Evaluation of a geriatrics primary care model using prospective matching to guide enrollment

Background Few definitive guidelines exist for rigorous large-scale prospective evaluation of nonrandomized programs and policies that require longitudinal primary data collection. In Veterans Affairs (VA) we identified a need to understand the impact of a geriatrics primary care model (referred to as GeriPACT); however, randomization of patients to GeriPACT vs. a traditional PACT was not feasible because GeriPACT has been rolled out nationally, and the decision to transition from PACT to GeriPACT is made jointly by a patient and provider. We describe our study design used to evaluate the comparative effectiveness of GeriPACT compared to a traditional primary care model (referred to as PACT) on patient experience and quality of care metrics. Methods We used prospective matching to guide enrollment of GeriPACT-PACT patient dyads across 57 VA Medical Centers. First, we identified matches based an array of administratively derived characteristics using a combination of coarsened exact and distance function matching on 11 identified key variables that may function as confounders. Once a GeriPACT patient was enrolled, matched PACT patients were then contacted for recruitment using pre-assigned priority categories based on the distance function; if eligible and consented, patients were enrolled and followed with telephone surveys for 18 months. Results We successfully enrolled 275 matched dyads in near real-time, with a median time of 7 days between enrolling a GeriPACT patient and a closely matched PACT patient. Standardized mean differences of < 0.2 among nearly all baseline variables indicates excellent baseline covariate balance. Exceptional balance on survey-collected baseline covariates not available at the time of matching suggests our procedure successfully controlled many known, but administratively unobserved, drivers of entrance to GeriPACT. Conclusions We present an important process to prospectively evaluate the effects of different treatments when randomization is infeasible and provide guidance to researchers who may be interested in implementing a similar approach. Rich matching variables from the pre-treatment period that reflect treatment assignment mechanisms create a high quality comparison group from which to recruit. This design harnesses the power of national administrative data coupled with collection of patient reported outcomes, enabling rigorous evaluation of non-randomized programs or policies.

Applied comparison of large‐scale propensity score matching and cardinality matching for causal inference in observational research

Background Cardinality matching (CM), a novel matching technique, finds the largest matched sample meeting prespecified balance criteria thereby overcoming limitations of propensity score matching (PSM) associated with limited covariate overlap, which are especially pronounced in studies with small sample sizes. The current study proposes a framework for large-scale CM (LS-CM); and compares large-scale PSM (LS-PSM) and LS-CM in terms of post-match sample size, covariate balance and residual confounding at progressively smaller sample sizes. Methods Evaluation of LS-PSM and LS-CM within a comparative cohort study of new users of angiotensin-converting enzyme inhibitor (ACEI) and thiazide or thiazide-like diuretic monotherapy identified from a U.S. insurance claims database. Candidate covariates included patient demographics, and all observed prior conditions, drug exposures and procedures. Propensity scores were calculated using LASSO regression, and candidate covariates with non-zero beta coefficients in the propensity model were defined as matching covariates for use in LS-CM. One-to-one matching was performed using progressively tighter parameter settings. Covariate balance was assessed using standardized mean differences. Hazard ratios for negative control outcomes perceived as unassociated with treatment (i.e., true hazard ratio of 1) were estimated using unconditional Cox models. Residual confounding was assessed using the expected systematic error of the empirical null distribution of negative control effect estimates compared to the ground truth. To simulate diverse research conditions, analyses were repeated within 10 %, 1 and 0.5 % subsample groups with increasingly limited covariate overlap. Results A total of 172,117 patients (ACEI: 129,078; thiazide: 43,039) met the study criteria. As compared to LS-PSM, LS-CM was associated with increased sample retention. Although LS-PSM achieved balance across all matching covariates within the full study population, substantial matching covariate imbalance was observed within the 1 and 0.5 % subsample groups. Meanwhile, LS-CM achieved matching covariate balance across all analyses. LS-PSM was associated with better candidate covariate balance within the full study population. Otherwise, both matching techniques achieved comparable candidate covariate balance and expected systematic error. Conclusions LS-CM found the largest matched sample meeting prespecified balance criteria while achieving comparable candidate covariate balance and residual confounding. We recommend LS-CM as an alternative to LS-PSM in studies with small sample sizes or limited covariate overlap.

Using a Satisficing Model of Experimenter Decision-Making to Guide Finite-Sample Inference for Compromised Experiments

This paper presents a simple decision-theoretic economic approach for analyzing social experiments with compromised random assignment protocols that are only partially documented. We model administratively constrained experimenters who satisfice in seeking covariate balance. We develop design-based small-sample hypothesis tests that use worst-case (least favorable) randomization null distributions. Our approach accommodates a variety of compromised experiments, including imperfectly documented re-randomization designs. To make our analysis concrete, we focus much of our discussion on the influential Perry Preschool Project. We reexamine previous estimates of program effectiveness using our methods. The choice of how to model reassignment vitally affects inference.

