Search

While randomized controlled trials (RCTs) are critical for establishing the efficacy of new therapies, there are limitations regarding what comparisons can be made directly from trial data. RCTs are limited to a small number of comparator arms and often compare a new therapeutic to a standard of care which has already proven efficacious. It is sometimes of interest to estimate the efficacy of the new therapy relative to a treatment that was not evaluated in the same trial, such as a placebo or an alternative therapy that was evaluated in a different trial. Such multi-study comparisons are challenging because of potential differences between trial populations that can affect the outcome. In this paper, two bridging estimators are considered that allow for comparisons of treatments evaluated in different trials using data fusion methods to account for measured differences in trial populations. A "multi-span'' estimator leverages a shared arm between two trials, while a "single-span'' estimator does not require a shared arm. A diagnostic statistic that compares the outcome in the standardized shared arms is provided. The two estimators are compared in simulations, where both estimators demonstrate minimal empirical bias and nominal confidence interval coverage when the identification assumptions are met. The estimators are applied to data from the AIDS Clinical Trials Group 320 and 388 to compare the efficacy of two-drug versus four-drug antiretroviral therapy on CD4 cell counts among persons with advanced HIV. The single-span approach requires fewer identification assumptions and was more efficient in simulations and the application.

Geographic mobility may disrupt continuity of care and contribute to poor clinical outcomes among people receiving treatment for tuberculosis (TB). This may occur especially where health services are not well coordinated across international borders, particularly in lower and middle income country settings. In this work, we describe mobility and the relationship between mobility and unfavorable TB treatment outcomes (i.e., death, loss to follow-up, or treatment failure) among a cohort of adults who initiated TB treatment at one of 12 health facilities near Lake Victoria. We abstracted data from health facility records for all 776 adults initiating TB treatment during a 6-month period at the selected facilities in Kenya, Tanzania, and Uganda. We interviewed 301 cohort members to assess overnight travel outside one’s residential district/sub-county. In our analyses, we estimated the proportion of cohort members traveling in 2 and 6 months following initiation of TB treatment, explored correlates of mobility, and examined the association between mobility and an unfavorable TB treatment outcome. We estimated that 40.7% (95% CI: 33.3%, 49.6%) of people on treatment for TB traveled overnight at least once in the 6 months following treatment initiation. Mobility was more common among people who worked in the fishing industry and among those with extra-pulmonary TB. Mobility was not strongly associated with other characteristics examined, however, suggesting that efforts to improve TB care for mobile populations should be broad ranging. We found that in this cohort, people who were mobile were not at increased risk of an unfavorable TB treatment outcome. Findings from this study can help inform development and implementation of mobility-competent health services for people with TB in East Africa.

Iterated conditional expectation (ICE) g-computation is an estimation approach for addressing time-varying confounding for both longitudinal and time-to-event data. Unlike other g-computation implementations, ICE avoids the need to specify models for each time-varying covariate. For variance estimation, previous work has suggested the bootstrap. However, bootstrapping can be computationally intense and sensitive to the number of resamples used. Here, we present ICE g-computation as a set of stacked estimating equations. Therefore, the variance for the ICE g-computation estimator can be estimated using the empirical sandwich variance estimator. Performance of the variance estimator was evaluated empirically with a simulation study. The proposed approach is also demonstrated with an illustrative example on the effect of cigarette smoking on the prevalence of hypertension. In the simulation study, the empirical sandwich variance estimator appropriately estimated the variance. When comparing runtimes between the sandwich variance estimator and the bootstrap for the applied example, the sandwich estimator was substantially faster, even when bootstraps were run in parallel. The empirical sandwich variance estimator is a viable option for variance estimation with ICE g-computation., Comment: 17 pages

Comparisons of treatments or exposures are of central interest in epidemiology, but direct comparisons are not always possible due to practical or ethical reasons. Here, we detail a fusion approach to compare treatments across studies. The motivating example entails comparing the risk of the composite outcome of death, AIDS, or greater than a 50% CD4 cell count decline in people with HIV when assigned triple versus mono antiretroviral therapy, using data from the AIDS Clinical Trial Group (ACTG) 175 (mono versus dual therapy) and ACTG 320 (dual versus triple therapy). We review a set of identification assumptions and estimate the risk difference using an inverse probability weighting estimator that leverages the shared trial arms (dual therapy). A fusion diagnostic based on comparing the shared arms is proposed that may indicate violation of the identification assumptions. Application of the data fusion estimator and diagnostic to the ACTG trials indicates triple therapy results in a reduction in risk compared to monotherapy in individuals with baseline CD4 counts between 50 and 300 cells/mm$^3$. Bridged treatment comparisons address questions that none of the constituent data sources could address alone, but valid fusion-based inference requires careful consideration., Comment: 20 pages, 3 figures, 5 tables

Additional file 1: Table S1. Prevalence estimates for each district council included in the study. Table S2. Estimated beta values and 95% uncertainty intervals for covariates included in adjusted models for land cover and log-odds of malaria prevalence, stratified by transmission intensity. Table S3. Posterior malaria odds ratio estimates for cropland, grassland, and forest land cover stratified by transmission intensity comparing adjusted, non-spatial model results and adjusted model results including a spatially varying intercept.

