Mediation analysis using incomplete information from publicly available data sources.

Our work was motivated by the question whether, and to what extent, well-established risk factors mediate the racial disparity observed for colorectal cancer (CRC) incidence in the United States. Mediation analysis examines the relationships between an exposure, a mediator and an outcome. All available methods require access to a single complete data set with these three variables. However, because population-based studies usually include few non-White participants, these approaches have limited utility in answering our motivating question. Recently, we developed novel methods to integrate several data sets with incomplete information for mediation analysis. These methods have two limitations: (i) they only consider a single mediator and (ii) they require a data set containing individual-level data on the mediator and exposure (and possibly confounders) obtained by independent and identically distributed sampling from the target population. Here, we propose a new method for mediation analysis with several different data sets that accommodates complex survey and registry data, and allows for multiple mediators. The proposed approach yields unbiased causal effects estimates and confidence intervals with nominal coverage in simulations. We apply our method to data from U.S. cancer registries, a U.S.-population-representative survey and summary level odds-ratio estimates, to rigorously evaluate what proportion of the difference in CRC risk between non-Hispanic Whites and Blacks is mediated by three potentially modifiable risk factors (CRC screening history, body mass index, and regular aspirin use).

Climate-driven vegetation greening further reduces water availability in drylands.

Climate change alters surface water availability (WA; precipitation minus evapotranspiration, P - ET) and consequently impacts agricultural production and societal water needs, leading to increasing concerns on the sustainability of water use. Although the direct effects of climate change on WA have long been recognized and assessed, indirect climate effects occurring through adjustments in terrestrial vegetation are more subtle and not yet fully quantified. To address this knowledge gap, here we investigate the interplay between climate-induced changes in leaf area index (LAI) and ET and quantify its ultimate effect on WA during the period 1982-2016 at the global scale, using an ensemble of data-driven products and land surface models. We show that ~44% of the global vegetated land has experienced a significant increase in growing season-averaged LAI and climate change explains 33.5% of this greening signal. Such climate-induced greening has enhanced ET of 0.051 ± 0.067 mm year-2 (mean ± SD), further amplifying the ongoing increase in ET directly driven by variations in climatic factors over 36.8% of the globe, and thus exacerbating the decline in WA prominently in drylands. These findings highlight the indirect impact of positive feedbacks in the land-climate system on the decline of WA, and call for an in-depth evaluation of these phenomena in the design of local mitigation and adaptation plans.

A generalist nematode destabilises plant competition: no evidence for direct effects, but strong evidence for indirect effects on rhizobium abundance.

Difficulties quantifying pathogen load and mutualist abundance limit our ability to connect disease dynamics to host community ecology. For example, specific predictions about how differential pathogen load is hypothesised to drive host competitive outcomes are rarely tested. Additionally, although infection is known to affect mutualists, we rarely measure the magnitude of pathogen effects on mutualist abundance across host competitive contexts. We tested for both mechanisms in a plant-rhizobia-nematode system. We paired the legume Medicago lupulina with intraspecific and interspecific plant competitors, with and without a generalist nematode parasite Meloidogyne sp. Relative change in plant biomass was used to determine how nematode inoculation affected plant competitive outcomes. We counted nematode galls to test for direct effects of parasitism on plant competition and rhizobia nodules to test for indirect effects of nematode presence on rhizobium abundance. Parasites were destabilising despite similar nematode load across competition treatments. During interspecific compared with intraspecific competition, nematode inoculation decreased nodulation on M. lupulina, increased nodulation on Trifolium repens and had no effect on nodulation on Chamaecrista fasciculata. We found no support for hypothesised direct effects of nematode load on competitive outcomes and strong but idiosyncratic indirect effects of nematode inoculation on rhizobium abundance.

Nonlinear mediation analysis with high-dimensional mediators whose causal structure is unknown.

