Distribution-free mediation analysis for nonlinear models with confounding

Older adults' online activities and cognition: Investigating the psychological mechanisms and age and gender differences

OBJECTIVES

This study aimed to investigate psychological mechanisms underlying the association between older adults' Internet use and cognition and examine potential age and gender group differences.

METHODS

2064 older participants were extracted from the Waves 2012, 2013, and 2016 Health and Retirement Study. Internet use was measured by two sets of variables: Internet access and different types of online activities (i.e., informational use, social use, online shopping, and online banking). Path analyses were applied to test the proposed mechanisms via three mediators (i.e., loneliness, depressive symptoms, and perceived control). Multi-group analyses were conducted to examine the potential group differences.

RESULTS

Internet use was positively associated with cognition. Despite the large direct effect, small but significant indirect effects via depressive symptoms and perceived control were identified across all online activities. Multi-group analyses revealed age-group differences in the mechanisms: depressive symptoms mediated the effects of all online activities on cognition among young-old adults, while perceived control mediated all the effects among old-old adults. Gender group differences were also identified: depressive symptoms mediated the effects of all online activities on cognition among older women and most online activities among older men, whereas perceived control mediated the associations between informational and instrumental (i.e., online shopping and banking) use and cognition among older men.

DISCUSSION

This study highlights the mediating effect of depressive symptoms and perceived control and age and gender differences regarding the Internet use-cognition association. Internet-based cognitive interventions should consider these psychological mediators and age and gender differences for the best results.

Causal mediation analysis: From simple to more robust strategies for estimation of marginal natural (in)direct effects

This paper aims to provide practitioners of causal mediation analysis with a better understanding of estimation options. We take as inputs two familiar strategies (weighting and model-based prediction) and a simple way of combining them (weighted models), and show how a range of estimators can be generated, with different modeling requirements and robustness properties. The primary goal is to help build intuitive appreciation for robust estimation that is conducive to sound practice. We do this by visualizing the target estimand and the estimation strategies. A second goal is to provide a "menu" of estimators that practitioners can choose from for the estimation of marginal natural (in)direct effects. The estimators generated from this exercise include some that coincide or are similar to existing estimators and others that have not previously appeared in the literature. We note several different ways to estimate the weights for cross-world weighting based on three expressions of the weighting function, including one that is novel; and show how to check the resulting covariate and mediator balance. We use a random continuous weights bootstrap to obtain confidence intervals, and also derive general asymptotic variance formulas for the estimators. The estimators are illustrated using data from an adolescent alcohol use prevention study. R-code is provided.

The Chasm in Percutaneous Coronary Intervention and In-Hospital Mortality Rates Among Acute Myocardial Infarction Patients in Rural and Urban Hospitals in China: A Mediation Analysis

Objectives: To determine to what extent the inequality in the ability to provide percutaneous coronary intervention (PCI) translates into outcomes for AMI patients in China.

Methods: We identified 82,677 patients who had primary diagnoses of AMI and were hospitalized in Shanxi Province, China, between 2013 and 2017. We applied logistic regressions with inverse probability weighting based on propensity scores and mediation analyses to examine the association of hospital rurality with in-hospital mortality and the potential mediating effects of PCI.

Results: In multivariate models where PCI was not adjusted for, rural hospitals were associated with a significantly higher risk of in-hospital mortality (odds ratio [OR]: 1.19, 95% confidence interval [CI]: 1.03–1.37). However, this association was nullified (OR: 0.94, 95% CI: 0.81–1.08) when PCI was included as a covariate. Mediation analyses revealed that PCI significantly mediated 132.3% (95% CI: 104.1–256.6%) of the effect of hospital rurality on in-hospital mortality. The direct effect of hospital rurality on in-hospital mortality was insignificant.

Conclusion: The results highlight the need to improve rural hospitals’ infrastructure and address the inequalities of treatments and outcomes in rural and urban hospitals.

Oxygen Saturation and Postoperative Mortality in Patients With Acute Ischemic Stroke Treated by Endovascular Thrombectomy

BACKGROUND

Monitored anesthesia care (MAC) and general anesthesia (GA) with endotracheal intubation are the 2 most used techniques for patients with acute ischemic stroke (AIS) undergoing endovascular thrombectomy. We aimed to test the hypothesis that increased arterial oxygen concentration during reperfusion period is a mechanism underlying the association between use of GA (versus MAC) and increased risk of in-hospital mortality.

