Tree-based subgroup discovery using electronic health record data: heterogeneity of treatment effects for DTG-containing therapies.

The rich longitudinal individual level data available from electronic health records (EHRs) can be used to examine treatment effect heterogeneity. However, estimating treatment effects using EHR data poses several challenges, including time-varying confounding, repeated and temporally non-aligned measurements of covariates, treatment assignments and outcomes, and loss-to-follow-up due to dropout. Here, we develop the subgroup discovery for longitudinal data algorithm, a tree-based algorithm for discovering subgroups with heterogeneous treatment effects using longitudinal data by combining the generalized interaction tree algorithm, a general data-driven method for subgroup discovery, with longitudinal targeted maximum likelihood estimation. We apply the algorithm to EHR data to discover subgroups of people living with human immunodeficiency virus who are at higher risk of weight gain when receiving dolutegravir (DTG)-containing antiretroviral therapies (ARTs) versus when receiving non-DTG-containing ARTs.

Recommendations on data sharing in HIV drug resistance research.

• Human immunodeficiency virus (HIV) drug resistance has implications for antiretroviral treatment strategies and for containing the HIV pandemic because the development of HIV drug resistance leads to the requirement for antiretroviral drugs that may be less effective, less well-tolerated, and more expensive than those used in first-line regimens. • HIV drug resistance studies are designed to determine which HIV mutations are selected by antiretroviral drugs and, in turn, how these mutations affect antiretroviral drug susceptibility and response to future antiretroviral treatment regimens. • Such studies collectively form a vital knowledge base essential for monitoring global HIV drug resistance trends, interpreting HIV genotypic tests, and updating HIV treatment guidelines. • Although HIV drug resistance data are collected in many studies, such data are often not publicly shared, prompting the need to recommend best practices to encourage and standardize HIV drug resistance data sharing. • In contrast to other viruses, sharing HIV sequences from phylogenetic studies of transmission dynamics requires additional precautions as HIV transmission is criminalized in many countries and regions. • Our recommendations are designed to ensure that the data that contribute to HIV drug resistance knowledge will be available without undue hardship to those publishing HIV drug resistance studies and without risk to people living with HIV.

Characterization of HIV-Related Stigma in Myanmar.

Characterizing HIV-related stigma and its impacts are important for interventions toward their elimination. A cross-sectional study was conducted in 2016 to evaluate enacted and internalized stigma among adult people living with HIV (PLWH) across four cities in Myanmar using the India Stigma Index questionnaire. Multivariable regression analyses were performed to determine differences in measured enacted and internalized stigma outcomes. Among 1,006 participants, 89% reported any stigma indicator, 47% enacted stigma, and 87% internalized stigma. In regression analysis, city and duration of illness were associated with higher enacted stigma, and younger age was associated with higher internalized stigma. Those with HIV duration > 7.4 years had mean enacted stigma nearly 2 units higher than the overall mean. Internalized stigma increased with duration of illness and leveled off at 5 years. PLWH from smaller cities experienced lower stigma. In Myanmar, nearly 90% of PLWH experience stigma, results that reflect a unique transition point.

Evaluation of peer-to-peer HIV counseling in Myanmar: a measure of knowledge, adherence, and barriers.

In Myanmar, an Asian country with one of the highest HIV-1 prevalence rates, counseling prior to initiating antiretroviral therapy (ART) is standard care, either by a healthcare worker (standard counselor, SC) or trained counselor who is also living with HIV (peer counselor, PC). PC is commonly utilized in Myanmar and other resource-limited settings. However, its benefit over SC is unclear. We conducted a cross-sectional survey of people living with HIV (PLWH), who completed either only PC or only SC before treatment initiation across four cities in Myanmar. Participants were evaluated for HIV knowledge, stigma, antiretroviral adherence, barriers to care, social support satisfaction and attitudes regarding both counseling processes. Bivariate analyses and multivariable mixed effects modeling were conducted to compare differences in these measures among PC and SC participants. Among 1006 participants (49% PC; 51% SC), 52% were females and median age was 37 years in those receiving PC and 40 years in those receiving SC. More than 70% of participants in both groups achieved up to grade school education. The average duration since HIV diagnosis was 4.6 years for PC and 5.7 years for SC participants. HIV knowledge and attitudes regarding counseling were good in both groups and more PC participants credited their HIV counselor for knowledge (75% vs 63%, p < 0.001). Compared to SC, PC participants had lower enacted stigma (Incidence Rate Ratio (IRR) 0.75, Confidence Interval (CI) [0.65, 0.86]), mean internalized stigma (-0.24, CI [-0.34, -0.14]), and risk of antiretroviral therapy non-adherence (Odds Ratio 0.59, CI [0.40, 0.88]), while reporting higher levels of barriers to care (9.63, CI [8.20, 11.75]). Our findings demonstrate potential benefits of PC compared to SC, and support the utilization of PC to enhance HIV health outcomes within the unique societal and geographical context of Myanmar, and possibly beyond.