In an analysis of randomized trials, use of efavirenz for treatment of human immunodeficiency virus (HIV) infection was associated with increased suicidal thoughts/behaviors. However, analyses of observational data have found no evidence of increased risk. To assess whether population differences might explain this divergence, we transported the effect of efavirenz use from these trials to a specific target population. Using inverse odds weights and multiple imputation, we transported the effect of efavirenz on suicidal thoughts/behaviors in these randomized trials (participants were enrolled in 2001-2007) to a trials-eligible cohort of US adults initiating antiretroviral therapy while receiving HIV clinical care at medical centers between 1999 and 2015. Overall, 8,291 cohort participants and 3,949 trial participants were eligible. Prescription of antidepressants (19% vs. 13%) and injection drug history (16% vs. 10%) were more frequent in the cohort than in the trial participants. Compared with the effect in trials, the estimated hazard ratio for efavirenz on suicidal thoughts/behaviors was attenuated in our target population (trials:hazard ratio (HR) =2.3 (95% confidence interval (CI): 1.2, 4.4); transported: HR=1.8 (95% CI: 0.9, 4.4)), whereas the incidence rate difference was similar (trials: HR=5.1 (95% CI: 1.6, 8.7); transported: HR=5.4 (95% CI: -0.4, 11.4)). In our target population, there was greater than 20% attenuation of the hazard ratio estimate as compared with the trials-only estimate. Transporting results from trials to a target population is informative for addressing external validity.

The article of record as published may be found at https://doi.org/10.1097/QAD.0000000000002502 Design: The Joint United Nations Programme on HIV/AIDS (UNAIDS) 90–90–90 and other cross-sectional metrics can lead to potentially counterintuitive conclusions when used to evaluate health systems’ performance. This study demonstrates how time and population dynamics impact UNAIDS 90–90–90 metrics in comparison with a longitudinal analogue. Methods: We simulated a large cohort over 15 years. Everyone started out at risk for HIV, and then transitioned through the HIV care continuum based on fixed daily probabilities of acquiring HIV, learning status, entering care, initiating antiretroviral therapy (ART), and becoming virally suppressed, or dying. We varied the probability of ART initiation over three five-year periods (low, high, and low). We repeated the simulation with an increased probability of death. Results: The cross-sectional probability of being on ART among persons who were diagnosed responded relatively slowly to changes in the rate of ART initiation. Increases in ART initiation rates caused apparent declines in the cross-sectional probability of being virally suppressed among persons who had initiated ART, despite no changes in the rate of viral suppression. In some cases, higher mortality resulted in the cross-sectional metrics implying improved healthcare system performance. The longitudinal continuum was robust to these issues. Conclusion: The UNAIDS 90–90–90 care continuum may lead to incorrect inference when used to evaluate health systems performance. We recommend that evaluation of HIV care delivery include longitudinal care continuum metrics wherever possible.

Supplemental material, HEB831755_Supplemental_Material for The Relationship Between School Dropout and Pregnancy Among Adolescent Girls and Young Women in South Africa: A HPTN 068 Analysis by Marie C. D. Stoner, Katherine B. Rucinski, Jessie K. Edwards, Amanda Selin, James P. Hughes, Jing Wang, Yaw Agyei, F. Xavier Gomez-Olive, Catherine MacPhail, Kathleen Kahn and Audrey Pettifor in Health Education & Behavior

Heavy drinking is prevalent among people living with HIV. Studies use tools like patient-reported outcomes (PROs) to quantify alcohol use in a detailed, timely manner. However, if alcohol misuse influences PRO completion, selection bias may result. Our study included 14,145 adult HIV patients (133,036 visits) from CNICS who were eligible to complete PROs at an HIV primary care visit. We compared PRO completion proportions between patients with and without a clinical diagnosis of at-risk alcohol use in the prior year. We accounted for confounding by baseline and visit-specific covariates. PROs were completed at 20.8% of assessed visits. The adjusted difference in PRO completion proportions was -3.2% (95% CI -5.6 to -0.8%). The small association between receipt of an at-risk alcohol use diagnosis and decreased PRO completion suggests there could be modest selection bias in studies using the PRO alcohol measure.

