Practical causal mediation analysis: extending nonparametric estimators to accommodate multiple mediators and multiple intermediate confounders.

Mediation analysis is appealing for its ability to improve understanding of the mechanistic drivers of causal effects, but real-world data complexities challenge its successful implementation, including (i) the existence of post-exposure variables that also affect mediators and outcomes (thus, confounding the mediator-outcome relationship), that may also be (ii) multivariate, and (iii) the existence of multivariate mediators. All three challenges are present in the mediation analysis we consider here, where our goal is to estimate the indirect effects of receiving a Section 8 housing voucher as a young child on the risk of developing a psychiatric mood disorder in adolescence that operate through mediators related to neighborhood poverty, the school environment, and instability of the neighborhood and school environments, considered together and separately. Interventional direct and indirect effects (IDE/IIE) accommodate post-exposure variables that confound the mediator-outcome relationship, but currently, no readily implementable nonparametric estimator for IDE/IIE exists that allows for both multivariate mediators and multivariate post-exposure intermediate confounders. The absence of such an IDE/IIE estimator that can easily accommodate both multivariate mediators and post-exposure confounders represents a significant limitation for real-world analyses, because when considering each mediator subgroup separately, the remaining mediator subgroups (or a subset of them) become post-exposure intermediate confounders. We address this gap by extending a recently developed nonparametric estimator for the IDE/IIE to allow for easy incorporation of multivariate mediators and multivariate post-exposure confounders simultaneously. We apply the proposed estimation approach to our analysis, including walking through a strategy to account for other, possibly co-occurring intermediate variables when considering each mediator subgroup separately.

DP2LM: leveraging deep learning approach for estimation and hypothesis testing on mediation effects with high-dimensional mediators and complex confounders.

Traditional linear mediation analysis has inherent limitations when it comes to handling high-dimensional mediators. Particularly, accurately estimating and rigorously inferring mediation effects is challenging, primarily due to the intertwined nature of the mediator selection issue. Despite recent developments, the existing methods are inadequate for addressing the complex relationships introduced by confounders. To tackle these challenges, we propose a novel approach called DP2LM (Deep neural network-based Penalized Partially Linear Mediation). This approach incorporates deep neural network techniques to account for nonlinear effects in confounders and utilizes the penalized partially linear model to accommodate high dimensionality. Unlike most existing works that concentrate on mediator selection, our method prioritizes estimation and inference on mediation effects. Specifically, we develop test procedures for testing the direct and indirect mediation effects. Theoretical analysis shows that the tests maintain the Type-I error rate. In simulation studies, DP2LM demonstrates its superior performance as a modeling tool for complex data, outperforming existing approaches in a wide range of settings and providing reliable estimation and inference in scenarios involving a considerable number of mediators. Further, we apply DP2LM to investigate the mediation effect of DNA methylation on cortisol stress reactivity in individuals who experienced childhood trauma, uncovering new insights through a comprehensive analysis.

A pantropical assessment of deforestation caused by industrial mining.

Growing demand for minerals continues to drive deforestation worldwide. Tropical forests are particularly vulnerable to the environmental impacts of mining and mineral processing. Many local- to regional-scale studies document extensive, long-lasting impacts of mining on biodiversity and ecosystem services. However, the full scope of deforestation induced by industrial mining across the tropics is yet unknown. Here, we present a biome-wide assessment to show where industrial mine expansion has caused the most deforestation from 2000 to 2019. We find that 3,264 km2 of forest was directly lost due to industrial mining, with 80% occurring in only four countries: Indonesia, Brazil, Ghana, and Suriname. Additionally, controlling for other nonmining determinants of deforestation, we find that mining caused indirect forest loss in two-thirds of the investigated countries. Our results illustrate significant yet unevenly distributed and often unmanaged impacts on these biodiverse ecosystems. Impact assessments and mitigation plans of industrial mining activities must address direct and indirect impacts to support conservation of the world's tropical forests.

