Search

ABSTRACT This paper addresses strategies for selecting variables for adjustment in non-experimental comparative effectiveness research and uses causal graphs to illustrate the causal network that relates treatment to outcome. Variables in the causal network take on multiple structural forms. Adjustment for a common cause pathway between treatment and outcome can remove confounding, whereas adjustment for other structural types may increase bias. For this reason, variable selection would ideally be based on an understanding of the causal network; however, the true causal network is rarely known. Therefore, we describe more practical variable selection approaches based on background knowledge when the causal structure is only partially known. These approaches include adjustment for all observed pretreatment variables thought to have some connection to the outcome, all known risk factors for the outcome, and all direct causes of the treatment or the outcome. Empirical approaches, such as forward and backward selection and automatic high-dimensional proxy adjustment, are also discussed. As there is a continuum between knowing and not knowing the causal, structural relations of variables, we recommend addressing variable selection in a practical way that involves a combination of background knowledge and empirical selection and that uses high-dimensional approaches. This empirical approach can be used to select from a set of a priori variables based on the researcher's knowledge to be included in the final analysis or to identify additional variables for consideration. This more limited use of empirically derived variables may reduce confounding while simultaneously reducing the risk of including variables that may increase bias. Copyright © 2013 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Purpose: Bootstrapping can account for uncertainty in propensity score (PS) estimation and matching processes in 1:1 PS‐matched cohort studies. While theory suggests that the classical bootstrap can fail to produce proper coverage, practical impact of this theoretical limitation in settings typical to pharmacoepidemiology is not well studied. Methods: In a plasmode‐based simulation study, we compared performance of the standard parametric approach, which ignores uncertainty in PS estimation and matching, with two bootstrapping methods. The first method only accounted for uncertainty introduced during the matching process (the observation resampling approach). The second method accounted for uncertainty introduced during both PS estimation and matching processes (the PS reestimation approach). Variance was estimated based on percentile and empirical standard errors, and treatment effect estimation was based on median and mean of the estimated treatment effects across 1000 bootstrap resamples. Two treatment prevalence scenarios (5% and 29%) across two treatment effect scenarios (hazard ratio of 1.0 and 2.0) were evaluated in 500 simulated cohorts of 10 000 patients each. Results: We observed that 95% confidence intervals from the bootstrapping approaches but not the standard approach, resulted in inaccurate coverage rates (98%‐100% for the observation resampling approach, 99%‐100% for the PS reestimation approach, and 95%‐96% for standard approach). Treatment effect estimation based on bootstrapping approaches resulted in lower bias than the standard approach (less than 1.4% vs 4.1%) at 5% treatment prevalence; however, the performance was equivalent at 29% treatment prevalence. Conclusion: Use of bootstrapping led to variance overestimation and inconsistent coverage, while coverage remained more consistent with parametric estimation. [ABSTRACT FROM AUTHOR]

ABSTRACT Purpose It is increasingly necessary to analyze data from multiple sources when conducting public health safety surveillance or comparative effectiveness research. However, security, privacy, proprietary, and legal concerns often reduce data holders' willingness to share highly granular information. We describe and compare two approaches that do not require sharing of patient-level information to adjust for confounding in multi-site studies. Methods We estimated the risks of angioedema associated with angiotensin-converting enzyme inhibitors (ACEIs), angiotensin receptor blockers (ARBs), and aliskiren in comparison with beta-blockers within Mini-Sentinel, which has created a distributed data system of 18 health plans. To obtain the adjusted hazard ratios (HRs) and 95% confidence intervals (CIs), we performed (i) a propensity score-stratified case-centered logistic regression analysis, a method identical to a stratified Cox regression analysis but needing only aggregated risk set data, and (ii) an inverse variance-weighted meta-analysis, which requires only the site-specific HR and variance. We also performed simulations to further compare the two methods. Results Compared with beta-blockers, the adjusted HR was 3.04 (95% CI: 2.81, 3.27) for ACEIs, 1.16 (1.00, 1.34) for ARBs, and 2.85 (1.34, 6.04) for aliskiren in the case-centered analysis. The corresponding HRs were 2.98 (2.76, 3.21), 1.15 (1.00, 1.33), and 2.86 (1.35, 6.04) in the meta-analysis. Simulations suggested that the two methods may produce different results under certain analytic scenarios. Conclusion The case-centered analysis and the meta-analysis produced similar results without the need to share patient-level data across sites in our empirical study, but may provide different results in other study settings. Copyright © 2013 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