Marginal structural models for time-fixed treatments fit using inverse-probability weighted estimating equations are increasingly popular. Nonetheless, the resulting effect estimates are subject to finite-sample bias when data are sparse, as is typical for large-sample procedures. Here we propose a semi-Bayes estimation approach which penalizes or shrinks the estimated model parameters to improve finite-sample performance. This approach uses simple symmetric data-augmentation priors. Limited simulation experiments indicate that the proposed approach reduces finite-sample bias and improves confidence-interval coverage when the true values lie within the central "hill" of the prior distribution. We illustrate the approach with data from a nonexperimental study of HIV treatments.

ObjectiveTo explore the effectiveness of raltegravir-based antiretroviral therapy (ART) on treatment response among ART-naive patients seeking routine clinical care.DesignCohort study of adults enrolled in HIV care in the United States.MethodsWe compared virologic suppression and CD4 cell count recovery over a 2.5 year period after initiation of an ART regimen containing raltegravir or efavirenz using observational data from a US clinical cohort, generalized to the US population of people with diagnosed HIV. We accounted for nonrandom treatment assignment, informative censoring, and nonrandom selection from the US target population using inverse probability weights.ResultsOf the 2843 patients included in the study, 2476 initiated the efavirenz-containing regimen and 367 initiated the raltegravir-containing regimen. In the weighted intent-to-treat analysis, patients spent an average of 74 (95% confidence interval: 41, 106) additional days alive with a suppressed viral load on the raltegravir regimen than on the efavirenz regimen over the 2.5-year study period. CD4 cell count recovery was also superior under the raltegravir regimen.ConclusionPatients receiving raltegravir spent more time alive and suppressed than patients receiving efavirenz, but the probability of viral suppression by 2.5 years after treatment was similar between groups. Optimizing the amount of time spent in a state of viral suppression is important to improve survival among people living with HIV and to reduce onward transmission.

Background and objective: Stacking is an ensemble machine learning method that averages predictions from multiple other algorithms, such as generalized linear models and regression trees. An implementation of stacking, called super learning, has been developed as a general approach to supervised learning and has seen frequent usage, in part due to the availability of an R package. We develop super learning in the SAS software system using a new macro, and demonstrate its performance relative to the R package. Methods: Following previous work using the R SuperLearner package we assess the performance of super learning in a number of domains. We compare the R package with the new SAS macro in a small set of simulations assessing curve fitting in a predictive model as well in a set of 14 publicly available datasets to assess cross-validated accuracy. Results: Across the simulated data and the publicly available data, the SAS macro performed similarly to the R package, despite a different set of potential algorithms available natively in R and SAS. Conclusions: Our super learner macro performs as well as the R package at a number of tasks. Further, by extending the macro to include the use of R packages, the macro can leverage both the robust, enterprise oriented procedures in SAS and the nimble, cutting edge packages in R. In the spirit of ensemble learning, this macro extends the potential library of algorithms beyond a single software system and provides a simple avenue into machine learning in SAS., Comment: 7 pages, 1 table, 3 figures

Background.The long-term effectiveness of human immunodeficiency virus (HIV) treatments containing integrase inhibitors is unknown.Methods.We use observational data from the Centers for AIDS Research Network of Integrated Clinical Systems and the Centers for Disease Control and Prevention to estimate 4-year risk of AIDS and all-cause mortality among 415 patients starting a raltegravir regimen compared to 2646 starting an efavirenz regimen (both regimens include emtricitabine and tenofovir disoproxil fumarate). We account for confounding and selection bias as well as generalizability by standardization for measured variables, and present both observational intent-to-treat and per-protocol estimates.Results.At treatment initiation, 12% of patients were female, 36% black, 13% Hispanic; median age was 37 years, CD4 count 321 cells/µL, and viral load 4.5 log10 copies/mL. Two hundred thirty-five patients incurred an AIDS-defining illness or died, and 741 patients left follow-up. After accounting for measured differences, the 4-year risk was similar among those starting both regimens (ie, intent-to treat hazard ratio [HR], 0.96 [95% confidence interval {CI}, .63-1.45]; risk difference, -0.9 [95% CI, -4.5 to 2.7]), as well as among those remaining on regimens (ie, per-protocol HR, 0.95 [95% CI, .59-1.54]; risk difference, -0.5 [95% CI, -3.8 to 2.9]).Conclusions.Raltegravir and efavirenz-based initial antiretroviral therapy have similar 4-year clinical effects. Vigilance regarding longer-term comparative effectiveness of HIV regimens using observational data is needed because large-scale experimental data are not forthcoming.

