Adaptive pair-matching in randomized trials with unbiased and efficient effect estimation

Causal inference in randomized trials with partial clustering

BACKGROUND

Participant dependence, if present, must be accounted for in the analysis of randomized trials. This dependence, also referred to as "clustering," can occur in one or more trial arms. This dependence may predate randomization or arise after randomization. We examine three trial designs: one "fully clustered" (where all participants are dependent within clusters or groups) and two "partially clustered" (where some participants are dependent within clusters and some participants are completely independent of all others).

METHODS

For these three designs, we (1) use causal models to non-parametrically describe the data generating process and formalize the dependence in the observed data distribution; (2) develop a novel implementation of targeted minimum loss-based estimation for analysis; (3) evaluate the finite-sample performance of targeted minimum loss-based estimation and common alternatives via a simulation study; and (4) apply the methods to real-data from the SEARCH-IPT trial.

RESULTS

We show that the two randomization schemes resulting in partially clustered trials have the same dependence structure, enabling use of the same statistical methods for estimation and inference of causal effects. Our novel targeted minimum loss-based estimation approach leverages covariate adjustment and machine learning to improve precision and facilitates estimation of a large set of causal effects. In simulations, we demonstrate that targeted minimum loss-based estimation achieves comparable or markedly higher statistical power than common alternatives for these partially clustered designs. Finally, application of targeted minimum loss-based estimation to real data from the SEARCH-IPT trial resulted in 20%-57% efficiency gains, demonstrating the real-world consequences of our proposed approach.ConclusionsPartially clustered trial analysis can be made more efficient by implementing targeted minimum loss-based estimation, assuming care is taken to account for the dependent nature of the observed data.

Sleeping for Two: a randomized controlled trial of cognitive behavioral therapy for insomnia in pregnancy

STUDY OBJECTIVES

Insomnia and sleep problems are common in pregnancy and have potentially negative impacts on both parental and infant health. This study examined the Sleeping for Two adaptation of cognitive behavioral therapy for insomnia (CBT-I) in pregnancy.

METHODS

A parallel (1:1) randomized controlled trial evaluated CBT-I (n = 32) compared to a treatment as usual waitlist (n = 32) among pregnant individuals from Alberta, Canada experiencing insomnia. Five weekly individual sessions of CBT-I pivoted from in-person delivery to telehealth due to the COVID-19 pandemic physical distancing regulations. Insomnia symptom severity (primary outcome), insomnia diagnosis by structured interview, self-reported sleep problems, as well as sleep parameters measured by diary and actigraphy were assessed pretreatment at 12-28 weeks gestation, 1-week posttreatment, and 6 months postpartum. Birth information (secondary outcomes) were collected via delivery record and parent report of infant sleep (exploratory outcome) was taken at 6 months postpartum.

RESULTS

Multilevel modeling using an intention-to-treat approach showed that CBT-I was associated with a decrease in insomnia symptoms and improved sleep quality across time compared to treatment as usual. The CBT-I group had fewer diagnoses of insomnia posttreatment, but the difference did not reach statistical significance until 6 months postpartum. Participants with worse sleep quality at baseline benefitted substantially more from CBT-I vs treatment as usual waitlist.

CONCLUSIONS

CBT-I delivered in pregnancy can reduce symptoms of insomnia and improve sleep quality, which could in turn minimize risk of negative consequences for birthing parent and infant health.

CLINICAL TRIAL REGISTRATION

Registry: ClinicalTrials.gov; Name: Sleeping for Two: RCT of CBT-Insomnia in Pregnancy; URL: https://www.clinicaltrials.gov/study/NCT03301727; Identifier: NCT03918057.

CITATION

MacKinnon AL, Silang K, Watts D, et al. Sleeping for Two: a randomized controlled trial of cognitive behavioral therapy for insomnia in pregnancy. J Clin Sleep Med. 2025;21(2):365-376.