"I have never talked to anyone to free my mind" - challenges surrounding status disclosure to adolescents contribute to their disengagement from HIV care: a qualitative study in western Kenya.

INTRODUCTION: Adolescents living with HIV (ALHIV, ages 10-19) experience complex barriers to care engagement. Challenges surrounding HIV status disclosure or non-disclosure to adolescents may contribute to adolescent disengagement from HIV care or non-adherence to ART. We performed a qualitative study to investigate the contribution of disclosure challenges to adolescent disengagement from HIV care. METHODS: This was a qualitative study performed with disengaged ALHIV and their caregivers, and with healthcare workers (HCW) in the Academic Model Providing Access to Healthcare (AMPATH) program in western Kenya. Inclusion criteria for ALHIV were ≥1 visit within the 18 months prior to data collection at one of two clinical sites and nonattendance ≥60 days following their last scheduled appointment. HCW were recruited from 10 clinics. Analysis was conducted by multiple independent coders, and narratives of disclosure and care disengagement were closely interrogated. Overarching themes were elucidated and summarized. RESULTS: Interviews were conducted with 42 disengaged ALHIV, 32 caregivers, and 28 HCW. ALHIV were average age 17.0 (range 12.9-20.9), and 95% indicated awareness of their HIV diagnosis. Issues surrounding disclosure to ALHIV presented important barriers to HIV care engagement. Themes centered on delays in HIV status disclosure; hesitancy and reluctance among caregivers to disclose; struggles for adolescents to cope with feelings of having been deceived prior to full disclosure; pervasive HIV stigma internalized in school and community settings prior to disclosure; and inadequate and unstructured support after disclosure, including for adolescent mental health burdens and for adolescent-caregiver relationships and communication. Both HCW and caregivers described feeling inadequately prepared to optimally handle disclosure and to manage challenges that may arise after disclosure. CONCLUSIONS: Complex challenges surrounding HIV status disclosure to adolescents contribute to care disengagement. There is need to enhance training and resources for HCW, and to empower caregivers to support children and adolescents before, during, and after HIV status disclosure. This should include counseling caregivers on how to provide children with developmentally-appropriate and accurate information about their health from an early age, and to support adolescent-caregiver communication and relationships. Optimally integrating peer support can further promote ALHIV wellbeing and retention in care.

Measurement error and variant-calling in deep Illumina sequencing of HIV.

MOTIVATION: Next-generation deep sequencing of viral genomes, particularly on the Illumina platform, is increasingly applied in HIV research. Yet, there is no standard protocol or method used by the research community to account for measurement errors that arise during sample preparation and sequencing. Correctly calling high and low-frequency variants while controlling for erroneous variants is an important precursor to downstream interpretation, such as studying the emergence of HIV drug-resistance mutations, which in turn has clinical applications and can improve patient care. RESULTS: We developed a new variant-calling pipeline, hivmmer, for Illumina sequences from HIV viral genomes. First, we validated hivmmer by comparing it to other variant-calling pipelines on real HIV plasmid datasets. We found that hivmmer achieves a lower rate of erroneous variants, and that all methods agree on the frequency of correctly called variants. Next, we compared the methods on an HIV plasmid dataset that was sequenced using Primer ID, an amplicon-tagging protocol, which is designed to reduce errors and amplification bias during library preparation. We show that the Primer ID consensus exhibits fewer erroneous variants compared to the variant-calling pipelines, and that hivmmer more closely approaches this low error rate compared to the other pipelines. The frequency estimates from the Primer ID consensus do not differ significantly from those of the variant-calling pipelines. AVAILABILITY AND IMPLEMENTATION: hivmmer is freely available for non-commercial use from https://github.com/kantorlab/hivmmer. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Classification using ensemble learning under weighted misclassification loss.