METHODS

In this retrospective cohort study, data were collected at the Cleveland Clinic between 2013 and 2018. To assess the potential mediation effect of time-weighted average oxygen saturation (Spo2) in first postoperative 48 hours between the association between GA versus MAC and in-hospital mortality, we assessed the association between anesthesia type and post-operative Spo2 tertiles (exposure-mediator relationship) through a cumulative logistic regression model and assessed the association between Spo2 and in-hospital mortality (mediator-outcome relationship) using logistic regression models. Confounding factors were adjusted for using propensity score methods. Both significant exposure-mediator and significant mediator-outcome relationships are needed to suggest potential mediation effect.

RESULTS

Among 358 patients included in the study, 104 (29%) patients received GA and 254 (71%) received MAC, with respective hospital mortality rate of 19% and 5% (unadjusted P value <.001). GA patients were 1.6 (1.2, 2.1) (P < .001) times more likely to have a higher Spo2 tertile as compared to MAC patients. Patients with higher Spo2 tertile had 3.8 (2.1, 6.9) times higher odds of mortality than patients with middle Spo2 tertile, while patients in the lower Spo2 tertile did not have significant higher odds compared to the middle tertile odds ratio (OR) (1.8 [0.9, 3.4]; overall P < .001). The significant exposure-mediator and mediator-outcome relationships suggest that Spo2 may be a mediator of the relationship between anesthetic method and mortality. However, the estimated direct effect of GA versus MAC on mortality (ie, after adjusting for Spo2; OR [95% confidence interval {CI}] of 2.1 [0.9-4.9]) was close to the estimated association ignoring Spo2 (OR [95% CI] of 2.2 [1.0-5.1]), neither statistically significant, suggesting that Spo2 had at most a modest mediator role.

CONCLUSIONS

GA was associated with a higher Spo2 compared to MAC among those treated by endovascular thrombectomy for AIS. Spo2 values that were higher than the middle tertile were associated with higher odds of mortality. However, GA was not significantly associated with higher odds of death. Spo2 at most constituted a modest mediator role in explaining the relationship between GA versus MAC and mortality.

Meaningful Causal Decompositions in Health Equity Research: Definition, Identification, and Estimation Through a Weighting Framework

Causal decomposition analyses can help build the evidence base for interventions that address health disparities (inequities). They ask how disparities in outcomes may change under hypothetical intervention. Through study design and assumptions, they can rule out alternate explanations such as confounding, selection bias, and measurement error, thereby identifying potential targets for intervention. Unfortunately, the literature on causal decomposition analysis and related methods have largely ignored equity concerns that actual interventionists would respect, limiting their relevance and practical value. This article addresses these concerns by explicitly considering what covariates the outcome disparity and hypothetical intervention adjust for (so-called allowable covariates) and the equity value judgments these choices convey, drawing from the bioethics, biostatistics, epidemiology, and health services research literatures. From this discussion, we generalize decomposition estimands and formulae to incorporate allowable covariate sets (and thereby reflect equity choices) while still allowing for adjustment of non-allowable covariates needed to satisfy causal assumptions. For these general formulae, we provide weighting-based estimators based on adaptations of ratio-of-mediator-probability and inverse-odds-ratio weighting. We discuss when these estimators reduce to already used estimators under certain equity value judgments, and a novel adaptation under other judgments.

Indirect Associations Between Commercial Television Exposure and Child Body Mass Index

OBJECTIVE

To formally test a hierarchy of effects model linking exposure to television (TV) advertising for unhealthy foods with child body weight through purchase requests, purchases, and consumption.

DESIGN

A nationally representative cross-sectional online study in the United Kingdom.

PARTICIPANTS

A total of 2,260 parent-child dyads (children aged 7-11 years) recruited via online research panel; 55.7% boys, mean age 8.9 ± 1.4 years, mean body mass index z-score 1.25 ± 2.1.

MAIN OUTCOME MEASURES

Parents reported on child TV exposure and child height and weight. Children self-reported their frequency of (1) pestering for advertised foods, (2) purchase of unhealthy foods, and (3) consumption of unhealthy foods.