The impacts of task shifting on the management and treatment of malnourished children in Northern Kenya: a cluster-randomized controlled trial

Treating children with acute malnutrition can be challenging, particularly regarding access to healthcare facilities during treatment. Task shifting, a strategy of transferring specific tasks to health workers with shorter training and fewer qualifications, is being considered as an effective approach to enhancing health outcomes in primary healthcare. This study aimed to assess the effectiveness of integrating the treatment of acute malnutrition by community health volunteers into integrated community case management in two sub-counties in northern Kenya (Loima and Isiolo). We conducted a two-arm non-inferiority cluster-randomized controlled trial across 20 community health units. Participants were children aged 6-59 months with uncomplicated acute malnutrition. In the intervention group, community health volunteers used simplified tools and protocols to identify and treat eligible children at home and provided the usual integrated community case management package. In the control group, community health volunteers provided the usual integrated community case management package only (screening and referral of the malnourished children to the health facilities). The primary outcome was recovery (MUAC ≥12.5 cm for 2 consecutive weeks). Results show that children in the intervention group were more likely to recover than those in the control group [73 vs 50; risk difference (RD) = 26% (95% CI 12 to 40) and risk ratio (RR) = 2 (95% CI 1.2 to 1.9)]. The probability of defaulting was lower in the intervention group than in the control group: RD = -21% (95% CI -31 to -10) and RR = 0.3 (95% CI 0.2 to 0.5). The intervention reduced the length of stay by about 13 days, although this was not statistically significant and varied substantially by sub-county. Integrating the treatment of acute malnutrition by community health volunteers into the integrated community case management programme led to better malnutrition treatment outcomes. There is a need to integrate acute malnutrition treatment into integrated community case management and review policies to allow community health volunteers to treat uncomplicated acute malnutrition.

Supporting Peer Support Workers and Their Supervisors: Cluster-Randomized Trial Evaluating a Systems-Level Intervention

OBJECTIVE

Peer support workers are a substantial and growing part of the mental health workforce. Because little research has investigated how to effectively support and supervise peer support workers, the authors evaluated the efficacy of a training program to strengthen the peer support workforce and the supervision of its workers.

METHODS

Mental health services sites with peer support workers and supervisors in Los Angeles County were recruited for this cluster-randomized trial and 10-month follow-up. Of 348 peer support workers and 143 supervisors at 85 sites, 251 (72%) peer support workers and 115 (80%) supervisors completed baseline surveys. SHARE! the Self-Help And Recovery Exchange, a peer-run organization, delivered four training sessions on strategies to reduce stigma and to build an effective peer workforce, cultural competence, and a trauma-informed developmental model of supervision. Primary outcomes were peer-supportive organizational climate, mental health stigma, and peer support worker recovery.

RESULTS

Intention-to-treat analyses indicated that sites receiving the training had significantly higher scores on peer-supportive organizational climate (Cohen's d=0.35, 95% CI=0.02-0.68, p=0.04) relative to sites not receiving the training. No significant differences were found between the two conditions for mental health stigma (Cohen's d=0.04) or peer support worker recovery (Cohen's d=0.14).

CONCLUSIONS

The training had no impact on mental health stigma or peer support worker recovery. However, the findings suggest that the training increased the value organizations gave to peer support work, which may help improve peer support worker retention and outcomes among those served. Efforts to incorporate principles of the training into practice may strengthen outcomes.

Application of targeted maximum likelihood estimation in public health and epidemiological studies: a systematic review

PURPOSE

The targeted maximum likelihood estimation (TMLE) statistical data analysis framework integrates machine learning, statistical theory, and statistical inference to provide a least biased, efficient, and robust strategy for estimation and inference of a variety of statistical and causal parameters. We describe and evaluate the epidemiological applications that have benefited from recent methodological developments.

METHODS

We conducted a systematic literature review in PubMed for articles that applied any form of TMLE in observational studies. We summarized the epidemiological discipline, geographical location, expertize of the authors, and TMLE methods over time. We used the Roadmap of Targeted Learning and Causal Inference to extract key methodological aspects of the publications. We showcase the contributions to the literature of these TMLE results.

