Kathleen E. Wirth

To the Editor—We read with interest the recent study by Butt et al [ 1]. This observational study, conducted in the Department of Veterans Affairs healthcare system, estimated the relative effectiveness of a booster dose for preventing confirmed severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection (19%), hospitalization (52%), and intensive care unit admission or death (83%). The authors state that they emulated a target trial, a framework that structures the research question and mitigates common sources of bias [ 2–5]. Unfortunately, the authors did not define or adhere to the target trial they sought to emulate, leading to several study design and analysis concerns.

The Southern African Journal of HIV Medicine is focused on HIV/AIDS treatment, prevention and related topics relevant to clinical and public health practice. The purpose of the journal is to disseminate original research results and support high-level learning related to HIV Medicine. It publishes original research articles, editorials, case reports/case series, reviews of state-of-the-art clinical practice and correspondence.

Background: Circumcised men may increase sexual risk-taking following voluntary medical male circumcision (VMMC) because of decreased perceptions of risk, which may negate the beneficial impact of VMMC in preventing new human immunodeficiency virus (HIV) infections.Objectives: We evaluated changes in sexual behaviour following VMMC.Method: We conducted a prospective cohort study amongst sexually active, HIV-negative adult men undergoing VMMC in Gaborone, Botswana, during 2013–2015. Risky sexual behaviour, defined as the number of sexual partners in the previous month and ≥ 1 concurrent sexual partnerships during the previous 3 months, was assessed at baseline (prior to VMMC) and 3 months post-VMMC. Change over time was assessed by using inverse probability weighted linear and conditional logistic regression models.Results: We enrolled 523 men; 509 (97%) provided sexual behaviour information at baseline. At 3 months post-VMMC, 368 (72%) completed the follow-up questionnaire. At baseline, the mean (95% confidence interval) number of sexual partners was 1.60 (1.48, 1.65), and 111 (31% of 353 with data) men reported engaging in concurrent partnerships. At 3 months post-VMMC, 70 (23% of 311 with data) reported fewer partners and 19% had more partners. Amongst 111 men with a concurrent partnership at baseline, 52% reported none post-VMMC. Amongst the 242 (69%) without a concurrent partnership at baseline, 19% reported initiating one post-VMMC. After adjustment for loss to follow-up, risky sexual behaviour post-VMMC (measured as mean changes in a number of partners and proportion engaging in concurrency) was similar to baseline levels.Conclusion: We found no evidence of sexual risk compensation in the 3 months following VMMC.

The Southern African Journal of HIV Medicine is focused on HIV/AIDS treatment, prevention and related topics relevant to clinical and public health practice. The purpose of the journal is to disseminate original research results and support high-level learning related to HIV Medicine. It publishes original research articles, editorials, case reports/case series, reviews of state-of-the-art clinical practice and correspondence.

Instrumental variables are routinely used to recover a consistent estimator of an exposure causal effect in the presence of unmeasured confounding. Instrumental variable approaches to account for nonignorable missing data also exist but are less familiar to epidemiologists. Like instrumental variables for exposure causal effects, instrumental variables for missing data rely on exclusion restriction and instrumental variable relevance assumptions. Yet these two conditions alone are insufficient for point identification. For estimation, researchers have invoked a third assumption, typically involving fairly restrictive parametric constraints. Inferences can be sensitive to these parametric assumptions, which are typically not empirically testable. The purpose of our article is to discuss another approach for leveraging a valid instrumental variable. Although the approach is insufficient for nonparametric identification, it can nonetheless provide informative inferences about the presence, direction, and magnitude of selection bias, without invoking a third untestable parametric assumption. An important contribution of this article is an Excel spreadsheet tool that can be used to obtain empirical evidence of selection bias and calculate bounds and corresponding Bayesian 95% credible intervals for a nonidentifiable population proportion. For illustrative purposes, we used the spreadsheet tool to analyze HIV prevalence data collected by the 2007 Zambia Demographic and Health Survey (DHS).

Design: Universal voluntary HIV counselling and testing followed by prompt initiation of antiretroviral therapy (ART) for all those diagnosed HIV‐infected (universal test and treat, UTT) is now a global health standard. However, its population‐level impact, feasibility and cost remain unknown. Five community‐based trials have been implemented in sub‐Saharan Africa to measure the effects of various UTT strategies at population level: BCPP/YaTsie in Botswana, MaxART in Swaziland, HPTN 071 (PopART) in South Africa and Zambia, SEARCH in Uganda and Kenya and ANRS 12249 TasP in South Africa. This report describes and contrasts the contexts, research methodologies, intervention packages, themes explored, evolution of study designs and interventions related to each of these five UTT trials.Methods: We conducted a comparative assessment of the five trials using data extracted from study protocols and collected during baseline studies, with additional input from study investigators. We organized differences and commonalities across the trials in five categories: trial contexts, research designs, intervention packages, trial themes and adaptations.Results: All performed in the context of generalized HIV epidemics, the trials highly differ in their social, demographic, economic, political and health systems settings. They share the common aim of assessing the impact of UTT on the HIV epidemic but differ in methodological aspects such as study design and eligibility criteria for trial populations. In addition to universal ART initiation, the trials deliver a wide range of biomedical, behavioural and structural interventions as part of their UTT strategies. The five studies explore common issues, including the uptake rates of the trial services and individual health outcomes. All trials have adapted since their initiation to the evolving political, economic and public health contexts, including adopting the successive national recommendations for ART initiation.Conclusions: We found substantial commonalities but also differences between the five UTT trials in their design, conduct and multidisciplinary outputs. As empirical literature on how UTT may improve efficiency and quality of HIV care at population level is still scarce, this article provides a foundation for more collaborative research on UTT and supports evidence‐based decision making for HIV care in country and internationally.

