Concerning the Consistency Assumption in Causal Inference

Harmful compared to what? The problem of gaming and ambiguous causal questions

BACKGROUND AND AIMS

There has been much concern regarding potential harmful effects of video game-play in the past 40 years, but limited progress in understanding its causal role. This paper discusses the basic requirements for identifying causal effects of video game-play and argues that most research to date has focused upon ambiguous causal questions.

METHODS

Video games and mental health are discussed from the perspective of causal inference with compound exposures; that is, exposures with multiple relevant variants that affect outcomes in different ways.

RESULTS

Not only does exposure to video games encompass multiple different factors, but also not playing video games is equally ambiguous. Estimating causal effects of a compound exposure introduces the additional challenge of exposure-version confounding.

CONCLUSIONS

Without a comparison of well-defined interventions, research investigating the effects of video game-play will be difficult to translate into actionable health interventions. Interventions that target games should be compared with other interventions aimed at improving the same outcomes.

Robust inference for causal mediation analysis of recurrent event data

Recurrent events, including cardiovascular events, are commonly observed in biomedical studies. Understanding the effects of various treatments on recurrent events and investigating the underlying mediation mechanisms by which treatments may reduce the frequency of recurrent events are crucial tasks for researchers. Although causal inference methods for recurrent event data have been proposed, they cannot be used to assess mediation. This study proposed a novel methodology of causal mediation analysis that accommodates recurrent outcomes of interest in a given individual. A formal definition of causal estimands (direct and indirect effects) within a counterfactual framework is given, and empirical expressions for these effects are identified. To estimate these effects, a semiparametric estimator with triple robustness against model misspecification was developed. The proposed methodology was demonstrated in a real-world application. The method was applied to measure the effects of two diabetes drugs on the recurrence of cardiovascular disease and to examine the mediating role of kidney function in this process.

Assessing Direct and Spillover Effects of Intervention Packages in Network-randomized Studies

BACKGROUND

Intervention packages may result in a greater public health impact than single interventions. Understanding the separate impact of each component on the overall package effectiveness can improve intervention delivery.

METHODS

We adapted an approach to evaluate the effects of a time-varying intervention package in a network-randomized study. In some network-randomized studies, only a subset of participants in exposed networks receive the intervention themselves. The spillover effect contrasts average potential outcomes if a person was not exposed to themselves under intervention in the network versus no intervention in a control network. We estimated the effects of components of the intervention package in HIV Prevention Trials Network 037, a Phase III network-randomized HIV prevention trial among people who inject drugs and their risk networks using marginal structural models to adjust for time-varying confounding. The index participant in an intervention network received a peer education intervention initially at baseline, then boosters at 6 and 12 months. All participants were followed to ascertain HIV risk behaviors.

RESULTS

There were 560 participants with at least one follow-up visit, 48% of whom were randomized to the intervention, and 1,598 participant visits were observed. The spillover effect of the boosters in the presence of initial peer education training was a 39% rate reduction (rate ratio = 0.61; 95% confidence interval = 0.43, 0.87).

CONCLUSIONS

These methods will be useful for evaluating intervention packages in studies with network features.

Errors in the Calculation of the Population Attributable Fraction

One of the common errors in the calculation of the population attributable fraction (PAF) is the use of an adjusted risk ratio in the Levin formula. In this article, we discuss the errors visually using wireframes by varying the standardized mortality ratio (SMR) and associational risk ratio (aRR) when the prevalence of exposure is fixed. When SMR >1 and SMR > aRR, the absolute bias is positive, and its magnitude increases as the difference between SMR and aRR increases. By contrast, when aRR > SMR > 1, the absolute bias is negative and its magnitude is relatively small. Moreover, when SMR > aRR, the relative bias is larger than one, whereas when SMR < aRR, the relative bias is smaller than one. Although the target population of the PAF is the total population, the target of causation of the PAF is not the total population but the exposed group.

On variance estimation of the inverse probability-of-treatment weighting estimator: A tutorial for different types of propensity score weights

Propensity score methods, such as inverse probability-of-treatment weighting (IPTW), have been increasingly used for covariate balancing in both observational studies and randomized trials, allowing the control of both systematic and chance imbalances. Approaches using IPTW are based on two steps: (i) estimation of the individual propensity scores (PS), and (ii) estimation of the treatment effect by applying PS weights. Thus, a variance estimator that accounts for both steps is crucial for correct inference. Using a variance estimator which ignores the first step leads to overestimated variance when the estimand is the average treatment effect (ATE), and to under or overestimated estimates when targeting the average treatment effect on the treated (ATT). In this article, we emphasize the importance of using an IPTW variance estimator that correctly considers the uncertainty in PS estimation. We present a comprehensive tutorial to obtain unbiased variance estimates, by proposing and applying a unifying formula for different types of PS weights (ATE, ATT, matching and overlap weights). This can be derived either via the linearization approach or M-estimation. Extensive R code is provided along with the corresponding large-sample theory. We perform simulation studies to illustrate the behavior of the estimators under different treatment and outcome prevalences and demonstrate appropriate behavior of the analytical variance estimator. We also use a reproducible analysis of observational lung cancer data as an illustrative example, estimating the effect of receiving a PET-CT scan on the receipt of surgery.

Causal Inference Challenges in the Relationship Between Social Determinants and Cardiovascular Outcomes

The effects of social determinants on cardiovascular outcomes are frequently estimated in epidemiologic analyses, but the profound causal and statistical challenges of this research program are not widely discussed. Here, we carefully review definitions and measures for social determinants of cardiovascular health and then examine the various assumptions required for valid causal inference in multivariable analyses of observational data, such as what one would typically encounter in cohorts, population surveys, health care databases, and vital statistics databases. We explain the necessity of the "well-defined exposure" and show how this goal relates to the "consistency assumption" that is necessary for valid causal inference. Well-defined exposure is especially challenging for social determinants of health because they are seldom simple atomistic interventions that are easily conceptualized and measured. We then review threats to valid inference that arise from confounding, selection bias, information bias, and positivity violations. Other causal considerations are reviewed and explained, such as correct model specification, absence of immortal time, and avoidance of the "Table 2 Fallacy," and their application to social determinants of cardiovascular outcomes are discussed. Fruitful approaches, including focusing on policy interventions and the "target trial" frameworks are proposed and provide a pathway for a more efficacious research program that can more reliably improve population health. Valid causal inference in this setting is quite challenging, but-with clever design and thoughtful analysis-the important role of social factors in patterning cardiovascular outcomes can be quantified and reported.

Enhancing inferences and conclusions in body image focused non-experimental research via a causal modelling approach: A tutorial

Causal inference is often the goal of psychological research. However, most researchers refrain from drawing causal conclusions based on non-experimental evidence. Despite the challenges associated with producing causal evidence from non-experimental data, it is crucial to address causal questions directly rather than avoiding them. Here we provide a clear, non-technical overview of the fundamental concepts (including the counterfactual framework and related assumptions) and tools that permit causal inference in non-experimental data, intended as a starting point for readers unfamiliar with the literature. Certain tools, such as the target trial framework and causal diagrams, have been developed to assist with the identification and reduction of potential biases in study design and analysis and the interpretation of findings. We apply these concepts and tools to a motivating example from the body image field. We assert that more precise and detailed elucidation of the barriers to causal inference within one's study is arguably a key first step in the enhancement of non-experimental research and future intervention development and evaluation.