BACKGROUND: Normative guidelines from the World Health Organization recommend tracking strategic information indicators among key populations. Monitoring progress in the global response to the HIV epidemic uses indicators put forward by the Joint United Nations Programme on HIV/AIDS. These include the 90-90-90 targets that require a realignment of surveillance data, routinely collected program data, and medical record data, which historically have developed separately. OBJECTIVE: The aim of this study was to describe current challenges for monitoring HIV-related strategic information indicators among key populations ((men who have sex with men [MSM], people in prisons and other closed settings, people who inject drugs, sex workers, and transgender people) and identify future opportunities to enhance the use of surveillance data, programmatic data, and medical record data to describe the HIV epidemic among key populations and measure the coverage of HIV prevention, care, and treatment programs. METHODS: To provide a historical perspective, we completed a scoping review of the expansion of HIV surveillance among key populations over the past three decades. To describe current efforts, we conducted a review of the literature to identify published examples of SI indicator estimates among key populations. To describe anticipated challenges and future opportunities to improve measurement of strategic information indicators, particularly from routine program and health data, we consulted participants of the Third Global HIV Surveillance Meeting in Bangkok, where the 2015 World Health Organization strategic information guidelines were launched. RESULTS: There remains suboptimal alignment of surveillance and programmatic data, as well as routinely collected medical records to facilitate the reporting of the 90-90-90 indicators for HIV among key populations. Studies (n=3) with estimates of all three 90-90-90 indicators rely on cross-sectional survey data. Programmatic data and medical record data continue to be insufficiently robust to provide estimates of the 90-90-90 targets for key populations. CONCLUSIONS: Current reliance on more active data collection processes, including key population-specific surveys, remains warranted until the quality and validity of passively collected routine program and medical record data for key populations is optimized.

BackgroundThe goal of targeted antiretroviral therapy initiation is to minimize disease progression among patients with human immunodeficiency virus while minimizing the therapeutic burden on these patients. We examine whether the effect of delaying therapy initiation from 500 cells/mm(3) to 350 or 200 cells/mm(3) is modified by age at entry into care.MethodsWe used the parametric g-formula to compare 10-year mortality under 3 CD4 cell count thresholds for therapy initiation among 3532 patients who entered care at 1 of 8 sites in the United States between 1998 and 2013. Results are reported separately for patients 18 to 34, 35 to 44, and 45 to 65 years of age at study entry.ResultsIn the observed data, 10-year mortality was 13% (165 deaths). Mortality increased from 11% under therapy initiation at 500 cells/mm(3) to 12% at 350 cells/mm(3) (risk difference [RD]: 0.87; 95% confidence interval [CI]: .56, 2.17) and to 14% at 200 cells/mm(3) (RD: 2.71; 95% CI: 1.79, 5.38). The effect of delaying therapy became greater with age: RDs comparing the 350-cells/mm(3) threshold with the 500-cells/mm(3) threshold ranged from -0.03 (95% CI: -0.15, 1.76) for patients 18 to 34 years of age to 0.99 (95% CI: -.27, 1.98) for patients 35 to 44 and to 2.30 (95% CI: 1.29, 5.42) for patients 45 to 65.ConclusionsDelaying therapy increased 10-year mortality in the full cohort. Subgroup analysis highlights that patients entering care at older ages may be more vulnerable to the consequences of delayed ART initiation than younger patients.

BackgroundUsing a dynamic visual display, we examine the changes in human immunodeficiency virus type 1 (HIV-1) plasma viral load and CD4 cell count for 5 years after antiretroviral therapy initiation in a large cohort of patients with HIV.MethodsPatients at a Centers for AIDS Research Network of Integrated Clinical Systems site who initiated combination antiretroviral therapy between 1 January 2000 and 31 December 2012 were followed for 5 years for HIV-1 plasma viral load, CD4 cell count, and mortality. The joint distribution of CD4 cell count and viral load over time was depicted in an animated display using a bivariate kernel smoother.ResultsWithin days of therapy initiation, many patients had a suppressed viral load and their median CD4 cell count had increased. However, the median CD4 cell count remained below normal levels throughout follow-up period and the proportion of patients with high viral load occasionally increased, even years after therapy initiation.ConclusionsThe dramatic changes in viral load and CD4 cell count after therapy initiation highlight the overwhelming effectiveness of antiretroviral therapy in the modern era. However, this work also emphasizes the need for pharmaceutical or behavioral interventions to prevent virologic failure and to stimulate complete recovery of normal CD4 cell count.