With multiple possible mediators on the causal pathway from a treatment to an outcome, we consider the problem of decomposing the effects along multiple possible causal path(s) through each distinct mediator. Under a path-specific effects framework, such fine-grained decompositions necessitate stringent assumptions, such as correctly specifying the causal structure among the mediators, and no unobserved confounding among the mediators. In contrast, interventional direct and indirect effects for multiple mediators can be identified under much weaker conditions, while providing scientifically relevant causal interpretations. Nonetheless, current estimation approaches require (correctly) specifying a model for the joint mediator distribution, which can be difficult when there is a high-dimensional set of possibly continuous and noncontinuous mediators. In this article, we avoid the need to model this distribution, by developing a definition of interventional effects previously suggested for longitudinal mediation. We propose a novel estimation strategy that uses nonparametric estimates of the (counterfactual) mediator distributions. Noncontinuous outcomes can be accommodated using nonlinear outcome models. Estimation proceeds via Monte Carlo integration. The procedure is illustrated using publicly available genomic data to assess the causal effect of a microRNA expression on the 3-month mortality of brain cancer patients that is potentially mediated by expression values of multiple genes.

Applied causal inference methods for sequential mediators.

BACKGROUND: Mediation analysis aims at estimating to what extent the effect of an exposure on an outcome is explained by a set of mediators on the causal pathway between the exposure and the outcome. The total effect of the exposure on the outcome can be decomposed into an indirect effect, i.e. the effect explained by the mediators jointly, and a direct effect, i.e. the effect unexplained by the mediators. However finer decompositions are possible in presence of independent or sequential mediators. METHODS: We review four statistical methods to analyse multiple sequential mediators, the inverse odds ratio weighting approach, the inverse probability weighting approach, the imputation approach and the extended imputation approach. These approaches are compared and implemented using a case-study with the aim to investigate the mediating role of adverse reproductive outcomes and infant respiratory infections in the effect of maternal pregnancy mental health on infant wheezing in the Ninfea birth cohort. RESULTS: Using the inverse odds ratio weighting approach, the direct effect of maternal depression or anxiety in pregnancy is equal to a 59% (95% CI: 27%,94%) increased prevalence of infant wheezing and the mediated effect through adverse reproductive outcomes is equal to a 3% (95% CI: -6%,12%) increased prevalence of infant wheezing. When including infant lower respiratory infections in the mediation pathway, the direct effect decreases to 57% (95% CI: 25%,92%) and the indirect effect increases to 5% (95% CI: -5%,15%). The estimates of the effects obtained using the weighting and the imputation approaches are similar. The extended imputation approach suggests that the small joint indirect effect through adverse reproductive outcomes and lower respiratory infections is due entirely to the contribution of infant lower respiratory infections, and not to an increased prevalence of adverse reproductive outcomes. CONCLUSIONS: The four methods revealed similar results of small mediating role of adverse reproductive outcomes and early respiratory tract infections in the effect of maternal pregnancy mental health on infant wheezing. The choice of the method depends on what is the effect of main interest, the type of the variables involved in the analysis (binary, categorical, count or continuous) and the confidence in specifying the models for the exposure, the mediators and the outcome.

Expanding the Direct and Indirect Effects Model of Writing (DIEW): Reading-Writing Relations, and Dynamic Relations As a Function of Measurement/Dimensions of Written Composition.

Within the context of the Direct and Indirect Effects model of Writing, we examined a dynamic relations hypothesis, which contends that the relations of component skills, including reading comprehension, to written composition vary as a function of dimensions of written composition. Specifically, we investigated (a) whether higher order cognitive skills (i.e., inference, perspective taking, and monitoring) are differentially related to three dimensions of written composition-writing quality, writing productivity, and correctness in writing; (b) whether reading comprehension is differentially related to the three dimensions of written composition after accounting for oral language, cognition, and transcription skills; and whether reading comprehension mediates the relations of discourse oral language and lexical literacy to the three dimensions of written composition; and (c) whether total effects of oral language, cognition, transcription, and reading comprehension vary for the three dimensions of written composition. Structural equation model results from 350 English-speaking second graders showed that higher order cognitive skills were differentially related to the three dimensions of written composition. Reading comprehension was related only to writing quality, but not to writing productivity or correctness in writing; and reading comprehension differentially mediated the relations of discourse oral language and lexical literacy to writing quality. Total effects of language, cognition, transcription, and reading comprehension varied largely for the three dimensions of written composition. These results support the dynamic relation hypothesis, role of reading in writing, and the importance of accounting for dimensions of written composition in a theoretical model of writing.

