Leveraging machine learning: Covariate-adjusted Bayesian adaptive randomization and subgroup discovery in multi-arm survival trials.

Clinical trials evaluate the safety and efficacy of treatments for specific diseases. Ensuring these studies are well-powered is crucial for identifying superior treatments. With the rise of personalized medicine, treatment efficacy may vary based on biomarker profiles. However, researchers often lack prior knowledge about which biomarkers are linked to varied treatment effects. Fixed or response-adaptive designs may not sufficiently account for heterogeneous patient characteristics, such as genetic diversity, potentially reducing the chance of selecting the optimal treatment for individuals. Recent advances in Bayesian nonparametric modeling pave the way for innovative trial designs that not only maintain robust power but also offer the flexibility to identify subgroups deriving greater benefits from specific treatments. Building on this inspiration, we introduce a Bayesian adaptive design for multi-arm trials focusing on time-to-event endpoints. We introduce a covariate-adjusted response adaptive randomization, updating treatment allocation probabilities grounded on causal effect estimates using a random intercept accelerated failure time BART model. After the trial concludes, we suggest employing a multi-response decision tree to pinpoint subgroups with varying treatment impacts. The performance of our design is then assessed via comprehensive simulations.

A genetically modulated Toll-like-receptor-tolerant phenotype in peripheral blood cells of children with multisystem inflammatory syndrome.

Dysregulated innate immune responses contribute to multisystem inflammatory syndrome in children (MIS-C), characterized by gastrointestinal, mucocutaneous, and/or cardiovascular injury occurring weeks after SARS-CoV-2 exposure. To investigate innate immune functions in MIS-C, we stimulated ex vivo peripheral blood cells from MIS-C patients with agonists of Toll-like receptors (TLR), key innate immune response initiators. We found severely dampened cytokine responses and elevated gene expression of negative regulators of TLR signaling. Increased plasma levels of zonulin, a gut leakage marker, were also detected. These effects were also observed in children enrolled months after MIS-C recovery. Moreover, cells from MIS-C children carrying rare genetic variants of lysosomal trafficking regulator (LYST) were less refractory to TLR stimulation and exhibited lysosomal and mitochondrial abnormalities with altered energy metabolism. Our results strongly suggest that MIS-C hyperinflammation and/or excessive or prolonged stimulation with gut-originated TLR ligands drive immune cells to a lasting refractory state. TLR hyporesponsiveness is likely beneficial, as suggested by excess lymphopenia among rare LYST variant carriers. Our findings point to cellular mechanisms underlying TLR hyporesponsiveness; identify genetic determinants that may explain the MIS-C clinical spectrum; suggest potential associations between innate refractory states and long COVID; and highlight the need to monitor long-term consequences of MIS-C.

A Bayesian nonparametric approach for multiple mediators with applications in mental health studies.

Mediation analysis with contemporaneously observed multiple mediators is a significant area of causal inference. Recent approaches for multiple mediators are often based on parametric models and thus may suffer from model misspecification. Also, much of the existing literature either only allow estimation of the joint mediation effect or estimate the joint mediation effect just as the sum of individual mediator effects, ignoring the interaction among the mediators. In this article, we propose a novel Bayesian nonparametric method that overcomes the two aforementioned drawbacks. We model the joint distribution of the observed data (outcome, mediators, treatment, and confounders) flexibly using an enriched Dirichlet process mixture with three levels. We use standardization (g-computation) to compute all possible mediation effects, including pairwise and all other possible interaction among the mediators. We thoroughly explore our method via simulations and apply our method to a mental health data from Wisconsin Longitudinal Study, where we estimate how the effect of births from unintended pregnancies on later life mental depression (CES-D) among the mothers is mediated through lack of self-acceptance and autonomy, employment instability, lack of social participation, and increased family stress. Our method identified significant individual mediators, along with some significant pairwise effects.

Place of care in the last three years of life for Medicare beneficiaries.