RESULTS

Of the 89 publications included, 33% originated from the University of California at Berkeley, where the framework was first developed by Professor Mark van der Laan. By 2022, 59% of the publications originated from outside the United States and explored up to seven different epidemiological disciplines in 2021-2022. Double-robustness, bias reduction, and model misspecification were the main motivations that drew researchers toward the TMLE framework. Through time, a wide variety of methodological, tutorial, and software-specific articles were cited, owing to the constant growth of methodological developments around TMLE.

CONCLUSIONS

There is a clear dissemination trend of the TMLE framework to various epidemiological disciplines and to increasing numbers of geographical areas. The availability of R packages, publication of tutorial papers, and involvement of methodological experts in applied publications have contributed to an exponential increase in the number of studies that understood the benefits and adoption of TMLE.

Statistical inference and power analysis for direct and spillover effects in two-stage randomized experiments

Two-stage randomized experiments become an increasingly popular experimental design for causal inference when the outcome of one unit may be affected by the treatment assignments of other units in the same cluster. In this paper, we provide a methodological framework for general tools of statistical inference and power analysis for two-stage randomized experiments. Under the randomization-based framework, we consider the estimation of a new direct effect of interest as well as the average direct and spillover effects studied in the literature. We provide unbiased estimators of these causal quantities and their conservative variance estimators in a general setting. Using these results, we then develop hypothesis testing procedures and derive sample size formulas. We theoretically compare the two-stage randomized design with the completely randomized and cluster randomized designs, which represent two limiting designs. Finally, we conduct simulation studies to evaluate the empirical performance of our sample size formulas. For empirical illustration, the proposed methodology is applied to the randomized evaluation of the Indian National Health Insurance Program. An open-source software package is available for implementing the proposed methodology.

Defining and estimating effects in cluster randomized trials: A methods comparison

Across research disciplines, cluster randomized trials (CRTs) are commonly implemented to evaluate interventions delivered to groups of participants, such as communities and clinics. Despite advances in the design and analysis of CRTs, several challenges remain. First, there are many possible ways to specify the causal effect of interest (eg, at the individual-level or at the cluster-level). Second, the theoretical and practical performance of common methods for CRT analysis remain poorly understood. Here, we present a general framework to formally define an array of causal effects in terms of summary measures of counterfactual outcomes. Next, we provide a comprehensive overview of CRT estimators, including the t-test, generalized estimating equations (GEE), augmented-GEE, and targeted maximum likelihood estimation (TMLE). Using finite sample simulations, we illustrate the practical performance of these estimators for different causal effects and when, as commonly occurs, there are limited numbers of clusters of different sizes. Finally, our application to data from the Preterm Birth Initiative (PTBi) study demonstrates the real-world impact of varying cluster sizes and targeting effects at the cluster-level or at the individual-level. Specifically, the relative effect of the PTBi intervention was 0.81 at the cluster-level, corresponding to a 19% reduction in outcome incidence, and was 0.66 at the individual-level, corresponding to a 34% reduction in outcome risk. Given its flexibility to estimate a variety of user-specified effects and ability to adaptively adjust for covariates for precision gains while maintaining Type-I error control, we conclude TMLE is a promising tool for CRT analysis.

Monetary Incentives and Early Initiation of Antenatal Care: A Matched-Pair, Parallel Cluster-Randomized Trial in Zambia

Monetary incentives are often used to increase the motivation and output of health service providers. However, the focus has generally been on frontline health service providers. Using a cluster randomized trial, we evaluate the effect of monetary incentives provided to community-based volunteers on early initiation of antenatal care (ANC) visits and deliveries in health facilities in communities in Zambia. Monetary incentives were assigned to community-based volunteers in treatment sites, and payments were made for every woman referred or accompanied in the first trimester of pregnancy during January-June 2020. We find a significant increase of about 32 percent in the number of women completing ANC visits in the first trimester but no effect on service coverage rates. The number of women accompanied by community-based volunteers for ANC in the first trimester increased by 33 percent. The number of deliveries in health facilities also increased by 22 percent. These findings suggest that the use of health facilities during the first trimester of pregnancy can be improved by providing community-based volunteers with monetary incentives and that such incentives can also increase deliveries in health facilities, which are key to improving the survival of women and newborns.