Exploring a Potential Interaction Between the Effect of Specific Maternal Smoking Patterns and Comorbid Antenatal Depression in Causing Postpartum Depression

Purpose

To explore a potential interaction between the effect of specific maternal smoking patterns and the presence of antenatal depression, as independent exposures, in causing postpartum depression (PPD).

Methods

This case-control study of participants with singleton term births (N = 51220) was based on data from the 2017-2018 Pregnancy Risk Assessment Monitoring System. Multivariable log-binomial regression models examined the main effects of smoking patterns and self-reported symptoms of antenatal depression on the risk of PPD on the adjusted risk ratio (aRR) scale and tested a two-way interaction adjusting for covariates selected in a directed acyclic graph (DAG). The interaction effects were measured on the additive scale using relative excess risk due to interaction (RERI), the attributable proportion of interaction (AP), and the synergy index (SI). Causal effects were defined in a counterfactual framework. The E-value quantified the potential impact of unobserved/unknown covariates, conditional on observed covariates.

Results

Among 6841 women in the sample who self-reported PPD, 35.7% also reported symptoms of antenatal depression. Out of 3921 (7.7%) women who reported smoking during pregnancy, 32.6% smoked at high intensity (≥10 cigarettes/day) in all three trimesters and 36.6% had symptoms of antenatal depression. The main effect of PPD was the strongest for women who smoked at high intensity throughout pregnancy (aRR 1.65; 95% CI: 1.63, 1.68). A synergistic interaction was detected, and the effect of all maternal smoking patterns was augmented, particularly in late pregnancy for Increasers and Reducers.

Conclusion

Strong associations and interaction effects between maternal smoking patterns and co-occurring antenatal depression support smoking prevention and cessation interventions during pregnancy to lower the likelihood of PPD.

Annual Research Review: Sex, gender, and internalizing conditions among adolescents in the 21st century - trends, causes, consequences

Internalizing conditions of psychopathology include depressive and anxiety disorders; they most often onset in adolescence, are relatively common, and contribute to significant population morbidity and mortality. In this research review, we present the evidence that internalizing conditions, including depression and anxiety, as well as psychological distress, suicidal thoughts and self-harm, and fatal suicide, are considerably increasing in adolescent populations across many countries. Evidence indicates that increases are currently greatest in female adolescents. We present an epidemiological framework for evaluating the causes of these increases, and synthesize research on whether several established risk factors (e.g., age of pubertal transition and stressful life events) and novel risk factors (e.g., digital technology and social media) meet conditions necessary to be plausible causes of increases in adolescent internalizing conditions. We conclude that there are a multitude of potential causes of increases in adolescent internalizing conditions, outline evidence gaps including the lack of research on nonbinary and gender nonconforming populations, and recommend necessary prevention and intervention foci from a clinical and public health perspective.

Childhood maltreatment and risk of endocrine diseases: an exploration of mediating pathways using sequential mediation analysis

BACKGROUND

Adverse childhood experiences (ACEs), including childhood maltreatment, have been linked with increased risk of diabetes and obesity during adulthood. A comprehensive assessment on the associations between childhood maltreatment and all major endocrine diseases, as well as the relative importance of different proposed mechanistic pathways on these associations, is currently lacking.

METHODS

Based on the UK Biobank, we constructed a cohort including 151,659 participants with self-reported data on childhood maltreatment who were 30 years of age or older on/after January 1, 1985. All participants were followed from the index date (i.e., January 1, 1985, or their 30th birthday, whichever came later) until the first diagnosis of any or specific (12 individual diagnoses and 9 subtypes) endocrine diseases, death, or the end of follow-up (December 31, 2019), whichever occurred first. We used Cox models to examine the association of childhood maltreatment, treated as continuous (i.e., the cumulative number of experienced childhood maltreatment), ordinal (i.e., 0, 1 and ≥ 2), or binary (< 2 and ≥ 2) variable, with any and specific endocrine diseases, adjusted for multiple covariates. We further examined the risk of having multiple endocrine diseases using Linear or Logistic Regression models. Then, sequential mediation analyses were performed to assess the contribution of four possible mechanisms (i.e., suboptimal socioeconomic status (SES), psychological adversities, unfavorable lifestyle, and biological alterations) on the observed associations.

RESULTS

During an average follow-up of 30.8 years, 20,885 participants received a diagnosis of endocrine diseases. We observed an association between the cumulative number of experienced childhood maltreatment and increased risk of being diagnosed with any endocrine disease (adjusted hazard ratio (HR) = 1.10, 95% confidence interval 1.09-1.12). The HR was 1.26 (1.22-1.30) when comparing individuals ≥ 2 with those with < 2 experienced childhood maltreatment. We further noted the most pronounced associations for type 2 diabetes (1.40 (1.33-1.48)) and hypothalamic-pituitary-adrenal (HPA)-axis-related endocrine diseases (1.38 (1.17-1.62)), and the association was stronger for having multiple endocrine diseases, compared to having one (odds ratio (95% CI) = 1.24 (1.19-1.30), 1.35 (1.27-1.44), and 1.52 (1.52-1.53) for 1, 2, and ≥ 3, respectively). Sequential mediation analyses showed that the association between childhood maltreatment and endocrine diseases was consistently and most distinctly mediated by psychological adversities (15.38 ~ 44.97%), while unfavorable lifestyle (10.86 ~ 25.32%) was additionally noted for type 2 diabetes whereas suboptimal SES (14.42 ~ 39.33%) for HPA-axis-related endocrine diseases.

CONCLUSIONS

Our study demonstrates that adverse psychological sequel of childhood maltreatment constitutes the main pathway to multiple endocrine diseases, particularly type 2 diabetes and HPA-axis-related endocrine diseases. Therefore, increased access to evidence-based mental health services may also be pivotal in reducing the risk of endocrine diseases among childhood maltreatment-exposed individuals.

Assessing Direct and Spillover Effects of Intervention Packages in Network-Randomized Studies

Intervention packages may result in a greater public health impact than single interventions. Understanding the separate impact of each component in the overall package effectiveness can improve intervention delivery. We adapted an approach to evaluate the effects of a time-varying intervention package in a network-randomized study. In some network-randomized studies, only a subset of participants in exposed networks receive the intervention themselves. The spillover effect contrasts average potential outcomes if a person was not exposed themselves under intervention in the network versus no intervention in a control network. We estimated effects of components of the intervention package in HIV Prevention Trials Network 037, a Phase III network-randomized HIV prevention trial among people who inject drugs and their risk networks using Marginal Structural Models to adjust for time-varying confounding. The index participant in an intervention network received a peer education intervention initially at baseline, then boosters at 6 and 12 months. All participants were followed to ascertain HIV risk behaviors. There were 560 participants with at least one follow-up visit, 48% of whom were randomized to the intervention, and 1,598 participant-visits were observed. The spillover effect of the boosters in the presence of initial peer education training was a 39% rate reduction (Rate Ratio = 0.61; 95% confidence interval= 0.43, 0.87). These methods will be useful to evaluate intervention packages in studies with network features.