The impact of adjusting for pure predictors of exposure, mediator, and outcome on the variance of natural direct and indirect effect estimators.

It is now well established that adjusting for pure predictors of the outcome, in addition to confounders, allows unbiased estimation of the total exposure effect on an outcome with generally reduced standard errors (SEs). However, no analogous results have been derived for mediation analysis. Considering the simplest linear regression setting and the ordinary least square estimator, we obtained theoretical results showing that adjusting for pure predictors of the outcome, in addition to confounders, allows unbiased estimation of the natural indirect effect (NIE) and the natural direct effect (NDE) on the difference scale with reduced SEs. Adjusting for pure predictors of the mediator increases the SE of the NDE's estimator, but may increase or decrease the variance of the NIE's estimator. Adjusting for pure predictors of the exposure increases the variance of estimators of the NIE and NDE. Simulation studies were used to confirm and extend these results to the case where the mediator or the outcome is binary. Additional simulations were conducted to explore scenarios featuring an exposure-mediator interaction as well as the relative risk and odds ratio scales for the case of binary mediator and outcome. Both a regression approach and an inverse probability weighting approach were considered in the simulation study. A real-data illustration employing data from the Canadian Study of Health and Aging is provided. This analysis is concerned with the mediating effect of vitamin D in the effect of physical activity on dementia and its results are overall consistent with the theoretical and empirical findings.

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.

Protective factors against suicidal ideation among community-dwelling older adults with experience of spousal physical abuse: focusing on direct and indirect protections.

Purpose of the Study: The current research investigated the impact of protective factors - problem-focused coping, emotion-focused coping, having as a confidant either family or friends, and social support from friends - on suicidal ideation among older adults using Wisconsin Longitudinal Study (WLS) data. This study also investigated the indirect effects of protective factors on suicidal ideation (hereafter SI) as mediated by depressive symptoms among older adults who have experienced physical abuse from a spouse or romantic partner.Method: Data from The Wisconsin Longitudinal Study (WLS) were used. The sampled respondents were 475 men and women ages 64 to 65 who reported a history of being physically abused by a spouse or romantic partner. A path analysis using Mplus was employed to identify protective factors against SI.Results: Problem-focused coping and social support had negative indirect effects on SI as mediated by depressive symptoms. Emotion-focused coping had positive direct and indirect effects on SI.Conclusion: Tailored services to boost protective factors and regular screening tests are imperative to reduce depressive symptoms and SI among older adults who have experienced spousal or romantic partner physical abuse.

How do climate change experiments alter plot-scale climate?

To understand and forecast biological responses to climate change, scientists frequently use field experiments that alter temperature and precipitation. Climate manipulations can manifest in complex ways, however, challenging interpretations of biological responses. We reviewed publications to compile a database of daily plot-scale climate data from 15 active-warming experiments. We find that the common practices of analysing treatments as mean or categorical changes (e.g. warmed vs. unwarmed) masks important variation in treatment effects over space and time. Our synthesis showed that measured mean warming, in plots with the same target warming within a study, differed by up to 1.6  ∘ C (63% of target), on average, across six studies with blocked designs. Variation was high across sites and designs: for example, plots differed by 1.1  ∘ C (47% of target) on average, for infrared studies with feedback control (n = 3) vs. by 2.2  ∘ C (80% of target) on average for infrared with constant wattage designs (n = 2). Warming treatments produce non-temperature effects as well, such as soil drying. The combination of these direct and indirect effects is complex and can have important biological consequences. With a case study of plant phenology across five experiments in our database, we show how accounting for drier soils with warming tripled the estimated sensitivity of budburst to temperature. We provide recommendations for future analyses, experimental design, and data sharing to improve our mechanistic understanding from climate change experiments, and thus their utility to accurately forecast species' responses.