BACKGROUND: Most older adults prefer aging in place; however, patients with advanced illness often need institutional care. Understanding place of care trajectory patterns may inform patient-centered care planning and health policy decisions. The purpose of this study was to characterize place of care trajectories during the last three years of life. METHODS: Linked administrative, claims, and assessment data were analyzed for a 10% random sample cohort of US Medicare beneficiaries who died in 2018, aged fifty or older, and continuously enrolled in Medicare during their last five years of life. A group-based trajectory modeling approach was used to classify beneficiaries based on the proportion of days of institutional care (hospital inpatient or skilled nursing facility) and skilled home care (home health care and home hospice) used in each quarter of the last three years of life. Associations between group membership and sociodemographic and clinical predictors were evaluated. RESULTS: The analytic cohort included 199,828 Medicare beneficiaries. Nine place of care trajectory groups were identified, which were categorized into three clusters: home, skilled home care, and institutional care. Over half (59%) of the beneficiaries were in the home cluster, spending their last three years mostly at home, with skilled home care and institutional care use concentrated in the final quarter of life. One-quarter (27%) of beneficiaries were in the skilled home care cluster, with heavy use of skilled home health care and home hospice; the remaining 14% were in the institutional cluster, with heavy use of nursing home and inpatient care. Factors associated with both the skilled home care and institutional care clusters were female sex, Black race, a diagnosis of dementia, and Medicaid insurance. Extended use of skilled home care was more prevalent in southern states, and extended institutional care was more prevalent in midwestern states. CONCLUSIONS: This study identified distinct patterns of place of care trajectories that varied in the timing and duration of institutional and skilled home care use during the last three years of life. Clinical, socioregional, and health policy factors influenced where patients received care. Our findings can help to inform personal and societal care planning.

Bayesian semiparametric model for sequential treatment decisions with informative timing.

We develop a Bayesian semiparametric model for the impact of dynamic treatment rules on survival among patients diagnosed with pediatric acute myeloid leukemia (AML). The data consist of a subset of patients enrolled in a phase III clinical trial in which patients move through a sequence of four treatment courses. At each course, they undergo treatment that may or may not include anthracyclines (ACT). While ACT is known to be effective at treating AML, it is also cardiotoxic and can lead to early death for some patients. Our task is to estimate the potential survival probability under hypothetical dynamic ACT treatment strategies, but there are several impediments. First, since ACT is not randomized, its effect on survival is confounded over time. Second, subjects initiate the next course depending on when they recover from the previous course, making timing potentially informative of subsequent treatment and survival. Third, patients may die or drop out before ever completing the full treatment sequence. We develop a generative Bayesian semiparametric model based on Gamma Process priors to address these complexities. At each treatment course, the model captures subjects' transition to subsequent treatment or death in continuous time. G-computation is used to compute a posterior over potential survival probability that is adjusted for time-varying confounding. Using our approach, we estimate the efficacy of hypothetical treatment rules that dynamically modify ACT based on evolving cardiac function.

Marshall Joffe's Contributions to Causal Inference, Biostatistics, and Epidemiology.

We pay tribute to Marshall Joffe, PhD, and his substantial contributions to the field of causal inference with focus in biostatistics and epidemiology. By compiling narratives written by us, his colleagues, we not only present highlights of Marshall's research and their significance for causal inference but also offer a portrayal of Marshall's personal accomplishments and character. Our discussion of Marshall's research notably includes (but is not limited to) handling of posttreatment variables such as noncompliance, employing G-estimation for treatment effects on failure-time outcomes, estimating effects of time-varying exposures subject to time-dependent confounding, and developing a causal framework for case-control studies. We also provide a description of some of Marshall's unpublished work, which is accompanied by a bonus anecdote. We discuss future research directions related to Marshall's research. While Marshall's impact in causal inference and the world outside of it cannot be wholly captured by our words, we hope nonetheless to present some of what he has done for our field and what he has meant to us and to his loved ones.

Ten-Year Trends of Persistent Mortality With Gallstone Disease: A Retrospective Cohort Study in New Jersey.

BACKGROUND AND AIMS: Recent trends in mortality with gallstone disease remain scarce in the United States. Yet multiple changes in clinical management, such as rates of endoscopy, cholecystectomy, and cholecystostomy, and insurance access at the state level, may have occurred. Thus, we evaluated recent secular trends of mortality with gallstone disease in New Jersey. METHODS: We performed a retrospective, cohort study of mortality from 2009 to 2018 using the National Center for Health Statistics, Restricted Mortality Files. The primary outcome was any death with an International Classifications of Disease, 10th Revision, Clinical Modification diagnosis code of gallstone disease in New Jersey. Simple linear regression was used to model trends of incidence of death. RESULTS: 1580 deaths with diagnosed gallstone disease (dGD) occurred from 2009 to 2018. The annual trend of incidence of death was flat over 10 years. The incidence of death with dGD relative to all death changed only from 0.21% to 0.20% over 10 years. These findings were consistent also in 18 of 20 subgroup combinations, although the trend of death with dGD in Latinos 65 years or older increased [slope estimate 0.93, 95% confidence limit 0.42-1.43, P = .003]. CONCLUSION: The rate of death with dGD showed little change over the recent 10 years in New Jersey. This needs to be reproduced in other states and nationally. A closer examination of the changes in clinical care and insurance access is needed to help understand why they did not result in a positive change in this avoidable cause of death.