Binary classification rules based on covariates typically depend on simple loss functions such as zero-one misclassification. Some cases may require more complex loss functions. For example, individual-level monitoring of HIV-infected individuals on antiretroviral therapy requires periodic assessment of treatment failure, defined as having a viral load (VL) value above a certain threshold. In some resource limited settings, VL tests may be limited by cost or technology, and diagnoses are based on other clinical markers. Depending on scenario, higher premium may be placed on avoiding false-positives, which brings greater cost and reduced treatment options. Here, the optimal rule is determined by minimizing a weighted misclassification loss/risk. We propose a method for finding and cross-validating optimal binary classification rules under weighted misclassification loss. We focus on rules comprising a prediction score and an associated threshold, where the score is derived using an ensemble learner. Simulations and examples show that our method, which derives the score and threshold jointly, more accurately estimates overall risk and has better operating characteristics compared with methods that derive the score first and the cutoff conditionally on the score especially for finite samples.

High discordance in blood and genital tract HIV-1 drug resistance in Indian women failing first-line therapy.

Objectives: Examine HIV-1 plasma viral load (PVL) and genital tract (GT) viral load (GVL) and drug resistance in India. Methods: At the YRG Centre for AIDS Research and Education, Chennai, we tested: PVL in women on first-line ART for ≥6 months; GVL when PVL >2000 copies/mL; and plasma, genital and proviral reverse transcriptase drug resistance when GVL >2000 copies/mL. Wilcoxon rank-sum and Fisher's exact tests were used to identify failure and resistance associations. Pearson correlations were calculated to evaluate PVL-GVL associations. Inter-compartmental resistance discordance was evaluated using generalized estimating equations. Results: Of 200 women, 37% had detectable (>400 copies/mL) PVL and 31% had PVL >1000 copies/mL. Of women with detectable PVL, 74% had PVL >2000 copies/mL, of which 74% had detectable GVL. Higher PVL was associated with higher GVL. Paired plasma and genital sequences were available for 21 women; mean age of 34 years, median ART duration of 33 months, median CD4 count of 217 cells/mm3, median PVL of 5.4 log10 copies/mL and median GVL of 4.6 log10 copies/mL. Drug resistance was detected in 81%-91% of samples and 67%-76% of samples had dual-class resistance. Complete three-compartment concordance was seen in only 10% of women. GT-proviral discordance was significantly larger than plasma-proviral discordance. GT or proviral mutations discordant from plasma led to clinically relevant resistance in 24% and 30%, respectively. Conclusions: We identified high resistance and high inter-compartmental resistance discordance in Indian women, which might lead to unrecognized resistance transmission and re-emergence compromising treatment outcomes, particularly relevant to countries like India, where sexual HIV transmission is predominant.

Regional Challenges in the Prevention of Human Immunodeficiency Virus Drug Resistance.

In diverse global regions with significant human immunodeficiency virus (HIV) burden, programmatic, cultural, and provider-, patient-, and virus-related factors may result in HIV drug resistance, with global implications. This article reviews such common and unique challenges in Russia, Latin America and the Caribbean, China, and India, to suggest potential solutions.

Current Status of Point-of-Care Testing for Human Immunodeficiency Virus Drug Resistance.

Healthcare delivery has advanced due to the implementation of point-of-care testing, which is often performed within minutes to hours in minimally equipped laboratories or at home. Technologic advances are leading to point-of-care kits that incorporate nucleic acid-based assays, including polymerase chain reaction, isothermal amplification, ligation, and hybridization reactions. As a limited number of single-nucleotide polymorphisms are associated with clinically significant human immunodeficiency virus (HIV) drug resistance, assays to detect these mutations have been developed. Early versions of these assays have been used in research. This review summarizes the principles underlying each assay and discusses strategic needs for their incorporation into the management of HIV infection.

Antiretroviral treatment failure, drug resistance, and subtype diversity in the only pediatric HIV clinic in Rhode Island.

