Instrumental variable estimation of the marginal structural Cox model for time-varying treatments

Summary Robins (1998) introduced marginal structural models, a general class of counterfactual models for the joint effects of time-varying treatments in complex longitudinal studies subject to time-varying confounding. Robins (1998) established the identification of marginal structural model parameters under a sequential randomization assumption, which rules out unmeasured confounding of treatment assignment over time. The marginal structural Cox model is one of the most popular marginal structural models to evaluate the causal effect of time-varying treatments on a censored failure time outcome. In this paper, we establish sufficient conditions for identification of marginal structural Cox model parameters with the aid of a time-varying instrumental variable, when sequential randomization fails to hold due to unmeasured confounding. Our instrumental variable identification condition rules out any interaction between an unmeasured confounder and the instrumental variable in its additive effects on the treatment process, the longitudinal generalization of the identifying condition of Wang & Tchetgen Tchetgen (2018). We describe a large class of weighted estimating equations that give rise to consistent and asymptotically normal estimators of the marginal structural Cox model, thereby extending the standard inverse probability of treatment weighted estimation of marginal structural models to the instrumental variable setting. Our approach is illustrated via extensive simulation studies and an application to estimate the effect of community antiretroviral therapy coverage on HIV incidence.

Enrollment and transition challenges in the International Maternal Pediatric and Adolescent AIDS Clinical Trials (IMPAACT) network’s PROMISE trial for resource-limited regions

Background: We describe enrollment and accrual challenges in the “Promoting Maternal and Infant Survival Everywhere” (PROMISE) trial conducted in resource-limited countries, as well as the challenges in transitioning participants from the antepartum to the postpartum components of the study. Methods: PROMISE was a large multi-national randomized controlled trial of the safety and efficacy of interventions to reduce perinatal transmission of HIV-1 (HIV) during pregnancy and breastfeeding and of interventions to preserve maternal health after cessation of perinatal transmission risk. The PROMISE study included two protocols for HIV-infected pregnant women in resource-limited countries who intended to either breastfeed or formula-feed their infants and did not meet country criteria for antiretroviral treatment. The PROMISE breastfeeding protocol (1077BF) used a sequential randomization design with up to three randomizations (Antepartum, Postpartum, and Maternal Health). The PROMISE formula-feeding protocol (1077FF) had two randomizations (Antepartum and Maternal Health). Women presenting to the clinic during early or active labor or in the immediate postpartum period were registered as Late Presenters and screened to determine whether eligible to participate in the Postpartum randomization. Results: The study was conducted at 14 sites in seven countries and opened to enrollment in April 2011. A total of 3259 pregnant women intending to breastfeed and an additional 284 pregnant women intending to formula feed were randomized in the Antepartum component. A total of 204 Late Presenters were registered during labor or after delivery. Enrollment was high among breastfeeding women (representing 96% of the target of 3400 women) but was lower than expected among women intending to formula feed (28% of 1000 expected) and late-presenting women (8% of 2500 expected). The successful overall enrollment and final primary study analyses results were attributed to substantial preparation before the study opened, collaboration among all stakeholders, close study monitoring during implementation and the flexibility to change and streamline the protocol. Conclusions: Experiences from the PROMISE study illustrate the challenges of enrolling in longer term studies in the setting of rapidly evolving prevention and treatment standards priorities. The lessons learned will help the community, site investigators, and study coordinators in the design and implementation of future clinical trials.