Two-Stage TMLE to reduce bias and improve efficiency in cluster randomized trials

Cluster randomized trials (CRTs) randomly assign an intervention to groups of individuals (e.g., clinics or communities) and measure outcomes on individuals in those groups. While offering many advantages, this experimental design introduces challenges that are only partially addressed by existing analytic approaches. First, outcomes are often missing for some individuals within clusters. Failing to appropriately adjust for differential outcome measurement can result in biased estimates and inference. Second, CRTs often randomize limited numbers of clusters, resulting in chance imbalances on baseline outcome predictors between arms. Failing to adaptively adjust for these imbalances and other predictive covariates can result in efficiency losses. To address these methodological gaps, we propose and evaluate a novel two-stage targeted minimum loss-based estimator to adjust for baseline covariates in a manner that optimizes precision, after controlling for baseline and postbaseline causes of missing outcomes. Finite sample simulations illustrate that our approach can nearly eliminate bias due to differential outcome measurement, while existing CRT estimators yield misleading results and inferences. Application to real data from the SEARCH community randomized trial demonstrates the gains in efficiency afforded through adaptive adjustment for baseline covariates, after controlling for missingness on individual-level outcomes.

Demystifying Statistical Inference When Using Machine Learning in Causal Research

In this issue, Naimi et al. (Am J Epidemiol. XXXX;XXX(XX):XXXX-XXXX) discuss a critical topic in public health and beyond: obtaining valid statistical inference when using machine learning in causal research. In doing so, the authors review recent prominent methodological work and recommend: (i) double robust estimators, such as targeted maximum likelihood estimation (TMLE); (ii) ensemble methods, such as Super Learner, to combine predictions from a diverse library of algorithms, and (iii) sample-splitting to reduce bias and improve inference. We largely agree with these recommendations. In this commentary, we highlight the critical importance of the Super Learner library. Specifically, in both simulation settings considered by the authors, we demonstrate that low bias and valid statistical inference can be achieved using TMLE without sample-splitting and with a Super Learner library that excludes tree-based methods but includes regression splines. Whether extremely data-adaptive algorithms and sample-splitting are needed depends on the specific problem and should be informed by simulations reflecting the specific application. More research is needed on practical recommendations for selecting among these options in common situations arising in epidemiology.

The Long-Term Dietitian and Psychological Support of Obese Patients Who Have Reduced Their Weight Allows Them to Maintain the Effects

The role of post-therapeutic support after weight loss in obesity treatment is not fully understood. Therefore, weight maintenance after a successful weight loss intervention is not very common, especially in obese individuals. This randomized controlled study was conducted to explore the efficacy of following dietary and psychological support in a group of 36 obese individuals. Participants (22 women, 14 men aged 35.58 ± 9.85 years, BMI 35.04 ± 3.80 kg/m²) who completed a 12-month weight loss phase (balanced energy-restricted diet) were randomly allocated to receive 18-month support (SG) or no additional care (CG). The support phase included some elements of Ten Top Tips (TTT), cognitive behavioral therapy (CBT), motivational interviewing (MI) in combination with nutritional education and assessment of the level of physical activity. The primary outcome was the maintenance of anthropometric parameters at an 18-month follow-up. The secondary outcomes included evaluation of biochemical parameters and single nucleotide polymorphisms (SNPs) in genes connected with obesity. A comparison of SG vs. CG after a 30-month period of the study revealed significant differences in weight changes (−3.83 ± 6.09 vs. 2.48 ± 6.24 kg), Body Mass Index (−1.27 ± 2.02 vs. 0.72 ± 2.12 kg/m²), visceral adipose tissue (−0.58 ± 0.63 vs. 0.45 ± 0.74 L), and waist circumference (−4.83 ± 4.05 vs. 1.83 ± 5.97 cm). Analysis of SNPs (rs9939609 FTO, rs987237 TFAP2B, and rs894160 PLIN1) provided further insight into the potential modulating effect of certain genotypes on weight loss and maintenance and extended the knowledge of the potential benefits of personalized medicine. Post-therapeutical support in current clinical practice may increase the chances of long-term weight loss maintenance in obesity treatment even in patients with a genetic predisposition to excessive weight.