Counterpoint: Mediation Formulas With Binary Mediators and Outcomes and the "Rare Outcome Assumption".

In their accompanying article, Samoilenko and Lefebvre (Am J Epidemiol. 2019;188(7):1203-1205) correctly note 2 typographical errors in the formulas presented in a 2011 paper on placental abruption by Ananth and VanderWeele (Am J Epidemiol. 2011;174(1):99-108). Fortunately, to the best of our knowledge, researchers are using our methods papers and Dr. VanderWeele's 2015 book on mediation analysis (Explanation in Causal Inference; Oxford University Press, New York, New York), rather than the paper on placental abruption, to carry out their direct and indirect effect analyses; and in our methods papers and the book, the formulas are correct. The formulas discussed by Samoilenko and Lefebvre and in our work make reference to a "rare outcome assumption." In evaluating this assumption, it is important to note that the outcome is to be relatively rare across all strata defined by the exposure and the mediator-a point that is often neglected.

A Regression Framework for Causal Mediation Analysis with Applications to Behavioral Science.

We introduce and extend the classical regression framework for conducting mediation analysis from the fit of only one model. Using the essential mediation components (EMCs) allows us to estimate causal mediation effects and their analytical variance. This single-equation approach reduces computation time and permits the use of a rich suite of regression tools that are not easily implemented on a system of three equations. Additionally, we extend this framework to non-nested mediation systems, provide a joint measure of mediation for complex mediation hypotheses, propose new visualizations for mediation effects, and explain why estimates of the total effect may differ depending on the approach used. Using data from social science studies, we also provide extensive illustrations of the usefulness of this framework and its advantages over traditional approaches to mediation analysis. The example data are freely available for download online and we include the R code necessary to reproduce our results.

Artificial light at night decreases metamorphic duration and juvenile growth in a widespread amphibian.

Artificial light at night (ALAN) affects over 20% of the earth's surface and is estimated to increase 6% per year. Most studies of ALAN have focused on a single mechanism or life stage. We tested for indirect and direct ALAN effects that occurred by altering American toads' (Anaxyrus americanus) ecological interactions or by altering toad development and growth, respectively. We conducted an experiment over two life stages using outdoor mesocosms and indoor terraria. In the first phase, the presence of ALAN reduced metamorphic duration and periphyton biomass. The effects of ALAN appeared to be mediated through direct effects on toad development, and we found no evidence for indirect effects of ALAN acting through altered ecological interactions or colonization. In the second phase, post-metamorphic toad growth was reduced by 15% in the ALAN treatment. Juvenile-stage ALAN also affected toad activity: in natural light, toads retreated into leaf litter at night whereas ALAN toads did not change behaviour. Carry-over effects of ALAN were also present; juvenile toads that had been exposed to larval ALAN exhibited marginally increased activity. In this time frame and system, our experiments suggested ALAN's effects act primarily through direct effects, rather than indirect effects, and can persist across life stages.

Climate change-induced vegetation shifts lead to more ecological droughts despite projected rainfall increases in many global temperate drylands.

Drylands occur worldwide and are particularly vulnerable to climate change because dryland ecosystems depend directly on soil water availability that may become increasingly limited as temperatures rise. Climate change will both directly impact soil water availability and change plant biomass, with resulting indirect feedbacks on soil moisture. Thus, the net impact of direct and indirect climate change effects on soil moisture requires better understanding. We used the ecohydrological simulation model SOILWAT at sites from temperate dryland ecosystems around the globe to disentangle the contributions of direct climate change effects and of additional indirect, climate change-induced changes in vegetation on soil water availability. We simulated current and future climate conditions projected by 16 GCMs under RCP 4.5 and RCP 8.5 for the end of the century. We determined shifts in water availability due to climate change alone and due to combined changes of climate and the growth form and biomass of vegetation. Vegetation change will mostly exacerbate low soil water availability in regions already expected to suffer from negative direct impacts of climate change (with the two RCP scenarios giving us qualitatively similar effects). By contrast, in regions that will likely experience increased water availability due to climate change alone, vegetation changes will counteract these increases due to increased water losses by interception. In only a small minority of locations, climate change-induced vegetation changes may lead to a net increase in water availability. These results suggest that changes in vegetation in response to climate change may exacerbate drought conditions and may dampen the effects of increased precipitation, that is, leading to more ecological droughts despite higher precipitation in some regions. Our results underscore the value of considering indirect effects of climate change on vegetation when assessing future soil moisture conditions in water-limited ecosystems.