Estimating Subgroup Effects in Generalizability and Transportability Analyses

Methods for extending-generalizing or transporting-inferences from a randomized trial to a target population involve conditioning on a large set of covariates that is sufficient for rendering the randomized and nonrandomized groups exchangeable. Yet, decision makers are often interested in examining treatment effects in subgroups of the target population defined in terms of only a few discrete covariates. Here, we propose methods for estimating subgroup-specific potential outcome means and average treatment effects in generalizability and transportability analyses, using outcome model--based (g-formula), weighting, and augmented weighting estimators. We consider estimating subgroup-specific average treatment effects in the target population and its nonrandomized subset, and we provide methods that are appropriate both for nested and non-nested trial designs. As an illustration, we apply the methods to data from the Coronary Artery Surgery Study (North America, 1975-1996) to compare the effect of surgery plus medical therapy versus medical therapy alone for chronic coronary artery disease in subgroups defined by history of myocardial infarction.

Mediation analysis with case-control sampling: Identification and estimation in the presence of a binary mediator

With reference to a stratified case-control (CC) procedure based on a binary variable of primary interest, we derive the expression of the distortion induced by the sampling design on the parameters of the logistic model of a secondary variable. This is particularly relevant when performing mediation analysis (possibly in a causal framework) with stratified case-control (SCC) data in settings where both the outcome and the mediator are binary. Despite being designed for parametric identification, our strategy is general and can be used also in a nonparametric context. With reference to parametric estimation, we derive the maximum likelihood (ML) estimator and the M-estimator of the joint outcome-mediator parameter vector. We then conduct a simulation study focusing on the main causal mediation quantities (i.e., natural effects) and comparing M- and ML estimation to existing methods, based on weighting. As an illustrative example, we reanalyze a German CC data set in order to investigate whether the effect of reduced immunocompetency on listeriosis onset is mediated by the intake of gastric acid suppressors.

Racial and Ethnic Disparities in Phthalate Exposure and Preterm Birth: A Pooled Study of Sixteen U.S. Cohorts

BACKGROUND

Phthalate exposures are ubiquitous during pregnancy and may contribute to racial and ethnic disparities in preterm birth.

OBJECTIVES

We investigated race and ethnicity in the relationship between biomarkers of phthalate exposure and preterm birth by examining: a) how hypothetical reductions in racial and ethnic disparities in phthalate metabolites might reduce the probability of preterm birth; and b) exposure-response models stratified by race and ethnicity.

METHODS

We pooled individual-level data on 6,045 pregnancies from 16 U.S. cohorts. We investigated covariate-adjusted differences in nine urinary phthalate metabolite concentrations by race and ethnicity [non-Hispanic White (White, 43%), non-Hispanic Black (Black, 13%), Hispanic/Latina (38%), and Asian/Pacific Islander (3%)]. Using g-computation, we estimated changes in the probability of preterm birth under hypothetical interventions to eliminate disparities in levels of urinary phthalate metabolites by proportionally lowering average concentrations in Black and Hispanic/Latina participants to be approximately equal to the averages in White participants. We also used race and ethnicity-stratified logistic regression to characterize associations between phthalate metabolites and preterm birth.

RESULTS

In comparison with concentrations among White participants, adjusted mean phthalate metabolite concentrations were consistently higher among Black and Hispanic/Latina participants by 23%-148% and 4%-94%, respectively. Asian/Pacific Islander participants had metabolite levels that were similar to those of White participants. Hypothetical interventions to reduce disparities in metabolite mixtures were associated with lower probabilities of preterm birth for Black [13% relative reduction; 95% confidence interval (CI): - 34 % , 8.6%] and Hispanic/Latina (9% relative reduction; 95% CI: - 19 % , 0.8%) participants. Odds ratios for preterm birth in association with phthalate metabolites demonstrated heterogeneity by race and ethnicity for two individual metabolites (mono-n-butyl and monoisobutyl phthalate), with positive associations that were larger in magnitude observed among Black or Hispanic/Latina participants.

CONCLUSIONS

Phthalate metabolite concentrations differed substantially by race and ethnicity. Our results show hypothetical interventions to reduce population-level racial and ethnic disparities in biomarkers of phthalate exposure could potentially reduce the probability of preterm birth. https://doi.org/10.1289/EHP12831.

Addressing Systematic Missing Data in the Context of Causally Interpretable Meta-analysis

Evidence synthesis involves drawing conclusions from trial samples that may differ from the target population of interest, and there is often heterogeneity among trials in sample characteristics, treatment implementation, study design, and assessment of covariates. Stitching together this patchwork of evidence requires subject-matter knowledge, a clearly defined target population, and guidance on how to weigh evidence from different trials. Transportability analysis has provided formal identifiability conditions required to make unbiased causal inference in the target population. In this manuscript, we review these conditions along with an additional assumption required to address systematic missing data. The identifiability conditions highlight the importance of accounting for differences in treatment effect modifiers between the populations underlying the trials and the target population. We perform simulations to evaluate the bias of conventional random effect models and multiply imputed estimates using the pooled trials sample and describe causal estimators that explicitly address trial-to-target differences in key covariates in the context of systematic missing data. Results indicate that the causal transportability estimators are unbiased when treatment effect modifiers are accounted for in the analyses. Results also highlight the importance of carefully evaluating identifiability conditions for each trial to reduce bias due to differences in participant characteristics between trials and the target population. Bias can be limited by adjusting for covariates that are strongly correlated with missing treatment effect modifiers, including data from trials that do not differ from the target on treatment modifiers, and removing trials that do differ from the target and did not assess a modifier.

An application of the Causal Roadmap in two safety monitoring case studies: Causal inference and outcome prediction using electronic health record data

Background

Real-world data, such as administrative claims and electronic health records, are increasingly used for safety monitoring and to help guide regulatory decision-making. In these settings, it is important to document analytic decisions transparently and objectively to assess and ensure that analyses meet their intended goals.

Methods

The Causal Roadmap is an established framework that can guide and document analytic decisions through each step of the analytic pipeline, which will help investigators generate high-quality real-world evidence.

Results

In this paper, we illustrate the utility of the Causal Roadmap using two case studies previously led by workgroups sponsored by the Sentinel Initiative - a program for actively monitoring the safety of regulated medical products. Each case example focuses on different aspects of the analytic pipeline for drug safety monitoring. The first case study shows how the Causal Roadmap encourages transparency, reproducibility, and objective decision-making for causal analyses. The second case study highlights how this framework can guide analytic decisions beyond inference on causal parameters, improving outcome ascertainment in clinical phenotyping.

Conclusion

These examples provide a structured framework for implementing the Causal Roadmap in safety surveillance and guide transparent, reproducible, and objective analysis.