Computer-Based 3D Simulation Method in Dental Occlusion Education: Student Response and Learning Effect

Occlusion is a fundamental subject in dental education, and occlusal adjustment is clinically essential in daily dental practices. This study aimed to assess the effects of computer-based 3D simulations on learner responses and learning effect on the principles of occlusal adjustment in undergraduate dental students in comparison with the traditional approach. Two teaching methods, i.e., paper-based 2D presentation and computer-based 3D simulation, were used for teaching the occlusal adjustment concepts. Sixty dental students were divided into two groups using a pair-matching randomization method. In the 2D presentation group, a textbook with 2D illustrations was used. 3D graphic dental models and computer design software were applied in the 3D simulation group. After the course, an attitudinal survey and examination were conducted to evaluate the participants’ feedback and the learning effects resulting from the teaching methods. The independent t test was used to compare the test scores between groups (with α = 0.5). Pearson’s correlation coefficient was calculated to investigate the agreement between the survey data and test scores. Most of the students’ feedback indicated that the 3D simulation method would be effective in acquiring knowledge on occlusion and jaw movement. The examination scores were significantly higher in the 3D simulation group compared with those in the 2D presentation group in the questions for centric relation (P = 0.034). Conversely, the scores were insignificant in the questions for eccentric relation (P = 0.403). There was no correlation observed between the survey data and the actual examination score. Computer-based 3D simulation could increase the participants’ expectations and learning effects in dental occlusion education. Further studies in diversified learning environments are required on the efficacy of digital educational modality.

Futility of Cluster Designs at Individual Hospitals to Study Surgical Site Infections and Interventions Involving the Installation of Capital Equipment in Operating Rooms

Anesthesia workspaces are integral components in the chains of many intraoperative bacterial transmission events resulting in surgical site infections (SSI). Matched cohort designs can be used to compare SSI rates among operating rooms (ORs) with or without capital equipment purchases (e.g., new anesthesia machines). Patients receiving care in intervention ORs (i.e., with installed capital equipment) are matched with similar patients receiving care in ORs lacking the intervention. We evaluate statistical power of an alternative design for clinical trials in which, instead, SSI incidences are compared directly among ORs (i.e., the ORs form the clusters) at single hospitals (e.g., the 5 ORs with bactericidal lights vs. the 5 other ORs). Data used for parameter estimates were SSI for 24 categories of procedures among 338 hospitals in the State of California, 2015. Estimated statistical power was ≅8.4% for detecting a reduction in the incidence of SSI from 3.6% to 2.4% over 1 year with 5 intervention ORs and 5 control ORs. For ≅80% statistical power, >20 such hospitals would be needed to complete a study in 1 year. Matched paired cluster designs pair similar ORs (e.g., 2 cardiac ORs, 1 to intervention and 1 to control). With 5 pairs, statistical power would be even less than the estimated 8.4%. Cluster designs (i.e., analyses by OR) are not suitable for comparing SSI among ORs at single hospitals. Even though matched cohort designs are non-randomized and thus have lesser validity, matching patients by their risk factors for SSI is more practical.

The PRogram In Support of Moms (PRISM): study protocol for a cluster randomized controlled trial of two active interventions addressing perinatal depression in obstetric settings

BACKGROUND

Perinatal depression, the most common pregnancy complication, is associated with negative maternal-offspring outcomes. Despite existence of effective treatments, it is under-recognized and under-treated. Professional organizations recommend universal screening, yet multi-level barriers exist to ensuring effective diagnosis, treatment, and follow-up. Integrating mental health and obstetric care holds significant promise for addressing perinatal depression. The overall study goal is to compare the effectiveness of two active interventions: (1) the Massachusetts Child Psychiatry Access Program (MCPAP) for Moms, a state-wide, population-based program, and (2) the PRogram In Support of Moms (PRISM) which includes MCPAP for Moms plus a proactive, multifaceted, practice-level intervention with intensive implementation support.