[Antiplatelet activity of flavonoids].

The review is devoted to the influence of the main subclasses of food flavonoids on the organism. The role of these polyphenolic compounds and containing them foods in reducing the risk of cardiovascular diseases is discussed. Particular attention is paid to the influence of flavonoids on the functional activity of platelets. The possible mechanisms of the effect of flavonoids on various links and stages of activation, adherence and aggregation of platelets are analyzed in detail. Most of the available data indicate the ability of these plant polyphenols to correct platelet abnormalities by affecting various receptors, mechanisms of intracellular mobilization of Ca2+, and multiple pathways of intra-platelet signaling. Problems that arise when transferring information obtained in in vitro experiments to the conditions of the whole organism are noted, which indicate the urgent need for further in-depth study of the effects of flavonoids. The data presented allow us to consider flavonoids as a basis for the creation of potentially effective and safe antiplatelet agents that can make a significant contribution to the treatment and prevention of a number of cardiovascular diseases.

Estimating causal contrasts involving intermediate variables in the presence of selection bias.

An important goal across the biomedical and social sciences is the quantification of the role of intermediate factors in explaining how an exposure exerts an effect on an outcome. Selection bias has the potential to severely undermine the validity of inferences on direct and indirect causal effects in observational as well as in randomized studies. The phenomenon of selection may arise through several mechanisms, and we here focus on instances of missing data. We study the sign and magnitude of selection bias in the estimates of direct and indirect effects when data on any of the factors involved in the analysis is either missing at random or not missing at random. Under some simplifying assumptions, the bias formulae can lead to nonparametric sensitivity analyses. These sensitivity analyses can be applied to causal effects on the risk difference and risk-ratio scales irrespectively of the estimation approach employed. To incorporate parametric assumptions, we also develop a sensitivity analysis for selection bias in mediation analysis in the spirit of the expectation-maximization algorithm. The approaches are applied to data from a health disparities study investigating the role of stage at diagnosis on racial disparities in colorectal cancer survival. Copyright © 2016 John Wiley & Sons, Ltd.

Biotic context and soil properties modulate native plant responses to enhanced rainfall.

BACKGROUND AND AIMS: The environmental and biotic context within which plants grow have a great potential to modify responses to climatic changes, yet few studies have addressed both the direct effects of climate and the modulating roles played by variation in the biotic (e.g. competitors) and abiotic (e.g. soils) environment. METHODS: In a grassland with highly heterogeneous soils and community composition, small seedlings of two native plants, Lasthenia californica and Calycadenia pauciflora, were transplanted into factorially watered and fertilized plots. Measurements were made to test how the effect of climatic variability (mimicked by the watering treatment) on the survival, growth and seed production of these species was modulated by above-ground competition and by edaphic variables. KEY RESULTS: Increased competition outweighed the direct positive impacts of enhanced rainfall on most fitness measures for both species, resulting in no net effect of enhanced rainfall. Both species benefitted from enhanced rainfall when the absence of competitors was accompanied by high soil water retention capacity. Fertilization did not amplify the watering effects; rather, plants benefitted from enhanced rainfall or competitor removal only in ambient nutrient conditions with high soil water retention capacity. CONCLUSIONS: The findings show that the direct effects of climatic variability on plant fitness may be reversed or neutralized by competition and, in addition, may be strongly modulated by soil variation. Specifically, coarse soil texture was identified as a factor that may limit plant responsiveness to altered water availability. These results highlight the importance of considering the abiotic as well as biotic context when making future climate change forecasts.