ANALYSIS

A structural equation model was applied to data.

RESULTS

As predicted, commercial TV exposure was indirectly associated with children's body mass index through purchasing and consumption through purchase requests. It was also directly associated with children's purchase requests, purchasing, and consumption of unhealthy foods. Associations between noncommercial TV and behavior or body weight outcomes, when found, were significantly weaker than for commercial exposure.

CONCLUSIONS AND IMPLICATIONS

This study provides insight into the likely behavioral pathways underpinning the effects of food marketing on diet and potentially body weight in children. Future longitudinal analyses would provide insight into causal inferences.

Generalized causal mediation and path analysis: Extensions and practical considerations

Causal mediation analysis seeks to decompose the effect of a treatment or exposure among multiple possible paths and provide casually interpretable path-specific effect estimates. Recent advances have extended causal mediation analysis to situations with a sequence of mediators or multiple contemporaneous mediators. However, available methods still have limitations, and computational and other challenges remain. The present paper provides an extended causal mediation and path analysis methodology. The new method, implemented in the new R package, gmediation (described in a companion paper), accommodates both a sequence (two stages) of mediators and multiple mediators at each stage, and allows for multiple types of outcomes following generalized linear models. The methodology can also handle unsaturated models and clustered data. Addressing other practical issues, we provide new guidelines for the choice of a decomposition, and for the choice of a reference group multiplier for the reduction of Monte Carlo error in mediation formula computations. The new method is applied to data from a cohort study to illuminate the contribution of alternative biological and behavioral paths in the effect of socioeconomic status on dental caries in adolescence.

Mediation analysis for count and zero-inflated count data

Different conventional and causal approaches have been proposed for mediation analysis to better understand the mechanism of a treatment. Count and zero-inflated count data occur in biomedicine, economics, and social sciences. This paper considers mediation analysis for count and zero-inflated count data under the potential outcome framework with nonlinear models. When there are post-treatment confounders which are independent of, or affected by, the treatment, we first define the direct, indirect, and total effects of our interest and then discuss various conditions under which the effects of interest can be identified. Proofs are provided for the sensitivity analysis proposed in the paper. Simulation studies show that the methods work well. We apply the methods to the Detroit Dental Health Project's Motivational Interviewing DVD trial for the direct and indirect effects of motivational interviewing on count and zero-inflated count dental caries outcomes.

Mediation analysis for count and zero-inflated count data without sequential ignorability and its application in dental studies

Mediation analysis seeks to understand the mechanism by which a treatment affects an outcome. Count or zero-inflated count outcomes are common in many studies in which mediation analysis is of interest. For example, in dental studies, outcomes such as the number of decayed, missing and filled teeth are typically zero inflated. Existing mediation analysis approaches for count data often assume sequential ignorability of the mediator. This is often not plausible because the mediator is not randomized so unmeasured confounders are associated with the mediator and the outcome. We develop causal methods based on instrumental variable approaches for mediation analysis for count data possibly with many 0s that do not require the assumption of sequential ignorability. We first define the direct and indirect effect ratios for those data, and then we propose estimating equations and use empirical likelihood to estimate the direct and indirect effects consistently. A sensitivity analysis is proposed for violations of the instrumental variables exclusion restriction assumption. Simulation studies demonstrate that our method works well for different types of outcome under various settings. Our method is applied to a randomized dental caries prevention trial and a study of the effect of a massive flood in Bangladesh on children's diarrhoea.

Estimating and testing high-dimensional mediation effects in epigenetic studies

MOTIVATION

High-dimensional DNA methylation markers may mediate pathways linking environmental exposures with health outcomes. However, there is a lack of analytical methods to identify significant mediators for high-dimensional mediation analysis.

RESULTS

Based on sure independent screening and minimax concave penalty techniques, we use a joint significance test for mediation effect. We demonstrate its practical performance using Monte Carlo simulation studies and apply this method to investigate the extent to which DNA methylation markers mediate the causal pathway from smoking to reduced lung function in the Normative Aging Study. We identify 2 CpGs with significant mediation effects.

AVAILABILITY AND IMPLEMENTATION

R package, source code, and simulation study are available at https://github.com/YinanZheng/HIMA CONTACT: lei.liu@northwestern.edu.