METHODS

This study is conducted in two phases: (1) a run-in phase which has been completed and involved practice and patient participant recruitment to demonstrate feasibility for the second phase, and (2) a cluster randomized controlled trial (RCT), which is ongoing, and will compare two active interventions 1:1 with ten Ob/Gyn practices as the unit of randomization. In phase 1, rates of depressive symptoms and other demographic and clinical features among patients were examined to inform practice randomization. Patient participants to be recruited in phase 2 will be followed longitudinally until 13 months postpartum; they will have 3-5 total study visits depending on whether their initial recruitment and interview was at 4-24 or 32-40 weeks gestation, or 1-3 months postpartum. Sampling throughout pregnancy and postpartum will ensure participants with different depressive symptom onset times. Differences in depression symptomatology and treatment participation will be compared between patient participants by intervention arm.

DISCUSSION

This manuscript describes the full two-phase study protocol. The study design is innovative because it combines effectiveness with implementation research designs and integrates critical components of participatory action research. Our approach assesses the feasibility, acceptance, efficacy, and sustainability of integrating a stepped-care approach to perinatal depression care into ambulatory obstetric settings; an approach that is flexible and can be tailored and adapted to fit unique workflows of real-world practices.

TRIAL REGISTRATION

ClinicalTrials.gov Identifier: NCT02760004, registered prospectively on May 3, 2016.

Hypertension control in integrated HIV and chronic disease clinics in Uganda in the SEARCH study

BACKGROUND

There is an increasing burden of hypertension (HTN) across sub-Saharan Africa where HIV prevalence is the highest in the world, but current care models are inadequate to address the dual epidemics. HIV treatment infrastructure could be leveraged for the care of other chronic diseases, including HTN. However, little data exist on the effectiveness of integrated HIV and chronic disease care delivery systems on blood pressure control over time.

METHODS

Population screening for HIV and HTN, among other diseases, was conducted in ten communities in rural Uganda as part of the SEARCH study (NCT01864603). Individuals with either HIV, HTN, or both were referred to an integrated chronic disease clinic. Based on Uganda treatment guidelines, follow-up visits were scheduled every 4 weeks when blood pressure was uncontrolled, and either every 3 months, or in the case of drug stock-outs more frequently, when blood pressure was controlled. We describe demographic and clinical variables among all patients and used multilevel mixed-effects logistic regression to evaluate predictors of HTN control.

RESULTS

Following population screening (2013-2014) of 34,704 adults age ≥ 18 years, 4554 individuals with HTN alone or both HIV and HTN were referred to an integrated chronic disease clinic. Within 1 year 2038 participants with HTN linked to care and contributed 15,653 follow-up visits over 3 years. HTN was controlled at 15% of baseline visits and at 46% (95% CI: 44-48%) of post-baseline follow-up visits. Scheduled visit interval more frequent than clinical indication among patients with controlled HTN was associated with lower HTN control at the subsequent visit (aOR = 0.89; 95% CI 0.79-0.99). Hypertension control at follow-up visits was higher among HIV-infected patients than uninfected patients to have controlled blood pressure at follow-up visits (48% vs 46%; aOR 1.28; 95% CI 0.95-1.71).

CONCLUSIONS

Improved HTN control was achieved in an integrated HIV and chronic care model. Similar to HIV care, visit frequency determined by drug supply chain rather than clinical indication is associated with worse HTN control.

TRIAL REGISTRATION

The SEARCH Trial was prospectively registered with ClinicalTrials.gov : NCT01864603.

Qualitative research on community experiences in large HIV research trials: what have we learned?

INTRODUCTION

Very few pragmatic and community-level effectiveness trials integrate the use of qualitative research over all stages of the trial, to inform trial design, implementation optimization, results interpretation and post-trial policy recommendations. This is despite the growing demand for mixed methods research from funding agencies and awareness of the vital importance of qualitative and mixed methods research for understanding trial successes and challenges.