BACKGROUND: Drug resistance development in the human immunodeficiency virus (HIV)-infected pediatric population in the United States can impact long-term antiretroviral therapy (ART) efficacy. Limited formularies and adherence constraints in children jeopardize lifelong-needed ART. METHODS: We examined treatment failure, drug resistance, and their correlates in ART-naive and ART-experienced children attending the pediatric HIV clinic in Rhode Island between 1991 and 2012. Pol sequences were obtained for phylogenetic, subtype, and resistance analyses. Associations between selected covariates and virologic failure and resistance were evaluated using generalized additive models and Fisher exact tests. RESULTS: Data were available for all 56 clinic-attending children. At diagnosis, 33% were aged <1 year, 31% aged 1-4 years, and 37% aged ≥ 5 years; 54% were male, 73% black or Hispanic, 55% US-born, 20% refugees, and 64% perinatally infected. Of 44 ART-experienced children, 57% had virologic failure, most never virologically suppressed. Failure was associated with missed appointments (P = .05) and missed doses (P < .01). Of 40 children with available genotypes, 35% were infected with non-B subtypes; 6% of ART-naive children had resistance; and 73% of ART-experienced children had ≥ 1 major mutation: (16% conferring triple-class, 47% dual-class, and 37% single-class resistance). An epidemiologically confirmed resistance transmission from a perinatally infected teenage male to a newly infected teenage female was demonstrated. CONCLUSIONS: We report high HIV type 1 diversity, extensive drug resistance among ART-experienced children, and horizontal transmission of resistance in the Rh ode Island pediatric HIV clinic. As HIV-infected children mature into adulthood, close monitoring of ART, adherence, and diagnosis disclosure are essential to optimize patient care.

Pretreatment HIV Drug Resistance and HIV-1 Subtype C Are Independently Associated With Virologic Failure: Results From the Multinational PEARLS (ACTG A5175) Clinical Trial.

BACKGROUND: Evaluation of pretreatment HIV genotyping is needed globally to guide treatment programs. We examined the association of pretreatment (baseline) drug resistance and subtype with virologic failure in a multinational, randomized clinical trial that evaluated 3 antiretroviral treatment (ART) regimens and included resource-limited setting sites. METHODS: Pol genotyping was performed in a nested case-cohort study including 270 randomly sampled participants (subcohort), and 218 additional participants failing ART (case group). Failure was defined as confirmed viral load (VL) >1000 copies/mL. Cox proportional hazards models estimated resistance-failure association. RESULTS: In the representative subcohort (261/270 participants with genotypes; 44% women; median age, 35 years; median CD4 cell count, 151 cells/µL; median VL, 5.0 log10 copies/mL; 58% non-B subtypes), baseline resistance occurred in 4.2%, evenly distributed among treatment arms and subtypes. In the subcohort and case groups combined (466/488 participants with genotypes), used to examine the association between resistance and treatment failure, baseline resistance occurred in 7.1% (9.4% with failure, 4.3% without). Baseline resistance was significantly associated with shorter time to virologic failure (hazard ratio [HR], 2.03; P = .035), and after adjusting for sex, treatment arm, sex-treatment arm interaction, pretreatment CD4 cell count, baseline VL, and subtype, was still independently associated (HR, 2.1; P = .05). Compared with subtype B, subtype C infection was associated with higher failure risk (HR, 1.57; 95% confidence interval [CI], 1.04-2.35), whereas non-B/C subtype infection was associated with longer time to failure (HR, 0.47; 95% CI, .22-.98). CONCLUSIONS: In this global clinical trial, pretreatment resistance and HIV-1 subtype were independently associated with virologic failure. Pretreatment genotyping should be considered whenever feasible. CLINICAL TRIALS REGISTRATION: NCT00084136.

Transmitted HIV drug resistance at the Thai Red Cross anonymous clinic in Bangkok: results from three consecutive years of annual surveillance.