Causal mediation analysis with a binary outcome and multiple continuous or ordinal mediators: Simulations and application to an alcohol intervention

We investigate a method to estimate the combined effect of multiple continuous/ordinal mediators on a binary outcome: 1) fit a structural equation model with probit link for the outcome and identity/probit link for continuous/ordinal mediators, 2) predict potential outcome probabilities, and 3) compute natural direct and indirect effects. Step 2 involves rescaling the latent continuous variable underlying the outcome to address residual mediator variance/covariance. We evaluate the estimation of risk-difference- and risk-ratio-based effects (RDs, RRs) using the ML, WLSMV and Bayes estimators in Mplus. Across most variations in path-coefficient and mediator-residual-correlation signs and strengths, and confounding situations investigated, the method performs well with all estimators, but favors ML/WLSMV for RDs with continuous mediators, and Bayes for RRs with ordinal mediators. Bayes outperforms WLSMV/ML regardless of mediator type when estimating RRs with small potential outcome probabilities and in two other special cases. An adolescent alcohol prevention study is used for illustration.

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.

G-computation demonstration in causal mediation analysis

Recent work has considerably advanced the definition, identification and estimation of controlled direct, and natural direct and indirect effects in causal mediation analysis. Despite the various estimation methods and statistical routines being developed, a unified approach for effect estimation under different effect decomposition scenarios is still needed for epidemiologic research. G-computation offers such unification and has been used for total effect and joint controlled direct effect estimation settings, involving different types of exposure and outcome variables. In this study, we demonstrate the utility of parametric g-computation in estimating various components of the total effect, including (1) natural direct and indirect effects, (2) standard and stochastic controlled direct effects, and (3) reference and mediated interaction effects, using Monte Carlo simulations in standard statistical software. For each study subject, we estimated their nested potential outcomes corresponding to the (mediated) effects of an intervention on the exposure wherein the mediator was allowed to attain the value it would have under a possible counterfactual exposure intervention, under a pre-specified distribution of the mediator independent of any causes, or under a fixed controlled value. A final regression of the potential outcome on the exposure intervention variable was used to compute point estimates and bootstrap was used to obtain confidence intervals. Through contrasting different potential outcomes, this analytical framework provides an intuitive way of estimating effects under the recently introduced 3- and 4-way effect decomposition. This framework can be extended to complex multivariable and longitudinal mediation settings.

Causal mediation analysis with a latent mediator

Health researchers are often interested in assessing the direct effect of a treatment or exposure on an outcome variable, as well as its indirect (or mediation) effect through an intermediate variable (or mediator). For an outcome following a nonlinear model, the mediation formula may be used to estimate causally interpretable mediation effects. This method, like others, assumes that the mediator is observed. However, as is common in structural equations modeling, we may wish to consider a latent (unobserved) mediator. We follow a potential outcomes framework and assume a generalized structural equations model (GSEM). We provide maximum-likelihood estimation of GSEM parameters using an approximate Monte Carlo EM algorithm, coupled with a mediation formula approach to estimate natural direct and indirect effects. The method relies on an untestable sequential ignorability assumption; we assess robustness to this assumption by adapting a recently proposed method for sensitivity analysis. Simulation studies show good properties of the proposed estimators in plausible scenarios. Our method is applied to a study of the effect of mother education on occurrence of adolescent dental caries, in which we examine possible mediation through latent oral health behavior.

Practical guidance for conducting mediation analysis with multiple mediators using inverse odds ratio weighting

Despite the recent flourishing of mediation analysis techniques, many modern approaches are difficult to implement or applicable to only a restricted range of regression models. This report provides practical guidance for implementing a new technique utilizing inverse odds ratio weighting (IORW) to estimate natural direct and indirect effects for mediation analyses. IORW takes advantage of the odds ratio's invariance property and condenses information on the odds ratio for the relationship between the exposure (treatment) and multiple mediators, conditional on covariates, by regressing exposure on mediators and covariates. The inverse of the covariate-adjusted exposure-mediator odds ratio association is used to weight the primary analytical regression of the outcome on treatment. The treatment coefficient in such a weighted regression estimates the natural direct effect of treatment on the outcome, and indirect effects are identified by subtracting direct effects from total effects. Weighting renders treatment and mediators independent, thereby deactivating indirect pathways of the mediators. This new mediation technique accommodates multiple discrete or continuous mediators. IORW is easily implemented and is appropriate for any standard regression model, including quantile regression and survival analysis. An empirical example is given using data from the Moving to Opportunity (1994-2002) experiment, testing whether neighborhood context mediated the effects of a housing voucher program on obesity. Relevant Stata code (StataCorp LP, College Station, Texas) is provided.