DISCUSSION

We offer examples from work we have been involved in to illustrate how qualitative research conducted within trials can reveal vital contextual factors that influence implementation and outcomes, can enable an informed adaptation of trials as they are being conducted and can lead to the formulation of theory regarding the social and behavioural pathways of intervention, while also enabling community engagement in trial design and implementation. These examples are based on published findings from qualitative studies embedded within two ongoing large-scale studies demonstrating the population-level impacts of universal HIV testing and treatment strategies in southern and eastern Africa, and a qualitative study conducted alongside a clinical trial testing the adaptation, acceptability and experience of short-cycle therapy in children and adolescents living with HIV.

CONCLUSIONS

We advocate for the integration of qualitative with clinical and survey research methods in pragmatic clinical and community-level trials and implementation studies, and for increasing visibility of qualitative and mixed methods research in medical journals. Qualitative research from trials ideally should be published along with clinical outcome data, either integrated into the "main" trial papers or published concurrently in the same journal issue. Integration of qualitative research within trials can help not only to understand the why behind success or failure of interventions in different contexts, but also inform the adaptation of interventions that can facilitate their success, and lead to new alternative strategies and to policy changes that may be vital for achieving public health goals, including the end of AIDS.

Expectations about future health and longevity in Kenyan and Ugandan communities receiving a universal test-and-treat intervention in the SEARCH trial

Expectations about future health and longevity are important determinants of individuals’ decisions to invest in physical and human capital. Few population-level studies have measured subjective expectations and examined how they are affected by scale-up of antiretroviral therapy (ART). We assessed these expectations in communities receiving annual HIV testing and universal ART. Longitudinal data on expectations were collected at baseline and one year later in 16 intervention communities participating in the Sustainable East Africa Research in Community Health (SEARCH) trial of the test and treat strategy in Kenya and Uganda (NCT01864603). A random sample of households with and without an HIV-positive adult was selected after baseline HIV testing. Individuals’ expectations about survival to 50, 60, 70, and 80 years of age, as well as future health status and economic well-being, were measured using a Likert scale. Primary outcomes were binary variables indicating participants who reported being very likely or almost certain to survive to advanced ages. Logistic regression analyses were used to examine trends in expectations as well as associations with HIV status and viral load for HIV-positive individuals. Data were obtained from 3126 adults at baseline and 3977 adults in year 1, with 2926 adults participating in both waves. HIV-negative adults were more likely to have favorable expectations about survival to 60 years than HIV-positive adults with detectable viral load (adjusted odds ratio [AOR] 1.87, 95% CI 1.53–2.30), as were HIV-positive adults with undetectable viral load (AOR 1.41, 95% CI 1.13–1.77). Favorable expectations about survival to 60 years were more likely for all groups in year 1 compared to baseline (AOR 1.53, 95% CI 1.31–1.77). These findings are consistent with the hypothesis that universal ART leads to improved population-level expectations about future health and well-being. Future research from the SEARCH trial will help determine whether these changes are causally driven by the provision of universal ART.

Adaptive pre-specification in randomized trials with and without pair-matching

In randomized trials, adjustment for measured covariates during the analysis can reduce variance and increase power. To avoid misleading inference, the analysis plan must be pre-specified. However, it is often unclear a priori which baseline covariates (if any) should be adjusted for in the analysis. Consider, for example, the Sustainable East Africa Research in Community Health (SEARCH) trial for HIV prevention and treatment. There are 16 matched pairs of communities and many potential adjustment variables, including region, HIV prevalence, male circumcision coverage, and measures of community-level viral load. In this paper, we propose a rigorous procedure to data-adaptively select the adjustment set, which maximizes the efficiency of the analysis. Specifically, we use cross-validation to select from a pre-specified library the candidate targeted maximum likelihood estimator (TMLE) that minimizes the estimated variance. For further gains in precision, we also propose a collaborative procedure for estimating the known exposure mechanism. Our small sample simulations demonstrate the promise of the methodology to maximize study power, while maintaining nominal confidence interval coverage. We show how our procedure can be tailored to the scientific question (intervention effect for the study sample vs. for the target population) and study design (pair-matched or not). Copyright © 2016 John Wiley & Sons, Ltd.