G-estimation of structural nested mean models for interval-censored data using pseudo-observations

Two large-scale randomized clinical trials compared fenofibrate and placebo in diabetic patients with pre-existing retinopathy (FIELD study) or risk factors (ACCORD trial) on an intention-to-treat basis and reported a significant reduction in the progression of diabetic retinopathy in the fenofibrate arms. However, their analyses involved complications due to intercurrent events, that is, treatment-switching and interval-censoring. This article addresses these problems involved in estimation of causal effects of long-term use of fibrates in a cohort study that followed patients with type 2 diabetes for 8 years. We propose structural nested mean models (SNMMs) of time-varying treatment effects and pseudo-observation estimators for interval-censored data. The first estimator for SNMMs uses a nonparametric maximum likelihood estimator (MLE) as a pseudo-observation, while the second estimator is based on MLE under a parametric piecewise exponential distribution. Through numerical studies with real and simulated datasets, the pseudo-observations estimators of causal effects using the nonparametric Wellner-Zhan estimator perform well even under dependent interval-censoring. Its application to the diabetes study revealed that the use of fibrates in the first 4 years reduced the risk of diabetic retinopathy but did not support its efficacy beyond 4 years.

Bicycle infrastructure and the incidence rate of crashes with cars: A case-control study with Strava data in Atlanta

Introduction

Bicycling has individual and collective health benefits. Safety concerns are a deterrent to bicycling. Incomplete data on bicycling volumes has limited epidemiologic research investigating safety impacts of bicycle infrastructure, such as protected bike lanes.

Methods

In this case-control study, set in Atlanta, Georgia, USA between 2016-10-01 and 2018-08-31, we estimated the incidence rate of police-reported crashes between bicyclists and motor vehicles (n = 124) on several types of infrastructure (off-street paved trails, protected bike lanes, buffered bike lanes, conventional bike lanes, and sharrows) per distance ridden and per intersection entered. To estimate underlying bicycling (the control series), we used a sample of high-resolution bicycling data from Strava, an app, combined with data from 15 on-the-ground bicycle counters to adjust for possible selection bias in the Strava data. We used model-based standardization to estimate effects of treatment on the treated.

Results

After adjustment for selection bias and confounding, estimated ratio effects on segments (excluding intersections) with protected bike lanes (incidence rate ratio [IRR] = 0.5 [95% confidence interval: 0.0, 2.5]) and buffered bike lanes (IRR = 0 [0,0]) were below 1, but were above 1 on conventional bike lanes (IRR = 2.8 [1.2, 6.0]) and near null on sharrows (IRR = 1.1 [0.2, 2.9]). Per intersection entry, estimated ratio effects were above 1 for entries originating from protected bike lanes (incidence proportion ratio [IPR] = 3.0 [0.0, 10.8]), buffered bike lanes (IPR = 16.2 [0.0, 53.1]), and conventional bike lanes (IPR = 3.2 [1.8, 6.0]), and were near 1 and below 1, respectively, for those originating from sharrows (IPR = 0.9 [0.2, 2.1]) and off-street paved trails (IPR = 0.7 [0.0, 2.9]).

Conclusions

Protected bike lanes and buffered bike lanes had estimated protective effects on segments between intersections but estimated harmful effects at intersections. Conventional bike lanes had estimated harmful effects along segments and at intersections.

Long-term causal effects of far-right terrorism in New Zealand

The Christchurch mosque attacks in 2019, committed by a radical right-wing extremist, resulted in the tragic loss of 51 lives. Following these events, there was a noticable rise in societal acceptance of Muslim minorities. Comparable transient reactions have been observed elsewhere. However, the critical questions remain: can these effects endure? Are enduring effects evident across the political spectrum? It is challenging to answer such questions because identifying long-term causal effects requires estimating unobserved attitudinal trajectories without the attacks. Here, we use six preattack waves of Muslim acceptance responses from the New Zealand Attitudes and Values Study (NZAVS) to infer missing counterfactual trajectories (NZAVS cohort 2012, N = 4,865 ; replicated in 2013 cohort, N = 7,894 ). We find (1) the attacks initially boosted Muslim acceptance; (2) the magnitude of the initial Muslim acceptance boost was similar across the political spectrum; (3) no changes were observed in negative control groups; and (4) two- and three-year effects varied by baseline political orientation: liberal acceptance was stable, conservative acceptance grew relative to the counterfactual trend. Overall, the attacks added five years of growth in Muslim acceptance, with no regression to preattack levels over time. Continued growth among conservatives highlights the attack's failure to divide society. These results demonstrate the utility of combining methods for causal inference with national-scale panel data to answer psychological questions of basic human concern.

Tell me what you want, what you really really want: Estimands in observational pharmacoepidemiologic comparative effectiveness and safety studies

PURPOSE

Ideally, the objectives of a pharmacoepidemiologic comparative effectiveness or safety study should dictate its design and data analysis. This paper discusses how defining an estimand is instrumental to this process.

METHODS

We applied the ICH-E9 (Statistical Principles for Clinical Trials) R1 addendum on estimands - which originally focused on randomized trials - to three examples of observational pharmacoepidemiologic comparative effectiveness and safety studies. Five key elements specify the estimand: the population, contrasted treatments, endpoint, intercurrent events, and population-level summary measure.

RESULTS

Different estimands were defined for case studies representing three types of pharmacological treatments: (1) single-dose treatments using a case study about the effect of influenza vaccination versus no vaccination on mortality risk in an adult population of ≥60 years of age; (2) sustained-treatments using a case study about the effect of dipeptidyl peptidase 4 inhibitor versus glucagon-like peptide-1 agonist on hypoglycemia risk in treatment of uncontrolled diabetes; and (3) as needed treatments using a case study on the effect of nitroglycerin spray as-needed versus no nitroglycerin on syncope risk in treatment of stabile angina pectoris.

CONCLUSIONS

The case studies illustrated that a seemingly clear research question can still be open to multiple interpretations. Defining an estimand ensures that the study targets a treatment effect that aligns with the treatment decision the study aims to inform. Estimand definitions further help to inform choices regarding study design and data-analysis and clarify how to interpret study findings.

Examining the Causal Mediating Role of Cardiovascular Disease on the Effect of Subclinical Cardiovascular Disease on Cognitive Impairment via Separable Effects

BACKGROUND

An important epidemiological question is understanding how vascular risk factors contribute to cognitive impairment. Using data from the Cardiovascular Health Cognition Study, we investigated how subclinical cardiovascular disease (sCVD) relates to cognitive impairment risk and the extent to which the hypothesized risk is mediated by the incidence of clinically manifested cardiovascular disease (CVD), both overall and within apolipoprotein E-4 (APOE-4) subgroups.

METHODS

We adopted a novel "separable effects" causal mediation framework that assumes that sCVD has separably intervenable atherosclerosis-related components. We then ran several mediation models, adjusting for key covariates.

RESULTS

We found that sCVD increased overall risk of cognitive impairment (risk ratio [RR] = 1.21, 95% confidence interval [CI]: 1.03, 1.44); however, there was little or no mediation by incident clinically manifested CVD (indirect effect RR = 1.02, 95% CI: 1.00, 1.03). We also found attenuated effects among APOE-4 carriers (total effect RR = 1.09, 95% CI: 0.81, 1.47; indirect effect RR = 0.99, 95% CI: 0.96, 1.01) and stronger findings among noncarriers (total effect RR = 1.29, 95% CI: 1.05, 1.60; indirect effect RR = 1.02, 95% CI: 1.00, 1.05). In secondary analyses restricting cognitive impairment to only incident dementia cases, we found similar effect patterns.