Background: In 2007, the World Health Organization and the Joint United Nations Programme on HIV/AIDS endorsed voluntary medical male circumcision (VMMC) as an add-on HIV-prevention strategy. Similar to many other sub-Saharan countries, VMMC uptake in Botswana has been low; as of February 2016, only 42.7% of the program target had been achieved. Previous work has examined how individual-level factors, such as knowledge and attitudes, influence the update of VMMC. This paper examines how factors related to the health system can be leveraged to maximize uptake of circumcision services, with a focus on demand creation, access to services, and service delivery. Methods: Twenty-seven focus group discussions with 238 participants were conducted in four communities in Botswana among men (stratified by circumcision status and age), women (stratified by age), and community leaders. A semi-structured guide was used by a trained same-gender interviewer to facilitate discussions, which were audio recorded, transcribed, translated to English, and analyzed using an inductive analytic approach. Results: Participants felt demand creation activities utilizing age- and gender-appropriate mobilizers and community leaders were more effective than mass media campaigns. Participants felt improved access to VMMC clinics would facilitate service uptake, as would designated men’s clinics with male-friendly providers for VMMC service delivery. Additionally, providing comprehensive pre-procedure counseling and education, outlining the benefits and disadvantages of the surgical procedure, and explaining the differences between the surgical and non-surgical procedures, were suggested by participants to increase understanding and uptake of VMMC. Conclusion: Cultural acceptability of circumcision services can be improved by engaging age- and gender-appropriate community mobilizers. Involving influential community leaders, providing a forum for men to discuss health issues, and bringing services closer to people can increase VMMC utilization. Service delivery can be improved by communicating the pros and cons of the procedure to the clients for informed decision-making.

Unprotected sexual intercourse after undergoing voluntary medical male circumcision but prior to complete wound healing can lead to major adverse events including HIV acquisition. To better understand perceptions related to early resumption of sex prior to wound healing, 27 focus group discussions were conducted among 238 adult men, women, and community leaders in Botswana. Median age among all participants was 31 years of whom 60% were male and 51% were either employed and receiving salary or self-employed. Only 12% reported being currently married. Pain, not risk of HIV acquisition, was perceived as the main adverse consequence of early resumption of sex. In fact, no participant mentioned that early resumption of sex could lead to an increase in HIV risk. Demonstrating masculinity and virility, fear of losing female partners, and misperception about post-operative wound healing also played key roles in the decision to resume sex prior to complete wound healing. Findings from this study highlight a potentially widespread lack of awareness of the increased risk of HIV acquisition during the wound healing period. Strengthening post-operative counseling and identifying strategies to discourage the early resumption of sex will be increasingly important as older men and HIV-positive men seek voluntary medical male circumcision services.

Randomized trials have shown that voluntary medical male circumcision (VMMC) significantly reduces the risk of HIV acquisition in men. However, the rate of complications associated with the surgical procedure varies from 0.7% to 37.4% in real-world settings. We assessed the frequency, type and severity of adverse events following VMMC among 427 adult men surgically circumcised in southeastern Botswana; 97% completed ≥1 follow-up visit within seven days post-circumcision. Thirty moderate AEs were observed in 28 men resulting in an overall AE rate of 6.7%. Patient satisfaction was high: >95% were very or somewhat satisfied with the procedure and subsequent follow-up care.

Summary: The instrumental variable (IV) design is a well-known approach for unbiased evaluation of causal effects in the presence of unobserved confounding. In this article, we study the IV approach to account for selection bias in regression analysis with outcome missing not at random. In such a setting, a valid IV is a variable which (i) predicts the nonresponse process, and (ii) is independent of the outcome in the underlying population. We show that under the additional assumption (iii) that the IV is independent of the magnitude of selection bias due to nonresponse, the population regression in view is nonparametrically identified. For point estimation under (i)–(iii), we propose a simple complete-case analysis which modifies the regression of primary interest by carefully incorporating the IV to account for selection bias. The approach is developed for the identity, log and logit link functions. For inferences about the marginal mean of a binary outcome assuming (i) and (ii) only, we describe novel and approximately sharp bounds which unlike Robins–Manski bounds, are smooth in model parameters, therefore allowing for a straightforward approach to account for uncertainty due to sampling variability. These bounds provide a more honest account of uncertainty and allows one to assess the extent to which a violation of the key identifying condition (iii) might affect inferences. For illustration, the methods are used to account for selection bias induced by HIV testing nonparticipation in the evaluation of HIV prevalence in the Zambian Demographic and Health Surveys.