Faculty Bibliography

OBJECTIVE:Observational data can be used to attempt to emulate a target trial of statin use and estimate analogues of intention-to-treat and per protocol effects on dementia risk. METHODS:Using data from a prospective cohort study in the Netherlands, we conceptualized a sequence of "trials" in which eligible individuals ages 55-80 years were classified as statin initiators or noninitiators for every consecutive month between 1993 and 2007 and were followed until diagnosis of dementia, death, loss to follow-up, or the end of follow-up. We estimated 2 types of effects of statin use on dementia and a combined endpoint of dementia or death: the effect of initiation vs no initiation and the effect of sustained use vs no use. We estimated risk by statin treatment strategy over time via pooled logistic regression. We used inverse-probability weighting to account for treatment-confounder feedback in estimation of per-protocol effects. RESULTS:Of 233,526 eligible person-trials (6,373 individuals), there were 622 initiators and 232,904 noninitiators. Comparing statin initiation with no initiation, the 10-year risk differences (95% confidence interval) were -0.1% (-2.3% to 1.8%) for dementia and 0.3% (-2.7% to 3.3%) for dementia or death. Comparing sustained statin use vs no use, the 10-year risk differences were -2.2% (-5.2% to 1.6%) for dementia and -5.1% (-10.5% to -1.1%) for dementia or death. CONCLUSIONS:Individuals with sustained statin use, but not statin initiation alone, had reduced 10-year risks of dementia and dementia or death. Our results should be interpreted with caution due to the small number of initiators and events and potential for residual confounding.

BACKGROUND:We aimed to investigate the impact of reducing drinking in patients with unhealthy alcohol use on improvement of chronic pain interference, substance use, and psychiatric symptoms. METHODS:We analyzed longitudinal data from 2003 to 2015 in the Veterans Aging Cohort Study, a prospective, multisite observational study of US veterans, by emulating a hypothetical randomized trial (a target trial). Alcohol use was assessed using the AUDIT questionnaire, and outcome conditions were assessed via validated survey items. Individuals were followed from the first time their AUDIT score was ≥ 8 (baseline), a threshold consistent with unhealthy alcohol use. We compared individuals who reduced drinking (AUDIT < 8) at the next follow-up visit with individuals who did not (AUDIT ≥ 8). We fit separate logistic regression models to estimate odds ratios for improvement of each condition 2 years postbaseline among individuals who had that condition at baseline: moderate or severe pain interference symptoms, tobacco smoking, cannabis use, cocaine use, depressive symptoms, and anxiety symptoms. Inverse probability weighting was used to account for potential selection bias and confounding. RESULTS:Adjusted 2-year odds ratios (95% confidence intervals) for associations between reducing drinking and improvement or resolution of each condition were as follows: 1.49 (0.91, 2.42) for pain interference symptoms, 1.57 (0.93, 2.63) for tobacco smoking, 1.65 (0.92, 2.95) for cannabis use, 1.83 (1.03, 3.27) for cocaine use, 1.11 (0.64, 1.92) for depressive symptoms, and 1.33 (0.80, 2.22) for anxiety symptoms. CONCLUSIONS:We found some evidence for improvement of pain interference symptoms and substance use after reducing drinking among US veterans with unhealthy alcohol use, but confidence intervals were wide.

When reporting results from randomized experiments, researchers often choose to present a per-protocol effect in addition to an intention-to-treat effect. However, these per-protocol effects are often described retrospectively, for example, comparing outcomes among individuals who adhered to their assigned treatment strategy throughout the study. This retrospective definition of a per-protocol effect is often confounded and cannot be interpreted causally because it encounters treatment-confounder feedback loops, where past confounders affect future treatment, and current treatment affects future confounders. Per-protocol effects estimated using this method are highly susceptible to the placebo paradox, also called the "healthy adherers" bias, where individuals who adhere to placebo appear to have better survival than those who don’t. This result is generally not due to a benefit of placebo, but rather is most often the result of uncontrolled confounding. Here, we aim to provide an overview to causal inference for survival outcomes with time-varying exposures for static interventions using inverse probability weighting. The basic concepts described here can also apply to other types of exposure strategies, although these may require additional design or analytic considerations. We provide a workshop guide with solutions manual, fully reproducible R, SAS, and Stata code, and a simulated dataset on a GitHub repository for the reader to explore.

Distance to care is a common exposure and proposed instrumental variable in health research, but it is vulnerable to violations of fundamental identifiability conditions for causal inference. We used data collected from the Botswana Birth Outcomes Surveillance study between 2014 and 2016 to outline 4 challenges and potential biases when using distance to care as an exposure and as a proposed instrument: selection bias, unmeasured confounding, lack of sufficiently well-defined interventions, and measurement error. We describe how these issues can arise, and we propose sensitivity analyses for estimating the degree of bias.

Decisions about when to start or switch a therapy often depend on the frequency with which individuals are monitored or tested. For example, the optimal time to switch antiretroviral therapy depends on the frequency with which HIV-positive individuals have HIV RNA measured. This paper describes an approach to use observational data for the comparison of joint monitoring and treatment strategies and applies the method to a clinically relevant question in HIV research: when can monitoring frequency be decreased and when should individuals switch from a first-line treatment regimen to a new regimen? We outline the target trial that would compare the dynamic strategies of interest and then describe how to emulate it using data from HIV-positive individuals included in the HIV-CAUSAL Collaboration and the Centers for AIDS Research Network of Integrated Clinical Systems. When, as in our example, few individuals follow the dynamic strategies of interest over long periods of follow-up, we describe how to leverage an additional assumption: no direct effect of monitoring on the outcome of interest. We compare our results with and without the "no direct effect" assumption. We found little differences on survival and AIDS-free survival between strategies where monitoring frequency was decreased at a CD4 threshold of 350 cells/μl compared with 500 cells/μl and where treatment was switched at an HIV-RNA threshold of 1000 copies/ml compared with 200 copies/ml. The "no direct effect" assumption resulted in efficiency improvements for the risk difference estimates ranging from an 7- to 53-fold increase in the effective sample size.