CONCLUSIONS

We found that the effect of sCVD on cognitive impairment does not seem to be mediated by CVD, both overall and within APOE-4 subgroups. Our results were critically assessed via sensitivity analyses, and they were found to be robust. Future work is needed to fully understand the relationship between sCVD, CVD, and cognitive impairment.

Causal inference in survival analysis using longitudinal observational data: Sequential trials and marginal structural models

Longitudinal observational data on patients can be used to investigate causal effects of time-varying treatments on time-to-event outcomes. Several methods have been developed for estimating such effects by controlling for the time-dependent confounding that typically occurs. The most commonly used is marginal structural models (MSM) estimated using inverse probability of treatment weights (IPTW) (MSM-IPTW). An alternative, the sequential trials approach, is increasingly popular, and involves creating a sequence of "trials" from new time origins and comparing treatment initiators and non-initiators. Individuals are censored when they deviate from their treatment assignment at the start of each "trial" (initiator or noninitiator), which is accounted for using inverse probability of censoring weights. The analysis uses data combined across trials. We show that the sequential trials approach can estimate the parameters of a particular MSM. The causal estimand that we focus on is the marginal risk difference between the sustained treatment strategies of "always treat" vs "never treat." We compare how the sequential trials approach and MSM-IPTW estimate this estimand, and discuss their assumptions and how data are used differently. The performance of the two approaches is compared in a simulation study. The sequential trials approach, which tends to involve less extreme weights than MSM-IPTW, results in greater efficiency for estimating the marginal risk difference at most follow-up times, but this can, in certain scenarios, be reversed at later time points and relies on modelling assumptions. We apply the methods to longitudinal observational data from the UK Cystic Fibrosis Registry to estimate the effect of dornase alfa on survival.

Efficient and robust methods for causally interpretable meta-analysis: Transporting inferences from multiple randomized trials to a target population

We present methods for causally interpretable meta-analyses that combine information from multiple randomized trials to draw causal inferences for a target population of substantive interest. We consider identifiability conditions, derive implications of the conditions for the law of the observed data, and obtain identification results for transporting causal inferences from a collection of independent randomized trials to a new target population in which experimental data may not be available. We propose an estimator for the potential outcome mean in the target population under each treatment studied in the trials. The estimator uses covariate, treatment, and outcome data from the collection of trials, but only covariate data from the target population sample. We show that it is doubly robust in the sense that it is consistent and asymptotically normal when at least one of the models it relies on is correctly specified. We study the finite sample properties of the estimator in simulation studies and demonstrate its implementation using data from a multicenter randomized trial.

Clarifying the effect of biodiversity on productivity in natural ecosystems with longitudinal data and methods for causal inference

Causal effects of biodiversity on ecosystem functions can be estimated using experimental or observational designs - designs that pose a tradeoff between drawing credible causal inferences from correlations and drawing generalizable inferences. Here, we develop a design that reduces this tradeoff and revisits the question of how plant species diversity affects productivity. Our design leverages longitudinal data from 43 grasslands in 11 countries and approaches borrowed from fields outside of ecology to draw causal inferences from observational data. Contrary to many prior studies, we estimate that increases in plot-level species richness caused productivity to decline: a 10% increase in richness decreased productivity by 2.4%, 95% CI [-4.1, -0.74]. This contradiction stems from two sources. First, prior observational studies incompletely control for confounding factors. Second, most experiments plant fewer rare and non-native species than exist in nature. Although increases in native, dominant species increased productivity, increases in rare and non-native species decreased productivity, making the average effect negative in our study. By reducing the tradeoff between experimental and observational designs, our study demonstrates how observational studies can complement prior ecological experiments and inform future ones.

Depression networks: a systematic review of the network paradigm causal assumptions

The network paradigm for psychiatric disorder nosology was proposed based on the hypothesis that mental disorders are caused by networks of symptoms that are themselves causally related. Researchers have widely applied and integrated this paradigm to examine a variety of mental disorders, particularly depression. Existing studies generally focus on the correlation structure of symptoms, inferring causal relationships. Thus, presumption of causality may not be justified. The goal of this review was to examine the assumptions necessary for causal inference in network studies of depression. Specifically, we examined whether and how network studies address common violations of causal assumptions (i.e. no measurement error, exchangeability, and positivity). Of the 41 studies reviewed, five (12%) studies discussed sources of confounding unrelated to measurement error; none discussed positivity; and five conducted post-hoc analysis for measurement error. Depression network studies, in principle, are conducted under the assumption that symptom relationships are causal. Yet, in practice, studies seldomly discussed or adequately tested assumptions required to infer causality. Researchers continue to design studies that are unable to support the credibility of the network paradigm for the study of depression. There is a critical need to ensure scientific efforts cease to perpetuate problematic designs and findings to a potentially unsubstantiated paradigm.

Disentangling causality: assumptions in causal discovery and inference

Causality has been a burgeoning field of research leading to the point where the literature abounds with different components addressing distinct parts of causality. For researchers, it has been increasingly difficult to discern the assumptions they have to abide by in order to glean sound conclusions from causal concepts or methods. This paper aims to disambiguate the different causal concepts that have emerged in causal inference and causal discovery from observational data by attributing them to different levels of Pearl’s Causal Hierarchy. We will provide the reader with a comprehensive arrangement of assumptions necessary to engage in causal reasoning at the desired level of the hierarchy. Therefore, the assumptions underlying each of these causal concepts will be emphasized and their concomitant graphical components will be examined. We show which assumptions are necessary to bridge the gaps between causal discovery, causal identification and causal inference from a parametric and a non-parametric perspective. Finally, this paper points to further research areas related to the strong assumptions that researchers have glibly adopted to take part in causal discovery, causal identification and causal inference.

Regression-based estimation of heterogeneous treatment effects when extending inferences from a randomized trial to a target population

Most work on extending (generalizing or transporting) inferences from a randomized trial to a target population has focused on estimating average treatment effects (i.e., averaged over the target population's covariate distribution). Yet, in the presence of strong effect modification by baseline covariates, the average treatment effect in the target population may be less relevant for guiding treatment decisions. Instead, the conditional average treatment effect (CATE) as a function of key effect modifiers may be a more useful estimand. Recent work on estimating target population CATEs using baseline covariate, treatment, and outcome data from the trial and covariate data from the target population only allows for the examination of heterogeneity over distinct subgroups. We describe flexible pseudo-outcome regression modeling methods for estimating target population CATEs conditional on discrete or continuous baseline covariates when the trial is embedded in a sample from the target population (i.e., in nested trial designs). We construct pointwise confidence intervals for the CATE at a specific value of the effect modifiers and uniform confidence bands for the CATE function. Last, we illustrate the methods using data from the Coronary Artery Surgery Study (CASS) to estimate CATEs given history of myocardial infarction and baseline ejection fraction value in the target population of all trial-eligible patients with stable ischemic heart disease.

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.