Estimated population-level impact of pneumococcal conjugate vaccines against all-cause pneumonia mortality among unvaccinated age groups in five Latin American countries.

BACKGROUND: Pneumococcal conjugate vaccines (PCVs) provide strong direct protection in children, while limited data are available on their indirect effect on mortality among older age groups. This multi-country study aimed to assess the population-level impact of pediatric PCVs on all-cause pneumonia mortality among ≥5 years of age, and invasive pneumococcal disease (IPD) cases in Chile. METHODS: Demographic and mortality data from Argentina, Brazil, Chile, Colombia, and Mexico were collected considering the ≥ 5-year-old population, from 2000-2019, with 1,795,789 deaths due to all-cause pneumonia. IPD cases in Chile were also evaluated. Time series models were employed to evaluate changes in all-cause pneumonia deaths during the post-vaccination period, with other causes of death used as synthetic controls for unrelated temporal trends. RESULTS: No significant change in death rates due to all-cause pneumonia was detected following PCV introduction among most age groups and countries. The proportion of IPD cases caused by vaccine serotypes decreased from 29% (2012) to 6% (2022) among ≥65 years in Chile. DISCUSSION: While an effect of PCV against pneumonia deaths (a broad clinical definition that may not be specific enough to measure indirect effects) was not detected, evidence of indirect PCV impact was observed among vaccine-type-specific IPD cases.

The Hierarchical Coevolutionary Units of Ecological Networks.

In ecological networks, cohesive groups of species may shape the evolution of interactions, serving as coevolutionary units. Ranging across network scales, from motifs to isolated components, elucidating which cohesive groups are more determinant for coevolution remains a challenge in ecology. We address this challenge by integrating 376 empirical mutualistic and antagonistic networks and coevolutionary models. We identified cohesive groups at four network scales containing a significant proportion of potential direct coevolutionary effects. Cohesive groups displayed hierarchical organisation, and potential coevolutionary effects overflowing lower-scale groups were contained by higher-scale groups, underscoring the hierarchy's impact. However, indirect coevolutionary effects blurred group boundaries and hierarchy, particularly under strong selection from ecological interactions. Thus, under strong selection, indirect effects render networks themselves, and not cohesive groups, as the likely coevolutionary units of ecological systems. We hypothesise hierarchical cohesive groups to also shape how other forms of direct and indirect effects propagate in ecological systems.

Food web context modifies predator foraging and weakens trophic interaction strength.

Trophic interaction modifications (TIM) are widespread in natural systems and occur when a third species indirectly alters the strength of a trophic interaction. Past studies have focused on documenting the existence and magnitude of TIMs; however, the underlying processes and long-term consequences remain elusive. To address this gap, we experimentally quantified the density-dependent effect of a third species on a predator's functional response. We conducted short-term experiments with ciliate communities composed of a predator, prey and non-consumable 'modifier' species. In both communities, increasing modifier density weakened the trophic interaction strength, due to a negative effect on the predator's space clearance rate. Simulated long-term dynamics indicate quantitative differences between models that account for TIMs or include only pairwise interactions. Our study demonstrates that TIMs are important to understand and predict community dynamics and highlights the need to move beyond focal species pairs to understand the consequences of species interactions in communities.

Estimating community-wide indirect effects of influenza vaccination: triangulation using mathematical models and bias analysis.

Understanding whether influenza vaccine promotion strategies produce community-wide indirect effects is important for establishing vaccine coverage targets and optimizing vaccine delivery. Empirical epidemiologic studies and mathematical models have been used to estimate indirect effects of vaccines but rarely for the same estimand in the same dataset. Using these approaches together could be a powerful tool for triangulation in infectious disease epidemiology because each approach is subject to distinct sources of bias. We triangulated evidence about indirect effects from a school-located influenza vaccination program using two approaches: a difference-in-difference (DID) analysis, and an age-structured, deterministic, compartmental model. The estimated indirect effect was substantially lower in the mathematical model than in the DID analysis (2.1% (95% Bayesian credible intervals 0.4 - 4.4%) vs. 22.3% (95% CI 7.6% - 37.1%)). To explore reasons for differing estimates, we used sensitivity analyses and probabilistic bias analyses. When we constrained model parameters such that projections matched the DID analysis, results only aligned with the DID analysis with substantially lower pre-existing immunity among school-age children and older adults. Conversely, DID estimates corrected for potential bias only aligned with mathematical model estimates under differential outcome misclassification. We discuss how triangulation using empirical and mathematical modelling approaches could strengthen future studies.