Saliva microbiome in relation to SARS-CoV-2 infection in a prospective cohort of healthy US adults.

BACKGROUND: The clinical outcomes of SARS-CoV-2 infection vary in severity, potentially influenced by the resident human microbiota. There is limited consensus on conserved microbiome changes in response to SARS-CoV-2 infection, with many studies focusing on severely ill individuals. This study aimed to assess the variation in the upper respiratory tract microbiome using saliva specimens in a cohort of individuals with primarily mild to moderate disease. METHODS: In early 2020, a cohort of 831 adults without known SARS-CoV-2 infection was followed over a six-month period to assess the occurrence and natural history of SARS-CoV-2 infection. From this cohort, 81 participants with a SARS-CoV-2 infection, along with 57 unexposed counterparts were selected with a total of 748 serial saliva samples were collected for analysis. Total bacterial abundance, composition, population structure, and gene function of the salivary microbiome were measured using 16S rRNA gene and shotgun metagenomic sequencing. FINDINGS: The salivary microbiome remained stable in unexposed individuals over the six-month study period, as evidenced by all measured metrics. Similarly, participants with mild to moderate SARS-CoV-2 infection showed microbiome stability throughout and after their infection. However, there were significant reductions in microbiome diversity among SARS-CoV-2-positive participants with severe symptoms early after infection. Over time, the microbiome diversity in these participants showed signs of recovery. INTERPRETATION: These findings demonstrate the resilience of the salivary microbiome in relation to SARS-CoV-2 infection. Mild to moderate infections did not significantly disrupt the stability of the salivary microbiome, suggesting its ability to maintain its composition and function. However, severe SARS-CoV-2 infection was associated with temporary reductions in microbiome diversity, indicating the limits of microbiome resilience in the face of severe infection. FUNDING: This project was supported in part by Danone North America and grants from the National Institutes of Health, United States.

Pneumonitis in advanced non-small cell lung cancer: no interaction between immune checkpoint inhibition and radiation therapy.

Background: Radiation pneumonitis and immune-related pneumonitis have been studied independently, but little information has emerged on the interactions between radiation therapy (RT) and immune checkpoint inhibition (ICI). We examine whether RT and ICI are synergistic in causing pneumonitis. Methods: A retrospective cohort was assembled using the Surveillance, Epidemiology, and End Results-Medicare database, including Medicare beneficiaries diagnosed with American Joint Committee on Cancer 7th ed. (AJCC) stages IIIB-IV NSCLC between 2013-2017. Exposures to RT and ICI were determined by evaluating for treatment within 12 months of diagnosis (RT group and ICI group) and for a second exposure (e.g., ICI after RT) within 3 months after the first exposure (RT + ICI group). Untreated controls were matched to treated patients who were diagnosed in the same three-month window. A validated algorithm for identifying cases of pneumonitis in claims data was used to evaluate for the outcome within 6 months after treatment. The primary outcome was the relative excess risk due to interaction (RERI), a quantitative measure of additive interaction between two treatments. Results: There were 18,780 patients included in the analysis with 9,345 (49.8%), 7,533 (40.2%), 1,332 (7.1%), and 550 (2.9%) in the control, RT, ICI, and RT + ICI groups, respectively. Relative to controls, the hazards ratios of pneumonitis were 11.5 (95% CI: 7.9 to 17.0), 6.2 (95% CI: 3.8 to 10.3), and 10.7 (95% CI: 6.0 to 19.2) in the RT, ICI, and RT-ICI groups, respectively. The RERIs were -6.1 (95% CI: -13.1 to -0.6, P=0.97) and -4.0 (95% CI: -10.7 to 1.5, P=0.91) in the unadjusted and adjusted analyses, respectively, consistent with no evidence of additive interaction (RERI ≤0) between RT and ICI. Conclusions: In this study of Medicare beneficiaries with advanced NSCLC, RT and ICI were, at most, additive rather than synergistic in causing pneumonitis. Pneumonitis risk in patients treated with RT and ICI is not more than could be expected from each therapy alone.

Decreased Emergency Cholecystectomy and Case Fatality Rate, Not Explained by Expansion of Medicaid.

