Search

An increasing number of recent studies have suggested that doubly robust estimators with cross-fitting should be used when estimating causal effects with machine learning methods. However, not all existing programs that implement doubly robust estimators support machine learning methods and cross-fitting, or provide estimates on multiplicative scales. To address these needs, we developed AIPW , a software package implementing augmented inverse probability weighting (AIPW) estimation of average causal effects in R (R Foundation for Statistical Computing, Vienna, Austria). Key features of the AIPW package include cross-fitting and flexible covariate adjustment for observational studies and randomized controlled trials (RCTs). In this paper, we use a simulated RCT to illustrate implementation of the AIPW estimator. We also perform a simulation study to evaluate the performance of the AIPW package compared with other doubly robust implementations, including CausalGAM , npcausal , tmle , and tmle3. Our simulation showed that the AIPW package yields performance comparable to that of other programs. Furthermore, we also found that cross-fitting substantively decreases the bias and improves the confidence interval coverage for doubly robust estimators fitted with machine learning algorithms. Our findings suggest that the AIPW package can be a useful tool for estimating average causal effects with machine learning methods in RCTs and observational studies. [ABSTRACT FROM AUTHOR]

A conventional Mendelian randomization analysis assesses the causal effect of a risk factor on an outcome by using genetic variants that are solely associated with the risk factor of interest as instrumental variables. However, in some cases, such as the case of triglyceride level as a risk factor for cardiovascular disease, it may be difficult to find a relevant genetic variant that is not also associated with related risk factors, such as other lipid fractions. Such a variant is known as pleiotropic. In this paper, we propose an extension of Mendelian randomization that uses multiple genetic variants associated with several measured risk factors to simultaneously estimate the causal effect of each of the risk factors on the outcome. This “multivariable Mendelian randomization” approach is similar to the simultaneous assessment of several treatments in a factorial randomized trial. In this paper, methods for estimating the causal effects are presented and compared using real and simulated data, and the assumptions necessary for a valid multivariable Mendelian randomization analysis are discussed. Subject to these assumptions, we demonstrate that triglyceride-related pathways have a causal effect on the risk of coronary heart disease independent of the effects of low-density lipoprotein cholesterol and high-density lipoprotein cholesterol. [ABSTRACT FROM PUBLISHER]

Longitudinal data sources offer new opportunities for the evaluation of sequential interventions. To adjust for time-dependent confounding in these settings, longitudinal targeted maximum likelihood based estimation (TMLE), a doubly robust method that can be coupled with machine learning, has been proposed. This paper provides a tutorial in applying longitudinal TMLE, in contrast to inverse probability of treatment weighting and g-computation based on iterative conditional expectations. We apply these methods to estimate the causal effect of nutritional interventions on clinical outcomes among critically ill children in a United Kingdom study (Control of Hyperglycemia in Paediatric Intensive Care, 2008-2011). We estimate the probability of a child's being discharged alive from the pediatric intensive care unit by a given day, under a range of static and dynamic feeding regimes. We find that before adjustment, patients who follow the static regime "never feed" are discharged by the end of the fifth day with a probability of 0.88 (95% confidence interval: 0.87, 0.90), while for the patients who follow the regime "feed from day 3," the probability of discharge is 0.64 (95% confidence interval: 0.62, 0.66). After adjustment for time-dependent confounding, most of this difference disappears, and the statistical methods produce similar results. TMLE offers a flexible estimation approach; hence, we provide practical guidance on implementation to encourage its wider use. [ABSTRACT FROM AUTHOR]

Influenza viruses undergo frequent antigenic changes. As a result, the viruses circulating change within and between seasons, and the composition of the influenza vaccine is updated annually. Thus, estimation of the vaccine's effectiveness is not constant across seasons. In order to provide annual estimates of the influenza vaccine's effectiveness, health departments have increasingly adopted the "test-negative design," using enhanced data from routine surveillance systems. In this design, patients presenting to participating general practitioners with influenza-like illness are swabbed for laboratory testing; those testing positive for influenza virus are defined as cases, and those testing negative form the comparison group. Data on patients' vaccination histories and confounder profiles are also collected. Vaccine effectiveness is estimated from the odds ratio comparing the odds of testing positive for influenza among vaccinated patients and unvaccinated patients, adjusting for confounders. The test-negative design is purported to reduce bias associated with confounding by health-care-seeking behavior and misclassification of cases. In this paper, we use directed acyclic graphs to characterize potential biases in studies of influenza vaccine effectiveness using the test-negative design. We show how studies using this design can avoid or minimize bias and where bias may be introduced with particular study design variations. [ABSTRACT FROM AUTHOR]