Targeted estimation and inference for the sample average treatment effect in trials with and without pair-matching

In cluster randomized trials, the study units usually are not a simple random sample from some clearly defined target population. Instead, the target population tends to be hypothetical or ill-defined, and the selection of study units tends to be systematic, driven by logistical and practical considerations. As a result, the population average treatment effect (PATE) may be neither well defined nor easily interpretable. In contrast, the sample average treatment effect (SATE) is the mean difference in the counterfactual outcomes for the study units. The sample parameter is easily interpretable and arguably the most relevant when the study units are not sampled from some specific super-population of interest. Furthermore, in most settings, the sample parameter will be estimated more efficiently than the population parameter. To the best of our knowledge, this is the first paper to propose using targeted maximum likelihood estimation (TMLE) for estimation and inference of the sample effect in trials with and without pair-matching. We study the asymptotic and finite sample properties of the TMLE for the sample effect and provide a conservative variance estimator. Finite sample simulations illustrate the potential gains in precision and power from selecting the sample effect as the target of inference. This work is motivated by the Sustainable East Africa Research in Community Health (SEARCH) study, a pair-matched, community randomized trial to estimate the effect of population-based HIV testing and streamlined ART on the 5-year cumulative HIV incidence (NCT01864603). The proposed methodology will be used in the primary analysis for the SEARCH trial. Copyright © 2016 John Wiley & Sons, Ltd.

A hybrid mobile approach for population-wide HIV testing in rural east Africa: an observational study

Background

Despite large investments in HIV testing, only 45% of HIV-infected persons in sub-Saharan Africa are estimated to know their status. Optimal methods for maximizing population-level testing remain unknown. We sought to demonstrate the effectiveness at achieving population-wide testing coverage of a hybrid mobile HIV testing approach.

Methods

From 2013–2014, we enumerated 168,772 adult (≥15 years) residents of 32 communities in Uganda (N=20), and Kenya (N=12) using a door-to-door census. “Stable” residence was defined as living in community for ≥6 months over the past year. In each community we performed 2-week multi-disease community health campaigns (CHC) that included HIV testing, counseling, and referral to care if HIV-infected; CHC non-participants were approached for home-based testing (HBT) over 1–2 months. We determined population HIV testing coverage, and predictors of testing via HBT (vs. CHC) and non-testing.

Findings

HIV testing was achieved in 89% of stable adult residents (131,307/146,906). HIV prevalence was 9.6% (13,043/136,033 stable and non-stable adults); median CD4⁺ T-cell count was 514 cells/μL (IQR: 355–703). Among stable adults tested, 43% (56,106/131,307) reported no prior testing. Among HIV-infected adults, 38% (4,932/13,043) were unaware of their status. Among stable CHC attendees, 99.5% (104,635/105,170) accepted HIV testing. Of stable adults tested, 80% (104,635/131,307, range: 60–93%) tested via CHCs. In multivariable analyses of stable adults, predictors of non-testing included male gender (risk ratio [RR]: 1.52, 95% CI: 1.48–1.56), single marital status (RR: 1.70, 95% CI: 1.66–1.75), Kenyan residence (RR: 1.46, 95% CI: 1.41–1.50, vs. Ugandan), and out-of-community migration for ≥1 month in past year (RR: 1.60, 95% CI: 1.53–1.68). Testing was more common among farmers (RR: 0.73, 95% CI: 0.67–0.79) and adults with primary education (RR: 0.84, 95% CI: 0.80–0.89).

Interpretation

High HIV testing coverage was achieved in rural Ugandan and Kenyan communities using a hybrid, mobile approach of multi-disease CHCs followed by HBT. This approach allowed for flexibility at the community and individual level in reaching testing coverage goals. Men and mobile populations remain challenges for universal testing.