Complete effect decomposition for an arbitrary number of multiple ordered mediators with time-varying confounders: A method for generalized causal multi-mediation analysis

Causal mediation analysis is advantageous for mechanism investigation. In settings with multiple causally ordered mediators, path-specific effects have been introduced to specify the effects of certain combinations of mediators. However, most path-specific effects are unidentifiable. An interventional analog of path-specific effects is adapted to address the non-identifiability problem. Moreover, previous studies only focused on cases with two or three mediators due to the complexity of the mediation formula in a large number of mediators. In this study, we provide a generalized definition of traditional path-specific effects and interventional path-specific effects with a recursive formula, along with the required assumptions for nonparametric identification. Subsequently, a general approach is developed with an arbitrary number of multiple ordered mediators and with time-varying confounders. All methods and software proposed in this study contribute to comprehensively decomposing a causal effect confirmed by data science and help disentangling causal mechanisms in the presence of complicated causal structures among multiple mediators.

Causal Mediation Analysis with Multiple Time-varying Mediators

In longitudinal studies with time-varying exposures and mediators, the mediational g-formula is an important method for the assessment of direct and indirect effects. However, current methodologies based on the mediational g-formula can deal with only one mediator. This limitation makes these methodologies inapplicable to many scenarios. Hence, we develop a novel methodology by extending the mediational g-formula to cover cases with multiple time-varying mediators. We formulate two variants of our approach that are each suited to a distinct set of assumptions and effect definitions and present nonparametric identification results of each variant. We further show how complex causal mechanisms (whose complexity derives from the presence of multiple time-varying mediators) can be untangled. We implemented a parametric method, along with a user-friendly algorithm, in R software. We illustrate our method by investigating the complex causal mechanism underlying the progression of chronic obstructive pulmonary disease. We found that the effects of lung function impairment mediated by dyspnea symptoms accounted for 14.6% of the total effect and that mediated by physical activity accounted for 11.9%. Our analyses thus illustrate the power of this approach, providing evidence for the mediating role of dyspnea and physical activity on the causal pathway from lung function impairment to health status. See video abstract at, http://links.lww.com/EDE/B988 .

Transportability Methods for Time-to-Event Outcomes: Application in Adjuvant Colon Cancer Trials

PURPOSE

Differences in the benefits of treatment on 5-year overall survival have been observed in 12 randomized phase III colon cancer adjuvant clinical trials from the ACCENT group. We investigated the reasons for these differences by incorporating the distribution of the observed covariates from each trial.

MATERIALS AND METHODS

We applied state-of-the-art transportability methods on the basis of causal inference, and compared them with a conventional meta-analysis approach to predict the treatment effect for the target population. Prediction errors were defined to evaluate whether the identifiability conditions necessary for causal inference were satisfied among the 12 trials, and to measure the performance of each method.

RESULTS

In the one-trial-at-a-time transportability analysis, the ranks of prediction errors for the target population were mostly consistent with the discrepancy in treatment effects among the 12 trials across the three models. The overall prediction errors between the leave-one-trial-out transportability method and the conventional individual participant data meta-analysis approach were very similar, and more than 40% lower than the overall prediction errors from the one-trial-at-a-time transportability method.

CONCLUSION

The discrepancy in treatment effects among the 12 trials is unlikely to arise from the choice of model specification or distribution of observed covariates but from the distribution of unobserved covariates or study-level features. The ability to quantify heterogeneity among the 12 trials was greatly reduced in both the leave-one-trial-out transportability method and the conventional meta-analysis approach compared with the one-trial-at-a-time transportability method.

Parametric G-computation for compatible indirect treatment comparisons with limited individual patient data

Population adjustment methods such as matching-adjusted indirect comparison (MAIC) are increasingly used to compare marginal treatment effects when there are cross-trial differences in effect modifiers and limited patient-level data. MAIC is based on propensity score weighting, which is sensitive to poor covariate overlap and cannot extrapolate beyond the observed covariate space. Current outcome regression-based alternatives can extrapolate but target a conditional treatment effect that is incompatible in the indirect comparison. When adjusting for covariates, one must integrate or average the conditional estimate over the relevant population to recover a compatible marginal treatment effect. We propose a marginalization method based on parametric G-computation that can be easily applied where the outcome regression is a generalized linear model or a Cox model. The approach views the covariate adjustment regression as a nuisance model and separates its estimation from the evaluation of the marginal treatment effect of interest. The method can accommodate a Bayesian statistical framework, which naturally integrates the analysis into a probabilistic framework. A simulation study provides proof-of-principle and benchmarks the method's performance against MAIC and the conventional outcome regression. Parametric G-computation achieves more precise and more accurate estimates than MAIC, particularly when covariate overlap is poor, and yields unbiased marginal treatment effect estimates under no failures of assumptions. Furthermore, the marginalized regression-adjusted estimates provide greater precision and accuracy than the conditional estimates produced by the conventional outcome regression, which are systematically biased because the measure of effect is non-collapsible.

Population attributable fraction based on marginal sufficient component cause model for mediation settings

PURPOSE

Population attributable fraction (PAF), defined as the proportion of the occurrence of a disease which will be reduced by eliminating risk factors in a population, is one of the most common measurements for evaluating the benefit of a health-related policy in epidemiologic study. In this article, we propose an alternative PAF defined based on sufficient cause framework, which decompose the occurrence of a disease into several pathways including mediation and mechanistic interaction.

METHODS

We propose a formal statistical definition and regression-based estimator for PAF based on sufficient cause framework within mediation settings. Under monotonicity assumption, the proposed method can decompose the occurrence of a disease into nine PAFs corresponding to all types of mechanisms attributing to exposure and the mediator, including the portion attributing to exposure directly, to mediator, to indirect effect through mediator, to the mechanistic interaction, to both of mediation and interaction, and to none of exposure or mediator.

RESULTS

We apply the proposed method to explore the mechanism of a hepatitis C virus (HCV)-induced hepatocellular carcinoma (HCC) mediated by and/or interacted with alanine aminotransferase (ALT) and hepatitis B virus (HBV). When treating ALT as mediator, 56.77% of diseased subjects can be attributable to either HCV or abnormal ALT. When treating HBV as mediator, HCC is mainly induced by an exogenous high HBV viral load directly.

CONCLUSIONS

The proposed method can identify the impact of exposure and pathway effects, and benefit to allocate the resources on intervention strategies.

Causal inference and effect estimation using observational data

Observational studies aiming to estimate causal effects often rely on conceptual frameworks that are unfamiliar to many researchers and practitioners. We provide a clear, structured overview of key concepts and terms, intended as a starting point for readers unfamiliar with the causal inference literature. First, we introduce theoretical frameworks underlying causal effect estimation methods: the counterfactual theory of causation, the potential outcomes framework, structural equations and directed acyclic graphs. Second, we define the most common causal effect estimands, and the issues of effect measure modification, interaction and mediation (direct and indirect effects). Third, we define the assumptions required to estimate causal effects: exchangeability, positivity, consistency and non-interference. Fourth, we define and explain biases that arise when attempting to estimate causal effects, including confounding, collider bias, selection bias and measurement bias. Finally, we describe common methods and study designs for causal effect estimation, including covariate adjustment, G-methods and natural experiment methods.