How modularity and heterotrophy complicate the understanding of the causes of thermal performance curves: the case of feeding rate in a filter-feeding animal.

Warming global temperatures have consequences for biological rates. Feeding rates reflect the intake of energy that fuels survival, growth and reproduction. However, temperature can also affect food abundance and quality, as well as feeding behavior, which all affect feeding rate, making it challenging to understand the pathways by which temperature affects the intake of energy. Therefore, we experimentally assessed how clearance rate varied across a thermal gradient in a filter-feeding colonial marine invertebrate (the bryozoan Bugula neritina). We also assessed how temperature affects phytoplankton as a food source, and zooid states within a colony that affect energy budgets and feeding behavior. Clearance rate increased linearly from 18°C to 32°C, a temperature range that the population experiences most of the year. However, temperature increased algal cell size, and decreased the proportion of feeding zooids, suggesting indirect effects of temperature on clearance rates. Temperature increased polypide regression, possibly as a stress response because satiation occurred quicker, or because phytoplankton quality declined. Temperature had a greater effect on clearance rate per feeding zooid than it did per total zooids. Together, these results suggest that the effect of temperature on clearance rate at the colony level is not just the outcome of individual zooids feeding more in direct response to temperature but also emerges from temperature increasing polypide regression and the remaining zooids increasing their feeding rates in response. Our study highlights some of the challenges for understanding why temperature affects feeding rates, especially for understudied, yet ecologically important, marine colonial organisms.

Dynamic path analysis for exploring treatment effect mediation processes in clinical trials with time-to-event endpoints.

Why does a beneficial treatment effect on a longitudinal biomarker not translate into overall treatment benefit on survival, when the biomarker is in fact a prognostic factor of survival? In a recent exploratory data analysis in oncology, we were faced with this seemingly paradoxical result. To address this problem, we applied a theoretically principled methodology called dynamic path analysis, which allows us to perform mediation analysis with a longitudinal mediator and survival outcome. The aim of the analysis is to decompose the total treatment effect into a direct treatment effect and an indirect treatment effect mediated through a carefully constructed mediation path. The dynamic nature of the underlying methodology enables us to describe how these effects evolve over time, which can add to the mechanistic understanding of the underlying processes. In this paper, we present a detailed description of the dynamic path analysis framework and illustrate its application to survival mediation analysis using simulated and real data. The use case analysis provides clarity on the specific exploratory question of interest while the methodology generalizes to a wide range of applications in drug development where time-to-event is the primary clinical outcome of interest.

Multiply robust estimation of natural indirect effects with multiple ordered mediators.

Multiple mediation analysis is a powerful methodology to assess causal effects in the presence of multiple mediators. Several methodologies, such as G-computation and inverse-probability-weighting, have been widely used to draw inferences about natural indirect effects (NIEs). However, a limitation of these methods is their potential for model misspecification. Although powerful semiparametric methods with high robustness and consistency have been developed for inferring average causal effects and for analyzing the effects of a single mediator, a comparably robust method for multiple mediation analysis is still lacking. Therefore, this theoretical study proposes a method of using multiply robust estimators of NIEs in the presence of multiple ordered mediators. We show that the proposed estimators not only enjoy the multiply robustness to model misspecification, they are also consistent and asymptotically normal under regular conditions. We also performed simulations for empirical comparisons of the finite-sample properties between our multiply robust estimators and existing methods. In an illustrative example, a dataset for liver disease patients in Taiwan is used to examine the mediating roles of liver damage and liver cancer in the pathway from hepatitis B/C virus infection to mortality. The model is implemented in the open-source R package "MedMR."

