Objectives: This study tested the impacts of peer specialists on housing stability, substance abuse, and mental health status for previously homeless Veterans with cooccurring mental health issues and substance abuse. Methods: Veterans living in the US Housing and Urban Development—Veterans Administration Supported Housing (HUD-VASH) program were randomized to peer specialist services that worked independently from HUD-VASH case managers (ie, not part of a case manager/peer specialist dyad) and to treatment as usual that included case management services. Peer specialist services were community-based, using a structured curriculum for recovery with up to 40 weekly sessions. Standardized self-report measures were collected at 3 timepoints. The intent-to-treat analysis tested treatment effects using a generalized additive mixed-effects model that allows for different nonlinear relationships between outcomes and time for treatment and control groups. A secondary analysis was conducted for Veterans who received services from peer specialists that were adherent to the intervention protocol. Results: Treated Veterans did not spend more days in housing compared with control Veterans during any part of the study at the 95% level of confidence. Veterans assigned to protocol adherent peer specialists showed greater housing stability between about 400 and 800 days postbaseline. Neither analysis detected significant effects for the behavioral health measures. Conclusions: Some impact of peer specialist services was found for housing stability but not for behavioral health problems. Future studies may need more sensitive measures for early steps in recovery and may need longer time frames to effectively impact this highly challenged population.

Background: Several metrics of glucose variability have been proposed to date, but an integrated approach that provides a complete and consistent assessment of glycemic variation is missing. As a consequence, and because of the tedious coding necessary during quantification, most investigators and clinicians have not yet adopted the use of multiple glucose variability metrics to evaluate glycemic variation. Methods: We compiled the most extensively used statistical techniques and glucose variability metrics, with adjustable hyper-and hypoglycemic limits and metric parameters, to create a user-friendly Continuous Glucose

Randomized trials are often conducted with separate randomizations across multiple sites such as schools, voting districts, or hospitals. These sites can differ in important ways, including the site's implementation, local conditions, and the composition of individuals. An important question in practice is whether-and under what assumptionsresearchers can leverage this cross-site variation to learn more about the intervention. We address these questions in the principal stratification framework, which describes causal effects for subgroups defined by post-treatment quantities. We show that researchers can estimate certain principal causal effects via the multi-site design if they are willing to impose the strong assumption that the site-specific effects are uncorrelated with the site-specific distribution of stratum membership. We motivate this approach with a multi-site trial of the Early College High School Initiative, a unique secondary education program with the goal of increasing high school graduation rates and college enrollment. Our analyses corroborate previous studies suggesting that the initiative had positive effects for students who would have otherwise attended a low-quality high school, although power is limited. This eliminates strata (A) and (B). To eliminate strata (C) and (D) we need an additional assumption: Assumption 3.2 (No Flip-Floppers). There are no individuals with {D i (1) = lq, D i (0) = hq} or {D i (1) = hq, D i (0) = lq}.This assumption states that individuals do not switch the type of non-

No abstract

In online experimentation, trigger-dilute analysis is an approach to obtain more precise estimates of intent-to-treat (ITT) effects when the intervention is only exposed, or "triggered", for a small subset of the population. Trigger-dilute analysis cannot be used for estimation when triggering is only partially observed. In this paper, we propose an unbiased ITT estimator with reduced variance for cases where triggering status is only observed in the treatment group. Our method is based on the efficiency augmentation idea of CUPED and draws upon identification frameworks from the principal stratification and instrumental variables literature. The unbiasedness of our estimation approach relies on a testable assumption that an augmentation term used for covariate adjustment equals zero in expectation. When this augmentation term fails a mean-zero test, we show how our estimator can incorporate in-experiment observations to reduce the augmentation's bias, by sacrificing the amount of variance reduced. This provides an explicit knob to trade off bias with variance. We demonstrate through simulations that our estimator can remain unbiased and achieve precision improvements as good as if triggering status were fully observed, and in some cases outperforms trigger-dilute analysis.