INTRODUCTION: Population data on longitudinal trends for cholecystectomies and their outcomes are scarce. We evaluated the incidence and case fatality rate of emergency and ambulatory cholecystectomies in New Jersey (NJ) and whether the Medicaid expansion changed trends. MATERIALS AND METHODS: A retrospective population cohort design was used to study the incidence of cholecystectomies and their case fatality rate from 2009 to 2018. Using linear and logistic regression we explored the trends of incidence and the odds of case fatality after versus before the January 1, 2014 Medicaid expansion. RESULTS: Overall, 93,423 emergency cholecystectomies were performed, with 644 fatalities; 87,239 ambulatory cholecystectomies were performed, with fewer than 10 fatalities. The 2009 to 2018 annual incidence of emergency cholecystectomies dropped markedly from 114.8 to 77.5 per 100,000 NJ population (P < 0.0001); ambulatory cholecystectomies increased from 93.5 to 95.6 per 100,000 (P = 0.053). The incidence of emergency cholecystectomies dropped more after than before Medicaid expansion (P < 0.0001). The odds ratio for case fatality among those undergoing emergency cholecystectomies after versus before expansion was 0.85 (95% CI, 0.72-0.99). This decrease in case fatality, apparent only in those over age 65, was not explained by the addition of Medicaid. CONCLUSIONS: A marked decrease in the incidence of emergency cholecystectomies occurred after Medicaid expansion, which was not accounted for by a minimal increase in the incidence of ambulatory cholecystectomies. Case fatality from emergency cholecystectomy decreased over time due to factors other than Medicaid. Further work is needed to reconcile these findings with the previously reported lack of decrease in overall gallstone disease mortality in NJ.

Rheumatoid Arthritis Disease Activity and Hospitalized Infection in a Large US Registry.

OBJECTIVES: The association between disease activity and infection risk among patients with rheumatoid arthritis (RA) is not clear, and it is challenging to determine because of confounding due to the effects of RA treatments and comorbidities. METHODS: Using patients with RA in the CorEvitas registry with Medicare coverage in 2006-2019, we identified eligible patients who had at least 1 visit with moderate disease activity based on the Clinical Disease Activity Index (CDAI; CDAI >10 and ≤22). Follow-up started at the subsequent CorEvitas visit. Hospitalized infection during follow-up was assessed in linked Medicare data. We calculated the incidence rate of hospitalized infection for patients in remission, and low and moderate disease activity, and estimated the effect of time-varying CDAI on hospitalized infection by controlling for baseline and time-dependent confounders using marginal structural models (MSMs). RESULTS: A total of 3,254 patients with RA were eligible for analysis, among which 529 hospitalized infections were identified during follow-up. The crude incidence of hospitalized infection was 3.8 per 100 person-years for patients in remission, 6.6 for low disease activity, and 8.0 for moderate disease activity. Using MSMs and compared with being in remission, the hazard ratio of hospitalized infection associated with low disease activity was 1.60 (95% confidence interval [95% CI] 1.13-2.28) and with moderate disease activity was 1.83 (95% CI 1.30-2.64). CONCLUSION: The risk of hospitalized infection was higher for patients with RA in low or moderate disease activity than for those in remission after accounting for the interplay of disease activity, RA treatments, treatment switching, and other potential confounders.

Addressing positivity violations in causal effect estimation using Gaussian process priors.

In observational studies, causal inference relies on several key identifying assumptions. One identifiability condition is the positivity assumption, which requires the probability of treatment be bounded away from 0 and 1. That is, for every covariate combination, it should be possible to observe both treated and control subjects the covariate distributions should overlap between treatment arms. If the positivity assumption is violated, population-level causal inference necessarily involves some extrapolation. Ideally, a greater amount of uncertainty about the causal effect estimate should be reflected in such situations. With that goal in mind, we construct a Gaussian process model for estimating treatment effects in the presence of practical violations of positivity. Advantages of our method include minimal distributional assumptions, a cohesive model for estimating treatment effects, and more uncertainty associated with areas in the covariate space where there is less overlap. We assess the performance of our approach with respect to bias and efficiency using simulation studies. The method is then applied to a study of critically ill female patients to examine the effect of undergoing right heart catheterization.

Hierarchical Bayesian bootstrap for heterogeneous treatment effect estimation.