OBJECTIVES: The aim of this study was to prospectively survey transmitted drug resistance (TDR) among recently infected individuals (mostly MSM). METHODS: TDR was determined in prospective annual cohorts of recently HIV-1-infected individuals consecutively recruited from 2008 to 2010. Resistance interpretation was carried out using Stanford Database tools and the WHO surveillance drug resistance mutation list. Kruskal-Wallis and Fisher's exact tests were used to compare demographic and laboratory outcomes. RESULTS: A total of 299 subjects were enrolled, with 89% MSM. Median viral load was significantly higher in 2010 than in 2008 (P=0.004). Of the 284 analysable reverse transcriptase/protease sequences, TDR to any drug was found in 14/284 (4.9%); 4.0% in 2008, 5.9% in 2009 and 5.3% in 2010, with an increasing trend of TDR to NRTIs and NNRTIs from 2008 to 2010 (P=0.07). Good correlation was found between our data and the WHO threshold surveillance method. Only rilpivirine had significantly higher (P<0.05) predicted resistance in 2010 than in 2008 and 2009. CONCLUSIONS: A trend towards an increase in TDR in Thailand where the major epidemic is among MSM was observed, but did not reach the WHO-defined high-level threshold (>15%). Attention to prevent the development and spread of drug resistance is needed.

We should not stop considering HIV drug resistance testing at failure of first-line antiretroviral therapy.

HIV drug resistance is a major global hurdle to successful and sustained antiretroviral therapy. Global guidelines recommend testing for antiretroviral drug resistance and results are used to inform treatment regimen design for patients at different stages of therapy. Several clinical trials investigated optimal regimens after failure of first-line antiretroviral therapy, yielding data that advanced knowledge and informed care. However, further interpretation of data from these studies questioned the benefit of antiretroviral drug resistance testing for cases in which first-line treatment is not effective and, furthermore, that relying on the results of antiretroviral drug resistance testing could be misleading and unnecessary. In this Viewpoint, which is largely focused on high-income settings, we review these data, reflect on the potential problems with their interpretation, and call for caution in their extrapolation. Without negating the importance of the data, and recognising the varied circumstances related to HIV drug resistance testing in different global settings, we advise caution before changing current practice and recommendations. We believe that we should not universally stop considering HIV drug resistance testing at failure of first-line antiretroviral therapy.

Added Value of Next Generation Sequencing in Characterizing the Evolution of HIV-1 Drug Resistance in Kenyan Youth.

Drug resistance remains a global challenge in children and adolescents living with HIV (CALWH). Characterizing resistance evolution, specifically using next generation sequencing (NGS) can potentially inform care, but remains understudied, particularly in antiretroviral therapy (ART)-experienced CALWH in resource-limited settings. We conducted reverse-transcriptase NGS and investigated short-and long-term resistance evolution and its predicted impact in a well-characterized cohort of Kenyan CALWH failing 1st-line ART and followed for up to ~8 years. Drug resistance mutation (DRM) evolution types were determined by NGS frequency changes over time, defined as evolving (up-trending and crossing the 20% NGS threshold), reverting (down-trending and crossing the 20% threshold) or other. Exploratory analyses assessed potential impacts of minority resistance variants on evolution. Evolution was detected in 93% of 42 participants, including 91% of 22 with short-term follow-up, 100% of 7 with long-term follow-up without regimen change, and 95% of 19 with long-term follow-up with regimen change. Evolving DRMs were identified in 60% and minority resistance variants evolved in 17%, with exploratory analysis suggesting greater rate of evolution of minority resistance variants under drug selection pressure and higher predicted drug resistance scores in the presence of minority DRMs. Despite high-level pre-existing resistance, NGS-based longitudinal follow-up of this small but unique cohort of Kenyan CALWH demonstrated continued DRM evolution, at times including low-level DRMs detected only by NGS, with predicted impact on care. NGS can inform better understanding of DRM evolution and dynamics and possibly improve care. The clinical significance of these findings should be further evaluated.

Not all clusters are equal: dynamics of molecular HIV-1 clusters in a statewide Rhode Island epidemic.