Body mass index trajectories in childhood and incidence rates of type 2 diabetes and coronary heart disease in adulthood: A cohort study

AIMS

We examined associations between five body mass index (BMI) trajectories from ages 6-15 years and register-based adult-onset type 2 diabetes mellitus (T2D) and coronary heart disease (CHD) with and without adjustment for adult BMI.

METHODS

Child and adult BMI came from two Danish cohorts and 13,205 and 13,438 individuals were included in T2D and CHD analyses, respectively. Trajectories were estimated by latent class modelling. Incidence rate ratios (IRRs) were estimated with Poisson regression.

RESULTS

In models without adult BMI, compared to the lowest trajectory, among men the T2D IRRs were 0.92 (95 %CI:0.77-1.09) for the second lowest trajectory and 1.51 (95 %CI:0.71-3.20) for the highest trajectory. The corresponding IRRs in women were 0.92 (95 %CI:0.74-1.16) and 3.58 (95 %CI:2.30-5.57). In models including adult BMI, compared to the lowest trajectory, T2D IRRs in men were 0.57 (95 %CI:0.47-0.68) for the second lowest trajectory and 0.26 (95 %CI:0.12-0.56) for the highest trajectory. The corresponding IRRs in women were 0.60 (95 %CI:0.48-0.75) and 0.59 (95 %CI:0.36-0.96). The associations were similar in direction, but not statistically significant, for CHD.

CONCLUSIONS

Incidence rates of adult-onset T2D were greater for a high child BMI trajectory than a low child BMI trajectory, but not in models that included adult BMI.

Toward a Clearer Definition of Selection Bias When Estimating Causal Effects

Selection bias remains a subject of controversy. Existing definitions of selection bias are ambiguous. To improve communication and the conduct of epidemiologic research focused on estimating causal effects, we propose to unify the various existing definitions of selection bias in the literature by considering any bias away from the true causal effect in the referent population (the population before the selection process), due to selecting the sample from the referent population, as selection bias. Given this unified definition, selection bias can be further categorized into two broad types: type 1 selection bias owing to restricting to one or more level(s) of a collider (or a descendant of a collider) and type 2 selection bias owing to restricting to one or more level(s) of an effect measure modifier. To aid in explaining these two types-which can co-occur-we start by reviewing the concepts of the target population, the study sample, and the analytic sample. Then, we illustrate both types of selection bias using causal diagrams. In addition, we explore the differences between these two types of selection bias, and describe methods to minimize selection bias. Finally, we use an example of "M-bias" to demonstrate the advantage of classifying selection bias into these two types.

Accounting for motion in resting-state fMRI: What part of the spectrum are we characterizing in autism spectrum disorder?

The exclusion of high-motion participants can reduce the impact of motion in functional Magnetic Resonance Imaging (fMRI) data. However, the exclusion of high-motion participants may change the distribution of clinically relevant variables in the study sample, and the resulting sample may not be representative of the population. Our goals are two-fold: 1) to document the biases introduced by common motion exclusion practices in functional connectivity research and 2) to introduce a framework to address these biases by treating excluded scans as a missing data problem. We use a study of autism spectrum disorder in children without an intellectual disability to illustrate the problem and the potential solution. We aggregated data from 545 children (8-13 years old) who participated in resting-state fMRI studies at Kennedy Krieger Institute (173 autistic and 372 typically developing) between 2007 and 2020. We found that autistic children were more likely to be excluded than typically developing children, with 28.5% and 16.1% of autistic and typically developing children excluded, respectively, using a lenient criterion and 81.0% and 60.1% with a stricter criterion. The resulting sample of autistic children with usable data tended to be older, have milder social deficits, better motor control, and higher intellectual ability than the original sample. These measures were also related to functional connectivity strength among children with usable data. This suggests that the generalizability of previous studies reporting naïve analyses (i.e., based only on participants with usable data) may be limited by the selection of older children with less severe clinical profiles because these children are better able to remain still during an rs-fMRI scan. We adapt doubly robust targeted minimum loss based estimation with an ensemble of machine learning algorithms to address these data losses and the resulting biases. The proposed approach selects more edges that differ in functional connectivity between autistic and typically developing children than the naïve approach, supporting this as a promising solution to improve the study of heterogeneous populations in which motion is common.

Toward evaluation of disseminated effects of medications for opioid use disorder within provider-based clusters using routinely-collected health data

Routinely-collected health data can be employed to emulate a target trial when randomized trial data are not available. Patients within provider-based clusters likely exert and share influence on each other's treatment preferences and subsequent health outcomes and this is known as dissemination or spillover. Extending a framework to replicate an idealized two-stage randomized trial using routinely-collected health data, an evaluation of disseminated effects within provider-based clusters is possible. In this article, we propose a novel application of causal inference methods for dissemination to retrospective cohort studies in administrative claims data and evaluate the impact of the normality of the random effects distribution for the cluster-level propensity score on estimation of the causal parameters. An extensive simulation study was conducted to study the robustness of the methods under different distributions of the random effects. We applied these methods to evaluate baseline prescription for medications for opioid use disorder among a cohort of patients diagnosed with opioid use disorder and adjust for baseline confounders using information obtained from an administrative claims database. We discuss future research directions in this setting to better address unmeasured confounding in the presence of disseminated effects.

Two-stage matching-adjusted indirect comparison

BACKGROUND

Anchored covariate-adjusted indirect comparisons inform reimbursement decisions where there are no head-to-head trials between the treatments of interest, there is a common comparator arm shared by the studies, and there are patient-level data limitations. Matching-adjusted indirect comparison (MAIC), based on propensity score weighting, is the most widely used covariate-adjusted indirect comparison method in health technology assessment. MAIC has poor precision and is inefficient when the effective sample size after weighting is small.

METHODS

A modular extension to MAIC, termed two-stage matching-adjusted indirect comparison (2SMAIC), is proposed. This uses two parametric models. One estimates the treatment assignment mechanism in the study with individual patient data (IPD), the other estimates the trial assignment mechanism. The first model produces inverse probability weights that are combined with the odds weights produced by the second model. The resulting weights seek to balance covariates between treatment arms and across studies. A simulation study provides proof-of-principle in an indirect comparison performed across two randomized trials. Nevertheless, 2SMAIC can be applied in situations where the IPD trial is observational, by including potential confounders in the treatment assignment model. The simulation study also explores the use of weight truncation in combination with MAIC for the first time.

RESULTS

Despite enforcing randomization and knowing the true treatment assignment mechanism in the IPD trial, 2SMAIC yields improved precision and efficiency with respect to MAIC in all scenarios, while maintaining similarly low levels of bias. The two-stage approach is effective when sample sizes in the IPD trial are low, as it controls for chance imbalances in prognostic baseline covariates between study arms. It is not as effective when overlap between the trials' target populations is poor and the extremity of the weights is high. In these scenarios, truncation leads to substantial precision and efficiency gains but induces considerable bias. The combination of a two-stage approach with truncation produces the highest precision and efficiency improvements.

CONCLUSIONS

Two-stage approaches to MAIC can increase precision and efficiency with respect to the standard approach by adjusting for empirical imbalances in prognostic covariates in the IPD trial. Further modules could be incorporated for additional variance reduction or to account for missingness and non-compliance in the IPD trial.