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).

Out of shape: Ocean acidification simplifies coral reef architecture and reshuffles fish assemblages.

Climate change stressors are progressively simplifying biogenic habitats in the terrestrial and marine realms, and consequently altering the structure of associated species communities. Here, we used a volcanic CO2 seep in Papua New Guinea to test in situ if altered reef architecture due to ocean acidification reshuffles associated fish assemblages. We observed replacement of branching corals by massive corals at the seep, with simplified coral architectural complexity driving abundance declines between 60% and 86% for an assemblage of damselfishes associated with branching corals. An experimental test of habitat preference for a focal species indicated that acidification does not directly affect habitat selection behaviour, with changes in habitat structural complexity consequently appearing to be the stronger driver of assemblage reshuffling. Habitat health affected anti-predator behaviour, with P. moluccensis becoming less bold on dead branching corals relative to live branching corals, irrespective of ocean acidification. We conclude that coral reef fish assemblages are likely to be more sensitive to changes in habitat structure induced by increasing pCO2 than any direct effects on behaviour, indicating that changes in coral architecture and live cover may act as important mediators of reef fish community structures in a future ocean.

Infection risk associated with carnivore carcasses may govern trophic interactions between maggots and insectivorous passerine birds.

Infection risk by pathogenic agents motivates hosts to avoid using resources with high risks. This, in turn, results in increased availability of these resources for other species that are more tolerant of infections. For instance, carcasses of mammalian carnivores are frequently avoided by conspecific or closely related carnivores, allowing them to be almost exclusively used by maggots. This may lead to novel interactions with other species. This study investigated the consumption of maggots from carnivore carcasses by non-corvid passerines. We successfully monitored 66 raccoon carcasses in Hokkaido, Japan, from 2016 to 2019. Vertebrates only scavenged 14 carcasses before maggot dispersal; the other 52 carcasses produced abundant maggots that regularly fed at least 12 species of non-corvid passerines. Surprisingly, predation occurred at a distance from the carcasses, mainly after maggot dispersal for pupation, despite the higher efficiency of feeding on maggot masses on the carcasses. Birds are likely to reduce the potential risk of infection from the carcass and/or from maggots on the carcasses. Overall, only 1% of maggots were consumed. Our results suggest that necrophagous flies could benefit from the infection risk associated with carnivore carcasses, which may decrease scavenging by other carnivores and constrain maggot consumption by insectivorous birds.

Quantifying reciprocal relationships between poverty and health: combining a causal loop diagram with longitudinal structural equation modelling.

BACKGROUND: This study takes on the challenge of quantifying a complex causal loop diagram describing how poverty and health affect each other, and does so using longitudinal data from The Netherlands. Furthermore, this paper elaborates on its methodological approach in order to facilitate replication and methodological advancement. METHODS: After adapting a causal loop diagram that was built by stakeholders, a longitudinal structural equation modelling approach was used. A cross-lagged panel model with nine endogenous variables, of which two latent variables, and three time-invariant exogenous variables was constructed. With this model, directional effects are estimated in a Granger-causal manner, using data from 2015 to 2019. Both the direct effects (with a one-year lag) and total effects over multiple (up to eight) years were calculated. Five sensitivity analyses were conducted. Two of these focus on lower-income and lower-wealth individuals. The other three each added one exogenous variable: work status, level of education, and home ownership. RESULTS: The effects of income and financial wealth on health are present, but are relatively weak for the overall population. Sensitivity analyses show that these effects are stronger for those with lower incomes or wealth. Physical capability does seem to have strong positive effects on both income and financial wealth. There are a number of other results as well, as the estimated models are extensive. Many of the estimated effects only become substantial after several years. CONCLUSIONS: Income and financial wealth appear to have limited effects on the health of the overall population of The Netherlands. However, there are indications that these effects may be stronger for individuals who are closer to the poverty threshold. Since the estimated effects of physical capability on income and financial wealth are more substantial, a broad recommendation would be that including physical capability in efforts that are aimed at improving income and financial wealth could be useful and effective. The methodological approach described in this paper could also be applied to other research settings or topics.