A major focus of causal inference is the estimation of heterogeneous average treatment effects (HTE) - average treatment effects within strata of another variable of interest such as levels of a biomarker, education, or age strata. Inference involves estimating a stratum-specific regression and integrating it over the distribution of confounders in that stratum - which itself must be estimated. Standard practice involves estimating these stratum-specific confounder distributions independently (e.g. via the empirical distribution or Rubin's Bayesian bootstrap), which becomes problematic for sparsely populated strata with few observed confounder vectors. In this paper, we develop a nonparametric hierarchical Bayesian bootstrap (HBB) prior over the stratum-specific confounder distributions for HTE estimation. The HBB partially pools the stratum-specific distributions, thereby allowing principled borrowing of confounder information across strata when sparsity is a concern. We show that posterior inference under the HBB can yield efficiency gains over standard marginalization approaches while avoiding strong parametric assumptions about the confounder distribution. We use our approach to estimate the adverse event risk of proton versus photon chemoradiotherapy across various cancer types.

Community- Versus Health Care Organization-Based Approaches to Expanding At-Home COVID-19 Testing in Black and Latino Communities, New Jersey, 2021.

At-home COVID-19 testing offers convenience and safety advantages. We evaluated at-home testing in Black and Latino communities through an intervention comparing community-based organization (CBO) and health care organization (HCO) outreach. From May through December 2021, 1100 participants were recruited, 94% through CBOs. The odds of COVID-19 test requests and completions were significantly higher in the HCO arm. The results showed disparities in test requests and completions related to age, race, language, insurance, comorbidities, and pandemic-related challenges. Despite the popularity of at-home testing, barriers exist in underresourced communities. (Am J Public Health. 2022;112(S9):S918-S922. https://doi.org/10.2105/AJPH.2022.306989).

Use of a New Prevention Model in Acute Care Surgery: A Population Approach to Preventing Emergency Surgical Morbidity and Mortality.

Whether patients undergo the more morbid and costly emergent rather than an elective type of surgery, may depend on many factors. Since tertiary prevention (preventing poor outcomes from emergency surgery) carries a much higher mortality than secondary prevention (preventing emergency surgery) or primary prevention (preventing the disease requiring surgery), the overall United States mortality might be reduced significantly, if emergency surgery could be avoided via high-quality primary prevention and non-surgical therapy or increasing elective surgery at the expense of emergency procedures, e.g., secondary prevention. The practice and study of acute care surgery then has the potential to broaden from a focus on the patient in the hospital emergency and operating rooms to the patient who no longer requires either, whose disease is treated or prevented in his/her/their community.

The Future of Causal Inference.

The past several decades have seen exponential growth in causal inference approaches and their applications. In this commentary, we provide our top-10 list of emerging and exciting areas of research in causal inference. These include methods for high-dimensional data and precision medicine, causal machine learning, causal discovery, and others. These methods are not meant to be an exhaustive list; instead, we hope that this list will serve as a springboard for stimulating the development of new research.

Dichotomy between the humoral and cellular responses elicited by mRNA and adenoviral vector vaccines against SARS-CoV-2.

BACKGROUND: Protection from severe disease and hospitalization by SARS-CoV-2 vaccination has been amply demonstrated by real-world data. However, the rapidly evolving pandemic raises new concerns. One pertains efficacy of adenoviral vector-based vaccines, particularly the single-dose Ad26.COV2.S, relative to mRNA vaccines. MAIN BODY: We investigated the immunogenicity of Ad26.COV2.S and mRNA vaccines in 33 subjects vaccinated with either vaccine class 5 months earlier on average. After controlling for the time since vaccination, Spike-binding antibody and neutralizing antibody levels were higher in the mRNA-vaccinated subjects, while no significant differences in antigen-specific B cell and T cell responses were observed between the two groups. CONCLUSIONS: A dichotomy exists between the humoral and cellular responses elicited by the two vaccine classes. Testing only for humoral responses to compare the durability of SARS-CoV-2 vaccine-induced responses, as typically performed for public health and research purposes, is insufficient.

Associations between neighborhood disinvestment and breast cancer outcomes within a populous state registry.