ABSTRACT Purpose It is often preferable to simplify the estimation of treatment effects on multiple outcomes by using a single propensity score (PS) model. Variable selection in PS models impacts the efficiency and validity of treatment effects. However, the impact of different variable selection strategies on the estimated treatment effects in settings involving multiple outcomes is not well understood. The authors use simulations to evaluate the impact of different variable selection strategies on the bias and precision of effect estimates to provide insight into the performance of various PS models in settings with multiple outcomes. Methods Simulated studies consisted of dichotomous treatment, two Poisson outcomes, and eight standard-normal covariates. Covariates were selected for the PS models based on their effects on treatment, a specific outcome, or both outcomes. The PSs were implemented using stratification, matching, and weighting (inverse probability treatment weighting). Results PS models including only covariates affecting a specific outcome (outcome-specific models) resulted in the most efficient effect estimates. The PS model that only included covariates affecting either outcome (generic-outcome model) performed best among the models that simultaneously controlled measured confounding for both outcomes. Similar patterns were observed over the range of parameter values assessed and all PS implementation methods. Conclusions A single, generic-outcome model performed well compared with separate outcome-specific models in most scenarios considered. The results emphasize the benefit of using prior knowledge to identify covariates that affect the outcome when constructing PS models and support the potential to use a single, generic-outcome PS model when multiple outcomes are being examined. Copyright © 2012 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

ABSTRACT Background A correctly specified propensity score (PS) estimated in a cohort ('cohort PS') should, in expectation, remain valid in a subgroup population. Objective We sought to determine whether using a cohort PS can be validly applied to subgroup analyses and, thus, add efficiency to studies with many subgroups or restricted data. Methods In each of three cohort studies, we estimated a cohort PS, defined five subgroups, and then estimated subgroup-specific PSs. We compared difference in treatment effect estimates for subgroup analyses adjusted by cohort PSs versus subgroup-specific PSs. Then, over 10 million times, we simulated a population with known characteristics of confounding, subgroup size, treatment interactions, and treatment effect and again assessed difference in point estimates. Results We observed that point estimates in most subgroups were substantially similar with the two methods of adjustment. In simulations, the effect estimates differed by a median of 3.4% (interquartile (IQ) range 1.3-10.0%). The IQ range exceeded 10% only in cases where the subgroup had < 1000 patients or few outcome events. Conclusions Our empirical and simulation results indicated that using a cohort PS in subgroup analyses was a feasible approach, particularly in larger subgroups. Copyright © 2011 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

ABSTRACT Background Among the large number of cohort studies that employ propensity score matching, most match patients 1:1. Increasing the matching ratio is thought to improve precision but may come with a trade-off with respect to bias. Objective To evaluate several methods of propensity score matching in cohort studies through simulation and empirical analyses. Methods We simulated cohorts of 20 000 patients with exposure prevalence of 10%-50%. We simulated five dichotomous and five continuous confounders. We estimated propensity scores and matched using digit-based greedy ('greedy'), pairwise nearest neighbor within a caliper ('nearest neighbor'), and a nearest neighbor approach that sought to balance the scores of the comparison patient above and below that of the treated patient ('balanced nearest neighbor'). We matched at both fixed and variable matching ratios and also evaluated sequential and parallel schemes for the order of formation of 1: n match groups. We then applied this same approach to two cohorts of patients drawn from administrative claims data. Results Increasing the match ratio beyond 1:1 generally resulted in somewhat higher bias. It also resulted in lower variance with variable ratio matching but higher variance with fixed. The parallel approach generally resulted in higher mean squared error but lower bias than the sequential approach. Variable ratio, parallel, balanced nearest neighbor matching generally yielded the lowest bias and mean squared error. Conclusions 1: n matching can be used to increase precision in cohort studies. We recommend a variable ratio, parallel, balanced 1: n, nearest neighbor approach that increases precision over 1:1 matching at a small cost in bias. Copyright © 2012 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]