Inverse probability weighting (IPW), a well-established method of controlling for confounding in observational studies with binary exposures, has been extended to analyses with continuous exposures. Methods developed for continuous exposures may not apply when the exposure is quasicontinuous because of irregular exposure distributions that violate key assumptions. We used simulations and cluster-randomized clinical trial data to assess 4 approaches developed for continuous exposures—ordinary least squares (OLS), covariate balancing generalized propensity scores (CBGPS), nonparametric covariate balancing generalized propensity scores (npCBGPS), and quantile binning (QB)—and a novel method, a cumulative probability model (CPM), in quasicontinuous exposure settings. We compared IPW stability, covariate balance, bias, mean squared error, and standard error estimation across 3,000 simulations with 6 different quasicontinuous exposures, varying in skewness and granularity. In general, CBGPS and npCBGPS resulted in excellent covariate balance, and npCBGPS was the least biased but the most variable. The QB and CPM approaches had the lowest mean squared error, particularly with marginally skewed exposures. We then successfully applied the IPW approaches, together with missing-data techniques, to assess how session attendance (out of a possible 15) in a partners-based clustered intervention among pregnant couples living with human immunodeficiency virus in Mozambique (2017–2022) influenced postpartum contraceptive uptake. [ABSTRACT FROM AUTHOR]

In this article, I provide a perspective on the tobacco epidemic and epidemiology, describing the impact of the tobacco-caused disease epidemic on the field of epidemiology. Although there is an enormous body of epidemiologic evidence on the associations of smoking with health, little systematic attention has been given to how decades of research have affected epidemiology and its practice. I address the many advances that resulted from epidemiologic research on smoking and health, such as demonstration of the utility of observational designs and important parameters (the odds ratio and the population attributable risk), guidelines for causal inference, and systematic review approaches. I also cover unintended and adverse consequences for the field, including the strategy of doubt creation and the recruitment of epidemiologists by the tobacco industry to serve its mission. The paradigm of evidence-based action for addressing noncommunicable diseases began with the need to address the epidemic of tobacco-caused disease, an imperative for action documented by epidemiologic research. [ABSTRACT FROM AUTHOR]

Machine learning is gaining prominence in the health sciences, where much of its use has focused on data-driven prediction. However, machine learning can also be embedded within causal analyses, potentially reducing biases arising from model misspecification. Using a question-and-answer format, we provide an introduction and orientation for epidemiologists interested in using machine learning but concerned about potential bias or loss of rigor due to use of "black box" models. We conclude with sample software code that may lower the barrier to entry to using these techniques. [ABSTRACT FROM AUTHOR]

Matching methods are assumed to reduce the likelihood of a biased inference compared with ordinary least squares (OLS) regression. Using simulations, we compared inferences from propensity score matching, coarsened exact matching, and unmatched covariate-adjusted OLS regression to identify which methods, in which scenarios, produced unbiased inferences at the expected type I error rate of 5%. We simulated multiple data sets and systematically varied common support, discontinuities in the exposure and/or outcome, exposure prevalence, and analytical model misspecification. Matching inferences were often biased in comparison with OLS, particularly when common support was poor; when analysis models were correctly specified and common support was poor, the type I error rate was 1.6% for propensity score matching (statistically inefficient), 18.2% for coarsened exact matching (high), and 4.8% for OLS (expected). Our results suggest that when estimates from matching and OLS are similar (i.e. confidence intervals overlap), OLS inferences are unbiased more often than matching inferences; however, when estimates from matching and OLS are dissimilar (i.e. confidence intervals do not overlap), matching inferences are unbiased more often than OLS inferences. This empirical "rule of thumb" may help applied researchers identify situations in which OLS inferences may be unbiased as compared with matching inferences. [ABSTRACT FROM AUTHOR]