Reciprocal facilitation between ants and small mammals in tidal marshes.

The role of facilitation in shaping natural communities has primarily been studied in the context of plant assemblages, while its relevance for mobile animals remains less understood. Our study investigates whether reciprocal interspecific facilitation may exist between fire ants (Solenopsis richteri) and cavies (Cavia aperea), two mobile animals, in the SW Atlantic coast brackish marshes. Field samples showed a spatial association between ant mounds and cavies, and that ants prefer to use cavy runways for movement within the marsh. Through experiments involving transplanting the dominant plant, cordgrass (Spartina densiflora), and manipulating cavy presence in areas with and without ant mounds, we observed that cavies forage extensively (and defecate more) near ant mounds. The ants actively remove cavy droppings in their mound vicinity. These ant activities and interactions with cavy droppings led to reduced moisture and organic content while increasing nitrate and phosphate levels in marsh sediment. Consequently, this enhanced plant growth, indirectly facilitating the cavies, which preferred consuming vegetation near ant mounds. These cascading indirect effects persisted over time; even four months after cavies left the marshes, transplanted plants near ant mounds remained larger and exhibited more leaf senescence when exposed to cavy herbivory. Therefore, the networks of positive interactions appear to generate simultaneous selection among species (populations), promoting coexistence within the community. Although complex, these reciprocal facilitative effects among mobile animals may be more common than currently believed and should be further studied to gain a better understanding of the underlying mechanisms driving species coexistence in natural communities.

Childhood maltreatment and resting-state network connectivity: The risk-buffering role of positive parenting.

Unraveling the neurobiological foundations of childhood maltreatment is important due to the persistent associations with adverse mental health outcomes. However, the mechanisms through which abuse and neglect disturb resting-state network connectivity remain elusive. Moreover, it remains unclear if positive parenting can mitigate the negative impact of childhood maltreatment on network connectivity. We analyzed a cohort of 194 adolescents and young adults (aged 14-25, 47.42% female) from the Neuroscience in Psychiatry Network (NSPN) to investigate the impact of childhood abuse and neglect on resting-state network connectivity. Specifically, we examined the SAN, DMN, FPN, DAN, and VAN over time. We also explored the moderating role of positive parenting. The results showed that childhood abuse was linked to stronger connectivity within the SAN and VAN, as well as between the DMN-DAN, DMN-VAN, DMN-SAN, SAN-DAN, FPN-DAN, SAN-VAN, and VAN-DAN networks about 18 months later. Positive parenting during childhood buffered the negative impact of childhood abuse on network connectivity. To our knowledge, this is the first study to demonstrate the protective effect of positive parenting on network connectivity following childhood abuse. These findings not only highlight the importance of positive parenting but also lead to a better understanding of the neurobiology and resilience mechanisms of childhood maltreatment.

Psychosocial and financial well-being mediated the effects of COVID-19 distress on suicidality: a serial mediation model among Hong Kong young adults.