Using the half normal distribution to quantify covariate balance in cluster-randomized pragmatic trials

Background Pragmatic trials often consist of cluster-randomized controlled trials (C-RCTs), where staff of existing clinics or sites deliver interventions and randomization occurs at the site level. Covariate-constrained randomization (CCR) methods are often recommended to minimize imbalance on important site characteristics across intervention and control arms because sizable imbalances can occur by chance in simple randomizations when the number of units to be randomized is relatively small. CCR methods involve multiple random assignments initially, an assessment of balance achieved on site-level covariates from each randomization, and the final selection of an allocation that produces acceptable balance. However, no clear consensus exists on how to assess imbalance or identify allocations with sufficient balance. In this article, we describe an overall imbalance index (I) that is based on the mean of the absolute value of the standardized differences in means on the site characteristics. Methods We derive the theoretical distribution of I, then conduct simulation studies to examine its empirical properties under the varying covariate distributions and inter-correlations. Results I has an expected value of 0.798 and, assuming independent site characteristics, a variance of 0.363/k, where k is the number of site characteristics being balanced. Simulations indicated that the properties of I are robust under varying covariate circumstances as long as k is greater than 3 and the covariates are not too highly inter-correlated. Conclusions We recommend that values of I below the 10th percentile indicate sufficient overall site balance in CCRs. Definitions of acceptable randomizations might also include individual covariate criteria specified in advance, in addition to overall balance criteria.

Conditional as-if analyses in randomized experiments

The injunction to “analyze the way you randomize” is well known to statisticians since Fisher advocated for randomization as the basis of inference. Yet even those convinced by the merits of randomization-based inference seldom follow this injunction to the letter. Bernoulli randomized experiments are often analyzed as completely randomized experiments, and completely randomized experiments are analyzed as if they had been stratified; more generally, it is not uncommon to analyze an experiment as if it had been randomized differently. This article examines the theoretical foundation behind this practice within a randomization-based framework. Specifically, we ask when is it legitimate to analyze an experiment randomized according to one design as if it had been randomized according to some other design. We show that a sufficient condition for this type of analysis to be valid is that the design used for analysis should be derived from the original design by an appropriate form of conditioning. We use our theory to justify certain existing methods, question others, and finally suggest new methodological insights such as conditioning on approximate covariate balance.

Propensity score matching in otolaryngologic literature: A systematic review and critical appraisal

Background Propensity score techniques can reduce confounding and bias in observational studies. Such analyses are able to measure and balance pre-determined covariates between treated and untreated groups, leading to results that can approximate those generated by randomized prospective studies when such trials are not feasible. The most commonly used propensity score -based analytic technique is propensity score matching (PSM). Although PSM popularity has continued to increase in medical literature, improper methodology or methodological reporting may lead to biased interpretation of treatment effects or limited scientific reproducibility and generalizability. In this study, we aim to characterize and assess the quality of PSM methodology reporting in high-impact otolaryngologic literature. Methods PubMed and Embase based systematic review of the top 20 journals in otolaryngology, as measured by impact factor from the Journal Citations Reports from 2012 to 2018, for articles using PSM analysis throughout their publication history. Eligible articles were reviewed and assessed for quality and reporting of PSM methodology. Results Our search yielded 101 studies, of which 92 were eligible for final analysis and review. The proportion of studies utilizing PSM increased significantly over time (p < 0.001). Nearly all studies (96.7%, n = 89) specified the covariates used to calculate propensity scores. Covariate balance was illustrated in 67.4% (n = 62) of studies, most frequently through p-values. A minority (17.4%, n = 16) of studies were found to be fully reproducible according to previously established criteria. Conclusions While PSM analysis is becoming increasingly prevalent in otolaryngologic literature, the quality of PSM methodology reporting can be improved. We provide potential recommendations for authors regarding optimal reporting for analyses using PSM.

Hemodynamic Response to Vasopressin Dosage of 0.03 Units/Min vs. 0.04 Units/Min in Patients With Septic Shock

Background: Arginine vasopressin (AVP) is suggested as an adjunct to norepinephrine in patients with septic shock. Guidelines recommend an AVP dosage up to 0.03 units/min, but 0.04 units/min is commonly used in practice based on initial studies. This study was designed to compare the incidence of hemodynamic response between initial fixed-dosage AVP 0.03 units/min and AVP 0.04 units/min. Methods: This retrospective, multi-hospital health system, cohort study included adult patients with septic shock receiving AVP as an adjunct to catecholamine vasopressors. Patients were excluded if they received an initial dosage other than 0.03 units/min or 0.04 units/min, or AVP was titrated within the first 6 hours of therapy. The primary outcome was hemodynamic response, defined as a mean arterial pressure ≥65 mm Hg and a decrease in catecholamine dosage at 6 hours after AVP initiation. Inverse probability of treatment weighting (IPTW) based on the propensity score for initial AVP dosage receipt was utilized to estimate adjusted exposure effects. Results: Of the 1536 patients included in the observed data, there was a nearly even split between initial AVP dosage of 0.03 units/min (n = 842 [54.8%]) and 0.04 units/min (n = 694 [45.2%]). Observed patients receiving AVP 0.03 units/min were more frequently treated at the main campus academic medical center (96.3% vs. 52.2%, p < 0.01) and in a medical intensive care unit (87.4% vs. 39.8%, p < 0.01). The IPTW analysis included 1379 patients with achievement of baseline covariate balance. There was no evidence for a difference between groups in the incidence of hemodynamic response (0.03 units/min 50.0% vs. 0.04 units/min 53.1%, adjusted relative risk 1.06 [95% CI 0.94, 1.20]). Conclusions: Initial AVP dosing varied by hospital and unit type. Although commonly used, an initial AVP dosage of 0.04 units/min was not associated with a higher incidence of early hemodynamic response to AVP in patients with septic shock.