The popular inverse probability weighting method in causal inference is often hampered by extreme propensity scores, resulting in biased estimates and excessive variance. A common remedy is to trim patients with extreme scores (i.e. remove them from the weighted analysis). However, such methods are often sensitive to the choice of cutoff points and discard a large proportion of the sample. The implications for bias and the precision of the treatment effect estimate are unclear. These problems are mitigated by a newly developed method, the overlap weighting method. Overlap weights emphasize the target population with the most overlap in observed characteristics between treatments, by continuously down-weighting the units in the tails of the propensity score distribution. Here we use simulations to compare overlap weights to standard inverse probability weighting with trimming, in terms of bias, variance, and 95% confidence interval coverage. A range of propensity score distributions are considered, including settings with substantial nonoverlap and extreme values. To facilitate practical implementation, we further provide a consistent estimator for the standard error of the treatment effect estimated using overlap weighting. [ABSTRACT FROM AUTHOR]

The advent of counterfactual-based mediation analysis has triggered enormous progress on how, and under what assumptions, one may disentangle path-specific effects upon combining arbitrary (possibly nonlinear) models for mediator and outcome. However, current developments have largely focused on single mediators because required identification assumptions prohibit simple extensions to settings with multiple mediators that may depend on one another. In this article, we propose a procedure for obtaining fine-grained decompositions that may still be recovered fromobserved data in such complex settings.We first show that existing analytical approaches target specific instances of a more general set of decompositions and may therefore fail to provide a comprehensive assessment of the processes that underpin cause-effect relationships between exposure and outcome. We then outline conditions for obtaining the remaining set of decompositions. Because the number of targeted decompositions increases rapidly with the number of mediators, we introduce natural effects models along with estimation methods that allow for flexible and parsimonious modeling. Our procedure can easily be implemented using off-the-shelf software and is illustrated using a reanalysis of theWorld Health Organization's Large Analysis and Review of European Housing and Health Status (WHO-LARES) study on the effect ofmold exposure onmental health (2002-2003). [ABSTRACT FROM AUTHOR]

A propensity score (PS) model's ability to control confounding can be assessed by evaluating covariate balance across exposure groups after PS adjustment. The optimal strategy for evaluating a disease risk score (DRS) model's ability to control confounding is less clear. DRS models cannot be evaluated through balance checks within the full population, and they are usually assessed through prediction diagnostics and goodness-of-fit tests. A proposed alternative is the "dry-run" analysis, which divides the unexposed population into "pseudo-exposed" and "pseudo-unexposed" groups so that differences on observed covariates resemble differences between the actual exposed and unexposed populations. With no exposure effect separating the pseudo-exposed and pseudounexposed groups, a DRS model is evaluated by its ability to retrieve an unconfounded null estimate after adjustment in this pseudo-population. We used simulations and an empirical example to compare traditional DRS performance metrics with the dry-run validation. In simulations, the dry run often improved assessment of confounding control, compared with the C statistic and goodness-of-fit tests. In the empirical example, PS and DRS matching gave similar results and showed good performance in terms of covariate balance (PS matching) and controlling confounding in the dry-run analysis (DRS matching). The dry-run analysis may prove useful in evaluating confounding control through DRS models. [ABSTRACT FROM AUTHOR]

Traditional epidemiologic approaches allow us to compare counterfactual outcomes under 2 exposure distributions, usually 100% exposed and 100% unexposed. However, to estimate the population health effect of a proposed intervention, one may wish to compare factual outcomes under the observed exposure distribution to counterfactual outcomes under the exposure distribution produced by an intervention. Here, we used inverse probability weights to compare the 5-year mortality risk under observed antiretroviral therapy treatment plans to the 5-year mortality risk that would had been observed under an intervention in which all patients initiated therapy immediately upon entry into care among patients positive for human immunodeficiency virus in the US Centers for AIDS Research Network of Integrated Clinical Systems multisite cohort study between 1998 and 2013. Therapy-naïve patients (n = 14,700) were followed from entry into care until death, loss to follow-up, or censoring at 5 years or on December 31, 2013. The 5-year cumulative incidence of mortality was 11.65% under observed treatment plans and 10.10% under the intervention, yielding a risk difference of -1.57% (95% confidence interval: -3.08, -0.06). Comparing outcomes under the intervention with outcomes under observed treatment plans provides meaningful information about the potential consequences of new US guidelines to treat all patients with human immunodeficiency virus regardless of CD4 cell count under actual clinical conditions. [ABSTRACT FROM AUTHOR]

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