PURPOSE: The COVID-19 pandemic has brought significant distress on not only the physical health but also mental health of individuals. The present study investigated the direct and indirect effects from COVID-19 distress to suicidality via psychosocial and financial well-being among young people. METHODS: This cross-sectional survey recruited 1472 Hong Kong young people via random sampling in 2021. The respondents completed a phone survey on COVID-19 distress, the four-item Patient Health Questionnaire and items on social well-being, financial well-being, and suicidality. Structural equation modeling (SEM) was conducted to examine the direct and indirect effects of COVID-19 distress on suicidality via psychosocial and financial well-being. RESULTS: The direct effect of COVID-19 distress on suicidality was not significant (ß = 0.022, 95% CI - 0.097-0.156). The total indirect effect from COVID-19 distress to suicidality was significant and positive (αßγ = 0.150, 95% CI = 0.085-0.245) and accounted for 87% of the total effect (B = 0.172, 95% CI = 0.043-0.341). There were significant specific indirect effects via social well-being and psychological distress, and financial well-being and psychological distress. CONCLUSION: The present findings support different pathways from COVID-19 distress to suicidality via functioning in different domains among young people in Hong Kong. Measures are needed to ameliorate the impact on their social and financial well-being to reduce their psychological distress and suicidality.

Predation risk-mediated indirect effects promote submerged plant growth: Implications for lake restoration.

Biological manipulation, involving fish stockings, is commonly used to counteract the deterioration of submerged vegetation in eutrophic lakes. Nevertheless, the non-consumptive effects (NCEs) of stocked carnivorous fish are often overlooked. Using a controlled experimental system, we investigated the NCEs of a native carnivorous fish, snakehead (Channa argus), on two key biological factors, herbivore-dominated grass carp (Ctenopharyngodon idella) and disturbance-dominated loach (Misgurnus anguillicaudatus), influencing submerged plants growth. Additionally, we conducted a meta-analysis on predation risk and primary productivity. The results reveal that predation risk induces oxidative stress damage and affects grass carp growth. Non-significant changes in cortisol and glucose may be linked to predation risk prediction. Simultaneously, predation risk reduces fish feeding and disturbance behavior, relieving pressure on submerged plants to be grazed and disturbed, thereby supporting plant development. The presence of submerged plants, in turn, enhances loach activity and influences water body characteristics through negative feedback. Furthermore, the meta-analysis results indicate the facilitative effect of predation risk on primary producers. Our findings contribute to the understanding of biological manipulation theory. We demonstrate that the predation risk associated with introducing carnivorous fish can promote the growth of submerged plants through behaviorally mediated indirect effects. This highlights the potential utility of predation risk in lake restoration efforts.

High importance of indirect evolutionary rescue in a small food web.

Evolutionary rescue may allow species to survive environmental change, but how this mechanism operates in food webs is poorly understood. Here, the evolutionary rescue was investigated in a small model food web, systematically allowing the evolution of each single species in order to reveal how its adaptation affects the persistence of itself and others. The impact of evolution was highly species-specific and not necessarily positive: only one species, the specialist predator, consistently had a positive impact on overall persistence. Most strikingly, evolution overwhelmingly affected other species: rescue of others (indirect rescue) was far more frequent than self-rescue, and negative effects were nearly always indirect. This demonstrates that evolutionary rescue in food webs is inextricably bound up with species interactions, as the effects of evolution in one species ripple through the entire community. It is therefore critically important to consider the food web context in efforts to understand how species may survive global change.

Predator-mediated interactions through changes in predator home range size can lead to local prey exclusion.

The strength of indirect biotic interactions is difficult to quantify in the wild and can alter community composition. To investigate whether the presence of a prey species affects the population growth rate of another prey species, we quantified predator-mediated interaction strength using a multi-prey mechanistic model of predation and a population matrix model. Models were parametrized using behavioural, demographic and experimental data from a vertebrate community that includes the arctic fox (Vulpes lagopus), a predator feeding on lemmings and eggs of various species such as sandpipers and geese. We show that the positive effects of the goose colony on sandpiper nesting success (due to reduction of search time for sandpiper nests) were outweighed by the negative effect of an increase in fox density. The fox numerical response was driven by changes in home range size. As a result, the net interaction from the presence of geese was negative and could lead to local exclusion of sandpipers. Our study provides a rare empirically based model that integrates mechanistic multi-species functional responses and behavioural processes underlying the predator numerical response. This is an important step forward in our ability to quantify the consequences of predation for community structure and dynamics.