Understanding treatment effect mechanisms of the CAMBRA randomized trial in reducing caries increment

The Caries Management By Risk Assessment (CAMBRA) randomized controlled trial showed that an intervention featuring combined antibacterial and fluoride therapy significantly reduced bacterial load and suggested reduced caries increment in adults with 1 to 7 baseline cavitated teeth. While trial results speak to the overall effectiveness of an intervention, insight can be gained from understanding the mechanism by which an intervention acts on putative intermediate variables (mediators) to affect outcomes. This study conducted mediation analyses on 109 participants who completed the trial to understand whether the intervention reduced caries increment through its action on potential mediators (oral bacterial load, fluoride levels, and overall caries risk based on the composite of bacterial challenge and salivary fluoride) between the intervention and dental outcomes. The primary outcome was the increment from baseline in decayed, missing, and filled permanent surfaces (ΔDMFS) 24 mo after completing restorations for baseline cavitated lesions. Analyses adjusted for baseline overall risk, bacterial challenge, and fluoride values under a potential outcome framework using generalized linear models. Overall, the CAMBRA intervention was suggestive in reducing the 24-mo DMFS increment (reduction in ΔDMFS: -0.96; 95% confidence interval [CI]: -2.01 to 0.08; P = 0.07); the intervention significantly reduced the 12-mo overall risk (reduction in overall risk: -19%; 95% CI, -7 to -41%;], P = 0.005). Individual mediators, salivary log10 mutans streptococci, log10 lactobacilli, and fluoride level, did not represent statistically significant pathways alone through which the intervention effect was transmitted. However, 36% of the intervention effect on 24-mo DMFS increment was through a mediation effect on 12-mo overall risk (P = 0.03). These findings suggest a greater intervention effect carried through the combined action on multiple aspects of the caries process rather than through any single factor. In addition, a substantial portion of the total effect of the CAMBRA intervention may have operated through unanticipated or unmeasured pathways not included among the potential mediators studied.

Mediation misgivings: ambiguous clinical and public health interpretations of natural direct and indirect effects

Recent methodological innovation is giving rise to an increasing number of applied papers in medical and epidemiological journals in which natural direct and indirect effects are estimated. However, there is a longstanding debate on whether such effects are relevant targets of inference in population health. In light of the repeated calls for a more pragmatic and consequential epidemiology, we review three issues often raised in this debate: (i) the use of composite cross-world counterfactuals and the need for cross-world independence assumptions; (ii) interventional vs non-interventional identifiability; and (iii) the interpretational ambiguity of natural direct and indirect effect estimates. We use potential outcomes notation and directed acyclic graphs to explain 'cross-world' assumptions, illustrate implications of this assumption via regression models and discuss ensuing issues of interpretation. We argue that the debate on the relevance of natural direct and indirect effects rests on whether one takes as a target of inference the mathematical object per se, or the change in the world that the mathematical object represents. We further note that public health questions may be better served by estimating controlled direct effects.

Mediation Analysis with Multiple Mediators

Recent advances in the causal inference literature on mediation have extended traditional approaches to direct and indirect effects to settings that allow for interactions and non-linearities. In this paper, these approaches from causal inference are further extended to settings in which multiple mediators may be of interest. Two analytic approaches, one based on regression and one based on weighting are proposed to estimate the effect mediated through multiple mediators and the effects through other pathways. The approaches proposed here accommodate exposure-mediator interactions and, to a certain extent, mediator-mediator interactions as well. The methods handle binary or continuous mediators and binary, continuous or count outcomes. When the mediators affect one another, the strategy of trying to assess direct and indirect effects one mediator at a time will in general fail; the approach given in this paper can still be used. A characterization is moreover given as to when the sum of the mediated effects for multiple mediators considered separately will be equal to the mediated effect of all of the mediators considered jointly. The approach proposed in this paper is robust to unmeasured common causes of two or more mediators.