Have Paved Trails and Protected Bike Lanes Led to More Bicycling in Atlanta?: A Generalized Synthetic-Control Analysis

BACKGROUND

Bicycling is an important form of physical activity in populations. Research assessing the effect of infrastructure on bicycling with high-resolution smartphone data is emerging in several places, but it remains limited in low-bicycling US settings, including the Southeastern US. The Atlanta area has been expanding its bicycle infrastructure, including off-street paved trails such as the Atlanta BeltLine and some protected bike lanes.

METHODS

Using the generalized synthetic-control method, we estimated effects of five groups of off-street paved trails and protected bike lanes on bicycle ridership in their corresponding areas. To measure bicycling, we used 2 years (October 1, 2016 to September 30, 2018) of monthly Strava data in Atlanta's urban core along with data from 15 on-the-ground counters to adjust for spatiotemporal variation in app use.

RESULTS

Considering all infrastructure as one joint intervention, an estimated 1.10 (95% confidence interval [CI]: 0.99, 1.18) times more bicycle-distance was ridden than would have been expected in the same areas had the infrastructure not been built, when defining treatment areas by the narrower of two definitions (defined in text). The Atlanta BeltLine Westside Trail and Proctor Creek Greenway had especially strong effect estimates, e.g., ratios of 1.45 (95% CI: 1.12, 1.86) and 1.55 (1.10, 2.14) under each treatment-area definition, respectively. We estimated that other infrastructure had weaker positive or no effects on bicycle-distance ridden.

CONCLUSIONS

This study advances research on the topic because of its setting in the US Southeast, simultaneous assessment of several infrastructure groups, and data-driven approach to estimating effects. See video abstract at, http://links.lww.com/EDE/B936.

Estimation of causal effects of a time-varying exposure at multiple time points through multivariable mendelian randomization

Mendelian Randomisation (MR) is a powerful tool in epidemiology that can be used to estimate the causal effect of an exposure on an outcome in the presence of unobserved confounding, by utilising genetic variants as instrumental variables (IVs) for the exposure. The effect estimates obtained from MR studies are often interpreted as the lifetime effect of the exposure in question. However, the causal effects of some exposures are thought to vary throughout an individual's lifetime with periods during which an exposure has a greater effect on a particular outcome. Multivariable MR (MVMR) is an extension of MR that allows for multiple, potentially highly related, exposures to be included in an MR estimation. MVMR estimates the direct effect of each exposure on the outcome conditional on all the other exposures included in the estimation. We explore the use of MVMR to estimate the direct effect of a single exposure at different time points in an individual's lifetime on an outcome. We use simulations to illustrate the interpretation of the results from such analyses and the key assumptions required. We show that causal effects at different time periods can be estimated through MVMR when the association between the genetic variants used as instruments and the exposure measured at those time periods varies. However, this estimation will not necessarily identify exact time periods over which an exposure has the most effect on the outcome. Prior knowledge regarding the biological basis of exposure trajectories can help interpretation. We illustrate the method through estimation of the causal effects of childhood and adult BMI on C-Reactive protein and smoking behaviour.

Risk of large for gestational age births at early, full and late term in relation to pre-pregnancy body mass index: Mediation by gestational diabetes status

BACKGROUND

Maternal pre-pregnancy body mass index (BMI) is strongly associated with infant birthweight and the risk differs in pregnancies complicated by gestational diabetes (GDM).

OBJECTIVES

To examine the risk of large for gestational age (LGA) (≥97th percentile) singleton births at early term, full term and late term in relation to maternal pre-pregnancy BMI status mediated through GDM.

METHODS

We analysed data from the 2018 U.S. National Vital Statistics Natality File restricted to singleton term births (N = 3,229,783). In counterfactual models for causal inference, we estimated the total effect (TE), natural direct effect (NDE) and natural indirect effect (NIE) for the association of pre-pregnancy BMI with subcategories of LGA births at early, full and late term mediated through GDM, using log-binomial regression and adjusting for race/ethnicity, age, education, parity and infant sex. Proportion mediated was calculated on the risk difference scale and potential unmeasured confounders were assessed using the E-value.

RESULTS

Overall, 6.4% of women had GDM, and there were 3.6% LGA singleton term births. The highest prevalence of GDM was among pre-gestational overweight/obesity that also had the highest rates of LGA births at term. The TE estimates for the risk of LGA births were the strongest across women with higher pre-pregnancy BMI compared to women with normal pre-pregnancy BMI. The NDE estimates were higher than the NIE estimates for overweight/obese BMI status. The proportion mediated, which answers the causal question to what extent the total effect of the association between pre-pregnancy BMI and LGA births is accounted for through GDM, was the highest (up to 16%) for early term births.

CONCLUSIONS

Term singleton births make up the largest proportion in a cohort of newborns. While the percentage mediated through GDM was relatively small, health risks arising from pre-pregnancy overweight, and obesity can be substantial to both mothers and their offspring.

Synthetic Negative Controls: Using Simulation to Screen Large-scale Propensity Score Analyses

The propensity score has become a standard tool to control for large numbers of variables in healthcare database studies. However, little has been written on the challenge of comparing large-scale propensity score analyses that use different methods for confounder selection and adjustment. In these settings, balance diagnostics are useful but do not inform researchers on which variables balance should be assessed or quantify the impact of residual covariate imbalance on bias. Here, we propose a framework to supplement balance diagnostics when comparing large-scale propensity score analyses. Instead of focusing on results from any single analysis, we suggest conducting and reporting results for many analytic choices and using both balance diagnostics and synthetically generated control studies to screen analyses that show signals of bias caused by measured confounding. To generate synthetic datasets, the framework does not require simulating the outcome-generating process. In healthcare database studies, outcome events are often rare, making it difficult to identify and model all predictors of the outcome to simulate a confounding structure closely resembling the given study. Therefore, the framework uses a model for treatment assignment to divide the comparator population into pseudo-treatment groups where covariate differences resemble those in the study cohort. The partially simulated datasets have a confounding structure approximating the study population under the null (synthetic negative control studies). The framework is used to screen analyses that likely violate partial exchangeability due to lack of control for measured confounding. We illustrate the framework using simulations and an empirical example.

Generalizing trial evidence to target populations in non-nested designs: Applications to AIDS clinical trials

Comparative effectiveness evidence from randomized trials may not be directly generalizable to a target population of substantive interest when, as in most cases, trial participants are not randomly sampled from the target population. Motivated by the need to generalize evidence from two trials conducted in the AIDS Clinical Trials Group (ACTG), we consider weighting, regression and doubly robust estimators to estimate the causal effects of HIV interventions in a specified population of people living with HIV in the USA. We focus on a non-nested trial design and discuss strategies for both point and variance estimation of the target population average treatment effect. Specifically in the generalizability context, we demonstrate both analytically and empirically that estimating the known propensity score in trials does not increase the variance for each of the weighting, regression and doubly robust estimators. We apply these methods to generalize the average treatment effects from two ACTG trials to specified target populations and operationalize key practical considerations. Finally, we report on a simulation study that investigates the finite-sample operating characteristics of the generalizability estimators and their sandwich variance estimators.