BACKGROUND: Breast cancer (BrCa) outcomes vary by social environmental factors, but the role of built-environment factors is understudied. The authors investigated associations between environmental physical disorder-indicators of residential disrepair and disinvestment-and BrCa tumor prognostic factors (stage at diagnosis, tumor grade, triple-negative [negative for estrogen receptor, progesterone receptor, and HER2 receptor] BrCa) and survival within a large state cancer registry linkage. METHODS: Data on sociodemographic, tumor, and vital status were derived from adult women who had invasive BrCa diagnosed from 2008 to 2017 ascertained from the New Jersey State Cancer Registry. Physical disorder was assessed through virtual neighborhood audits of 23,276 locations across New Jersey, and a personalized measure for the residential address of each woman with BrCa was estimated using universal kriging. Continuous covariates were z scored (mean ± standard deviation [SD], 0 ± 1) to reduce collinearity. Logistic regression models of tumor factors and accelerated failure time models of survival time to BrCa-specific death were built to investigate associations with physical disorder adjusted for covariates (with follow-up through 2019). RESULTS: There were 3637 BrCa-specific deaths among 40,963 women with a median follow-up of 5.3 years. In adjusted models, a 1-SD increase in physical disorder was associated with higher odds of late-stage BrCa (odds ratio, 1.09; 95% confidence interval, 1.02-1.15). Physical disorder was not associated with tumor grade or triple-negative tumors. A 1-SD increase in physical disorder was associated with a 10.5% shorter survival time (95% confidence interval, 6.1%-14.6%) only among women who had early stage BrCa. CONCLUSIONS: Physical disorder is associated with worse tumor prognostic factors and survival among women who have BrCa diagnosed at an early stage.

Respiratory Microbiome Disruption and Risk for Ventilator-Associated Lower Respiratory Tract Infection.

BACKGROUND: Ventilator-associated lower respiratory tract infection (VA-LRTI) is common among critically ill patients and has been associated with increased morbidity and mortality. In acute critical illness, respiratory microbiome disruption indices (MDIs) have been shown to predict risk for VA-LRTI, but their utility beyond the first days of critical illness is unknown. We sought to characterize how MDIs previously shown to predict VA-LRTI at initiation of mechanical ventilation change with prolonged mechanical ventilation, and if they remain associated with VA-LRTI risk. METHODS: We developed a cohort of 83 subjects admitted to a long-term acute care hospital due to their prolonged dependence on mechanical ventilation; performed dense, longitudinal sampling of the lower respiratory tract, collecting 1066 specimens; and characterized the lower respiratory microbiome by 16S rRNA sequencing as well as total bacterial abundance by 16S rRNA quantitative polymerase chain reaction. RESULTS: Cross-sectional MDIs, including low Shannon diversity and high total bacterial abundance, were associated with risk for VA-LRTI, but associations had wide posterior credible intervals. Persistent lower respiratory microbiome disruption showed a more robust association with VA-LRTI risk, with each day of (base e) Shannon diversity <2.0 associated with a VA-LRTI odds ratio of 1.36 (95% credible interval, 1.10-1.72). The observed association was consistent across multiple clinical definitions of VA-LRTI. CONCLUSIONS: Cross-sectional MDIs have limited ability to discriminate VA-LRTI risk during prolonged mechanical ventilation, but persistent lower respiratory tract microbiome disruption, best characterized by consecutive days with low Shannon diversity, may identify a population at high risk for infection and may help target infection-prevention interventions.

Participants Switching to Second-Line Antiretroviral Therapy with Susceptible Virus Display Inferior Adherence and Worse Outcomes: An Observational Analysis.

Evidence on the impact of human immunodeficiency virus (HIV) drug resistance on regimens following treatment failure is varied and inconclusive. Differential medication adherence may explain this variation. We aimed to test the association between drug resistance at first-line antiretroviral therapy (ART) switch and adherence to and virologic failure on subsequent ART. We conducted a secondary analysis of data from an open-labeled randomized trial of second-line ART (ACTG A5234). ART susceptibility was determined from study entry plasma using the Stanford Drug Resistance database version 8.7. Adherence was measured with microelectronic monitors. Three adherence variables and rates of virologic failure (HIV-1 RNA ≥1000 copies/mL) on second-line ART were compared between participants with and without resistance at first-line ART failure. Of 214 participants switching to second-line ART with baseline resistance results, 113 (53%) were men, mean age was 39 years (standard deviation 10.3), and 37 (17%) had susceptible virus at study entry. Cumulative genotypic susceptibility score (cGSS) was inversely associated with adherence, adjusted odds ratio (aOR) 0.15, 95% confidence interval (CI) (0.05-0.40), p < 0.001. The aOR of virologic failure for a one-unit increase in cGSS was 1.72, 95% CI (1.22-2.41), p < 0.001. Participants switched to second-line ART without resistance displayed inferior adherence and had higher rates of virologic failure. Therefore, these individuals warrant additional adherence interventions to help them achieve virologic success. Clinical Trial Registration number: NCT00608569.