OBJECTIVES: Molecular epidemiology is a powerful tool to characterize HIV epidemics and prioritize public health interventions. Typically, HIV clusters are assumed to have uniform patterns over time. We hypothesized that assessment of cluster evolution would reveal distinct cluster behavior, possibly improving molecular epidemic characterization, towards disrupting HIV transmission. DESIGN: Retrospective cohort. METHODS: Annual phylogenies were inferred by cumulative aggregation of all available HIV-1 pol sequences of individuals with HIV-1 in Rhode Island (RI) between 1990 and 2020, representing a statewide epidemic. Molecular clusters were detected in annual phylogenies by strict and relaxed cluster definition criteria, and the impact of annual newly-diagnosed HIV-1 cases to the structure of individual clusters was examined over time. RESULTS: Of 2153 individuals, 31% (strict criteria) - 47% (relaxed criteria) clustered. Longitudinal tracking of individual clusters identified three cluster types: normal, semi-normal and abnormal. Normal clusters (83-87% of all identified clusters) showed predicted growing/plateauing dynamics, with approximately three-fold higher growth rates in large (15-18%) vs. small (∼5%) clusters. Semi-normal clusters (1-2% of all clusters) temporarily fluctuated in size and composition. Abnormal clusters (11-16% of all clusters) demonstrated collapses and re-arrangements over time. Borderline values of cluster-defining parameters explained dynamics of non-normal clusters. CONCLUSIONS: Comprehensive tracing of molecular HIV clusters over time in a statewide epidemic identified distinct cluster types, likely missed in cross-sectional analyses, demonstrating that not all clusters are equal. This knowledge challenges current perceptions of consistent cluster behavior over time and could improve molecular surveillance of local HIV epidemics to better inform public health strategies.

HIV-Related Stigma Shapes Research Participation for Youth Living With HIV in Kenya.

Introduction: HIV stigma affects medication adherence, psychosocial outcomes, and clinical management for youth living with HIV (YLWH). We explored the impact of HIV stigma on research participation, to inform the ethical engagement of this vulnerable group. Methods: We interviewed 40 YLWH, 20 caregivers, and 39 subject matter experts (SMEs); transcripts were analyzed by HK and EG, with emerging themes confirmed by JA and AC. Results: All categories of participants identified the impacts of stigma on YLWH research participation, suggesting implementing privacy protections, considering recruitment locations carefully, and developing supportive relationships with YLWH. SMEs suggested that YLWH experience uniquely high risks from stigma due to the compounding effects of developmental challenges and transitionary life period. Accidental HIV disclosure and subsequent stigma were identified as a risk of research participation; some viewed the creation of community through research as a benefit. Conclusion: Participants provided insights into stigma-related considerations for research with YLWH, which may guide engagement protocols.

Ethical Considerations for Engaging Children and Adolescents Living with HIV in Research in African Countries: A Systematic Review.

Research engaging children and adolescents living with HIV (CALWH) is critical for youth-friendly services and HIV care, and researchers need to ensure that such engagement is ethical. We conducted a systematic review to identify key ethical considerations for the engagement of CALWH in research. The review focused on primary research articles conducted in African countries that examined ethical issues in CALWH engaged in research. Ten studies met the inclusion criteria; the following seven key domains were extracted: 1) justifications for engaging CALWH in research; 2) community involvement; 3) informed consent/assent; 4) caregiver involvement; 5) perceptions of benefits; 6) perception of the risks of involvement; and 7) confidentiality. These domains can inform the ethical engagement of CALWH in research.

An Automated Bioinformatics Pipeline Informing Near-Real-Time Public Health Responses to New HIV Diagnoses in a Statewide HIV Epidemic.

Molecular HIV cluster data can guide public health responses towards ending the HIV epidemic. Currently, real-time data integration, analysis, and interpretation are challenging, leading to a delayed public health response. We present a comprehensive methodology for addressing these challenges through data integration, analysis, and reporting. We integrated heterogeneous data sources across systems and developed an open-source, automatic bioinformatics pipeline that provides molecular HIV cluster data to inform public health responses to new statewide HIV-1 diagnoses, overcoming data management, computational, and analytical challenges. We demonstrate implementation of this pipeline in a statewide HIV epidemic and use it to compare the impact of specific phylogenetic and distance-only methods and datasets on molecular HIV cluster analyses. The pipeline was applied to 18 monthly datasets generated between January 2020 and June 2022 in Rhode Island, USA, that provide statewide molecular HIV data to support routine public health case management by a multi-disciplinary team. The resulting cluster analyses and near-real-time reporting guided public health actions in 37 phylogenetically clustered cases out of 57 new HIV-1 diagnoses. Of the 37, only 21 (57%) clustered by distance-only methods. Through a unique academic-public health partnership, an automated open-source pipeline was developed and applied to prospective, routine analysis of statewide molecular HIV data in near-real-time. This collaboration informed public health actions to optimize disruption of HIV transmission.