OBJECTIVE:To compare the effect of preconception initiation of zidovudine, lamivudine, nevirapine (ZDV/3TC/NVP) versus tenofovir, emtricitabine, efavirenz (TDF/FTC/EFV) on adverse birth outcomes. DESIGN/METHODS:Emulation of a hypothetical (target) trial using a birth surveillance study in Botswana during an era of CD4-based antiretroviral therapy (ART) initiation. METHODS:In women who initiated ART less than 3 years from HIV diagnosis, conceived 0.5-5 years after ART initiation, and delivered at least 24-week gestation, we estimated risk ratios for stillbirth, preterm delivery (<37 weeks), very preterm delivery (<32 weeks), small-for-gestational-age (SGA) (<10 percentile), very SGA (<3 percentile), and any adverse or severe birth outcome for first-line ZDV/3TC/NVP versus TDF/FTC/EFV. We conducted a historical comparison in women who initiated TDF/FTC/EFV in 2012-2015 and ZDV/3TC/NVP in 2004-2011, and a contemporaneous comparison in an era of overlapping use from 2009 to 2013. RESULTS:In the historical comparison, 1108 women initiated TDF/FTC/EFV and 637 initiated ZDV/3TC/NVP. In the contemporaneous comparison, 1052 initiated TDF/FTC/EFV and 298 initiated ZDV/3TC/NVP. TDF/FTC/EFV initiators were younger and more likely to be nulliparous than ZDV/3TC/NVP initiators in both comparisons. In the historical comparison, the adjusted risk ratios (95% confidence interval) comparing ZDV/3TC/NVP with TDF/FTC/EFV were 2.95 (1.76, 4.96) for stillbirth, 1.40 (1.17, 1.67) for preterm delivery, 2.58 (1.70, 3.91) for very preterm delivery, 1.96 (1.64, 2.34) for SGA, 2.32 (1.73, 3.09) for very SGA, 1.54 (1.38, 1.72) for any adverse birth outcome, and 2.20 (1.76, 2.75) for any severe birth outcome, and were similar in the contemporaneous comparison. CONCLUSION/CONCLUSIONS:Preconception initiation of ZDV/3TC/NVP compared with TDF/FTC/EFV may increase the risk of adverse birth outcomes.

There is strong evidence linking stimulant use, namely methamphetamine use, to sexual risk behavior among sexual minority men (SMM); we do not, however, have a good understanding of this relationship among other at-risk populations. In this study, we systematically reviewed associations between stimulant use (i.e., methamphetamine, crack cocaine, cocaine) and sexual risk behaviors among populations facing elevated risk of HIV transmission and acquisition (i.e., SMM, people who inject drugs (PWID), and people living with HIV/AIDS (PLWH)). Random-effects meta-analyses and sensitivity analyses that included crude and adjusted estimates separately were conducted to evaluate the impact of potential confounding variables. The results showed strong relationships between stimulant use and condomless sex, transactional sex, and multiple sexual partners. Results were broadly consistent when analyses were stratified by type of stimulant (methamphetamine, crack cocaine, and other stimulants) and risk group. Sensitivity analyses with confounding variables did not greatly impact results. The results indicate that stimulant use is associated with numerous sexual risk behaviors regardless of risk group, suggesting prevention efforts focused on reducing methamphetamine-related HIV risk should target a range of at-risk populations.

BACKGROUND:There has been a significant increase in methamphetamine use among persons who use drugs in Vietnam in the last 5-10 years. We examined the degree to which adherence to antiretroviral therapy (ART) mediates the relationship between recent methamphetamine use and unsuppressed HIV viral load among people who inject drugs (PWID) in Hai Phong, Vietnam. METHODS:We recruited PWID from October 2016-October 2018 and enrolled HIV positive PWID into a cohort, with up to three years of total follow-up. We assessed relationships among recent methamphetamine use frequency, ART adherence and unsuppressed HIV viral load. Mediation analysis was used to estimate the total and natural direct effects of recent methamphetamine use on unsuppressed HIV viral load and the indirect effect proportion. RESULTS:We enrolled 792 HIV seropositive PWID into the cohort; approximately 75.9% reported high/perfect ART adherence at baseline and 81.3% were virally suppressed. In mediation analysis, the total effect for the association between methamphetamine use and unsuppressed HIV viral load (1000 copies/mL) was 3.94 (95% CI: 1.95, 7.96); the natural direct effect was 2.14 (95% CI: 1.29, 3.55); the proportion mediated by self-reported ART adherence was 0.444. Similar results were found when examining lower unsuppressed HIV viral load cutpoints of 250 copies/mL and 500 copies/mL. CONCLUSIONS:Methamphetamine use is associated with unsuppressed HIV viral load among PWID despite high levels of ART adherence. Further research is needed to better understand these relationships, with emphasis on potential biological pathways that may interact with ART.