A principal stratification approach for evaluating natural direct and indirect effects in the presence of treatment-induced intermediate confounding.

Recently, several authors have shown that natural direct and indirect effects (NDEs and NIEs) can be identified under the sequential ignorability assumptions, as long as there is no mediator-outcome confounder that is affected by the treatment. However, if such a confounder exists, NDEs and NIEs will generally not be identified without making additional identifying assumptions. In this article, we propose novel identification assumptions and estimators for evaluating NDEs and NIEs under the usual sequential ignorability assumptions, using the principal stratification framework. It is assumed that the treatment and the mediator are dichotomous. We must impose strong assumptions for identification. However, even if these assumptions were violated, the bias of our estimator would be small under typical conditions, which can be easily evaluated from the observed data. This conjecture is confirmed for binary outcomes by deriving the bounds of the bias terms. In addition, the advantage of our estimator is illustrated through a simulation study. We also propose a method of sensitivity analysis that examines what happens when our assumptions are violated. We apply the proposed method to data from the National Center for Health Statistics.

Effect of thiram and of a hydrocarbon mixture on freshwater macroinvertebrate communities in outdoor stream and pond mesocosms: II. Biological and ecological trait responses and leaf litter breakdown.

Higher-tier ecological risk assessment of chemicals often relies upon studies in dynamic and/or static mesocosms. Physico-chemical and hydrological properties of each type of mesocosm result in specific chemicals fate, community functioning, and potential recovery. In the present study, macroinvertebrate abundance- and biomass-weighted biological and ecological trait matrices were used to assess the effects of a dithiocarbamate fungicide, thiram (35 and 170 µg l(-1)), and of a petroleum middle distillate (0.01, 0.4, 2 and 20 mg l(-1)) in outdoor stream and pond mesocosms. Trait sensitivity was characterized using functional diversity indices and trait modality distributions to assess the influence of the type of experimental systems and the ability of traits to disentangle chemical-induced effects from temporal and stochastic variations. In addition, leaf litter breakdown was used as an integrative functional endpoint. Regardless to the substance, treatments had a direct effect on the functional structure of benthic macroinvertebrate communities in streams but not in ponds, suggesting that global functional responses to chemicals are system-specific. Although both substances had an effect in streams, differences were noticed in the nature of the affected traits suggesting that chemical mode of action plays a role in functional alterations. This was illustrated by the link between negative effects of chemical exposure on detritivorous taxa and reduced litter breakdown rate in streams. Therefore, characterisation of macroinvertebrate biological traits associated with the measurement of a functional process such as litter breakdown may provide a comprehensive understanding of the effects occurring in mesocosms exposed to organic chemicals.

Indirect effects drive evolutionary responses to global change.

Anthropogenic environmental changes pose significant threats to plant and animal populations. These changes also may affect the evolution of natural populations either directly or indirectly by altering the outcome of species interactions that are important drivers of evolution. This latter indirect pathway may be especially important for evolutionary responses to elevated atmospheric CO2 concentrations (eCO2), which appear to have minimal direct effects on plant evolution but have large effects on interspecific interactions, such as competition. We manipulated competitive and CO2 environments of experimental Arabidopsis thaliana populations to test whether eCO2 alters evolutionary trajectories indirectly by altering selection imposed by competitors. We found that interspecific competition increased selection on growth traits, reduced heritabilities, and altered genetic covariances between traits and that the magnitude of these effects depended upon the CO2 environment. Although eCO2 had minimal direct effects on evolutionary processes, eCO2 typically reduced the strength of selection imposed by competitors and, therefore, relaxed selection on plant traits when competitors were present. Our results indicate that global changes may affect plant evolution indirectly by altering competitive interactions and underscore the importance of conducting research in natural communities when attempting to predict population responses to global change.