disparity (risk difference = -0.87 per 1,000 births; 95% confidence interval: -1.39, -0.35). In contrast, doubly robust methods suggested a reduction in the disparity to 2.45 (95% confidence interval: 2.20, 2.71) more infant deaths per 1,000 births among non-Hispanic black women. Standard approaches for mediation analysis in health disparities research can yield misleading results.

Quantifying the mediating effects of smoking and occupational exposures in the relation between education and lung cancer: the ICARE study

Smoking only partly explains the higher lung cancer incidence observed among socially deprived people. Occupational exposures may account for part of these inequalities, but this issue has been little investigated. We investigated the extent to which smoking and occupational exposures to asbestos, silica and diesel motor exhaust mediated the association between education and lung cancer incidence in men. We analyzed data from a large French population-based case-control study (1976 lung cancers, 2648 controls). Detailed information on lifelong tobacco consumption and occupational exposures to various carcinogens was collected. We conducted inverse probability-weighted marginal structural models. A strong association was observed between education and lung cancer. The indirect effect through smoking varied by educational level, with the strongest indirect effect observed for those with the lowest education (OR = 1.34 (1.14-1.57)). The indirect effect through occupational exposures was substantial among men with primary (OR = 1.22 (1.15-1.30) for asbestos and silica) or vocational secondary education (OR = 1.18 (1.12-1.25)). The contribution of smoking to educational differences in lung cancer incidence ranged from 22 % (10-34) for men with primary education to 31 % (-3 to 84) for men with a high school degree. The contribution of occupational exposures to asbestos and silica ranged from 15 % (10-20) for men with a high school degree to 20 % (13-28) for men with vocational secondary education. Our results highlight the urgent need for public health policies that aim at decreasing exposure to carcinogens at work, in addition to tobacco control policies, if we want to reduce socioeconomic inequalities in the cancer field.

Clarifying Associations between Childhood Adversity, Social Support, Behavioral Factors, and Mental Health, Health, and Well-Being in Adulthood: A Population-Based Study

Previous studies have shown that socio-demographic factors, childhood socioeconomic status (CSES), childhood traumatic experiences (CTEs), social support and behavioral factors are associated with health and well-being in adulthood. However, the relative importance of these factors for mental health, health, and well-being has not been studied. Moreover, the mechanisms by which CTEs affect mental health, health, and well-being in adulthood are not clear. Using data from a representative sample (n = 12,981) of the adult population in Tromsø, Norway, this study examines (i) the relative contribution of structural conditions (gender, age, CSES, psychological abuse, physical abuse, and substance abuse distress) to social support and behavioral factors in adulthood; (ii) the relative contribution of socio-demographic factors, CSES, CTEs, social support, and behavioral factors to three multi-item instruments of mental health (SCL-10), health (EQ-5D), and subjective well-being (SWLS) in adulthood; (iii) the impact of CTEs on mental health, health, and well-being in adulthood, and; (iv) the mediating role of adult social support and behavioral factors in these associations. Instrumental support (24.16%, p < 0.001) explained most of the variation in mental health, while gender (21.32%, p < 0.001) explained most of the variation in health, and emotional support (23.34%, p < 0.001) explained most of the variation in well-being. Psychological abuse was relatively more important for mental health (12.13%), health (7.01%), and well-being (9.09%), as compared to physical abuse, and substance abuse distress. The subjective assessment of childhood financial conditions was relatively more important for mental health (6.02%), health (10.60%), and well-being (20.60%), as compared to mother's and father's education. CTEs were relatively more important for mental health, while, CSES was relatively more important for health and well-being. Respondents exposed to all three types of CTEs had a more than two-fold increased risk of being mentally unhealthy (RR Total Effect = 2.75, 95% CI: 2.19-3.10), an 89% increased risk of being unhealthy (RR Total Effect = 1.89, 95% CI: 1.47-1.99), and a 42% increased risk of having a low level of well-being in adulthood (RR Total Effect = 1.42, 95% CI: 1.29-1.52). Social support and behavioral factors mediate 11-18% (p < 0.01) of these effects. The study advances the theoretical understanding of how CTEs influence adult mental health, health, and well-being.

Mediation Analysis: A Practitioner's Guide

This article provides an overview of recent developments in mediation analysis, that is, analyses used to assess the relative magnitude of different pathways and mechanisms by which an exposure may affect an outcome. Traditional approaches to mediation in the biomedical and social sciences are described. Attention is given to the confounding assumptions required for a causal interpretation of direct and indirect effect estimates. Methods from the causal inference literature to conduct mediation in the presence of exposure-mediator interactions, binary outcomes, binary mediators, and case-control study designs are presented. Sensitivity analysis techniques for unmeasured confounding and measurement error are introduced. Discussion is given to extensions to time-to-event outcomes and multiple mediators. Further flexible modeling strategies arising from the precise counterfactual definitions of direct and indirect effects are also described. The focus throughout is on methodology that is easily implementable in practice across a broad range of potential applications.

Invited commentary: boundless science--putting natural direct and indirect effects in a clearer empirical context

Epidemiologists are increasingly using natural effects for applied mediation analyses, yet 1 key identifying assumption is unintuitive and subject to some controversy. In this issue of the Journal, Jiang and VanderWeele (Am J Epidemiol. 2015;182(2):105-108) formalize the conditions under which the difference method can be used to estimate natural indirect effects. In this commentary, I discuss implications of the controversial "cross-worlds" independence assumption needed to identify natural effects. I argue that with a binary mediator, a simple modification of the authors' approach will provide bounds for natural direct and indirect effect estimates that better reflect the capacity of the available data to support empirical statements on the presence of mediated effects. I discuss complications encountered when odds ratios are used to decompose effects, as well as the implications of incorrectly assuming the absence of exposure-induced mediator-outcome confounders. I note that the former problem can be entirely resolved using collapsible measures of effect, such as risk ratios. In the Appendix, I use previous derivations for natural direct effect bounds on the risk difference scale to provide bounds on the odds ratio scale that accommodate 1) uncertainty due to the cross-world independence assumption and 2) uncertainty due to the cross-world independence assumption and the presence of exposure-induced mediator-outcome confounders.