Phase 3 Safety and Efficacy of AZD1222 (ChAdOx1 nCoV-19) Covid-19 Vaccine.

BACKGROUND: The safety and efficacy of the AZD1222 (ChAdOx1 nCoV-19) vaccine in a large, diverse population at increased risk for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection in the United States, Chile, and Peru has not been known. METHODS: In this ongoing, double-blind, randomized, placebo-controlled, phase 3 clinical trial, we investigated the safety, vaccine efficacy, and immunogenicity of two doses of AZD1222 as compared with placebo in preventing the onset of symptomatic and severe coronavirus disease 2019 (Covid-19) 15 days or more after the second dose in adults, including older adults, in the United States, Chile, and Peru. RESULTS: A total of 32,451 participants underwent randomization, in a 2:1 ratio, to receive AZD1222 (21,635 participants) or placebo (10,816 participants). AZD1222 was safe, with low incidences of serious and medically attended adverse events and adverse events of special interest; the incidences were similar to those observed in the placebo group. Solicited local and systemic reactions were generally mild or moderate in both groups. Overall estimated vaccine efficacy was 74.0% (95% confidence interval [CI], 65.3 to 80.5; P<0.001) and estimated vaccine efficacy was 83.5% (95% CI, 54.2 to 94.1) in participants 65 years of age or older. High vaccine efficacy was consistent across a range of demographic subgroups. In the fully vaccinated analysis subgroup, no severe or critical symptomatic Covid-19 cases were observed among the 17,662 participants in the AZD1222 group; 8 cases were noted among the 8550 participants in the placebo group (<0.1%). The estimated vaccine efficacy for preventing SARS-CoV-2 infection (nucleocapsid antibody seroconversion) was 64.3% (95% CI, 56.1 to 71.0; P<0.001). SARS-CoV-2 spike protein binding and neutralizing antibodies increased after the first dose and increased further when measured 28 days after the second dose. CONCLUSIONS: AZD1222 was safe and efficacious in preventing symptomatic and severe Covid-19 across diverse populations that included older adults. (Funded by AstraZeneca and others; ClinicalTrials.gov number, NCT04516746.).

Vaccine Efficacy of ALVAC-HIV and Bivalent Subtype C gp120-MF59 in Adults.

BACKGROUND: A safe, effective vaccine is essential to eradicating human immunodeficiency virus (HIV) infection. A canarypox-protein HIV vaccine regimen (ALVAC-HIV plus AIDSVAX B/E) showed modest efficacy in reducing infection in Thailand. An analogous regimen using HIV-1 subtype C virus showed potent humoral and cellular responses in a phase 1-2a trial in South Africa. Efficacy data and additional safety data were needed for this regimen in a larger population in South Africa. METHODS: In this phase 2b-3 trial, we randomly assigned 5404 adults without HIV-1 infection to receive the vaccine (2704 participants) or placebo (2700 participants). The vaccine regimen consisted of injections of ALVAC-HIV at months 0 and 1, followed by four booster injections of ALVAC-HIV plus bivalent subtype C gp120-MF59 adjuvant at months 3, 6, 12, and 18. The primary efficacy outcome was the occurrence of HIV-1 infection from randomization to 24 months. RESULTS: In January 2020, prespecified criteria for nonefficacy were met at an interim analysis; further vaccinations were subsequently halted. The median age of the trial participants was 24 years; 70% of the participants were women. The incidence of adverse events was similar in the vaccine and placebo groups. During the 24-month follow-up, HIV-1 infection was diagnosed in 138 participants in the vaccine group and in 133 in the placebo group (hazard ratio, 1.02; 95% confidence interval, 0.81 to 1.30; P = 0.84). CONCLUSIONS: The ALVAC-gp120 regimen did not prevent HIV-1 infection among participants in South Africa despite previous evidence of immunogenicity. (HVTN 702 ClinicalTrials.gov number, NCT02968849.).

Association of Treated and Untreated Chronic Hepatitis C With the Incidence of Active Tuberculosis Disease: A Population-Based Cohort Study.

BACKGROUND: Hepatitis C virus (HCV) infection causes dysregulation and suppression of immune pathways involved in the control of tuberculosis (TB) infection. However, data on the role of chronic hepatitis C as a risk factor for active TB are lacking. We sought to evaluate the association between HCV infection and the development of active TB. METHODS: We conducted a cohort study in Georgia among adults tested for HCV antibodies (January 2015-September 2020) and followed longitudinally for the development of newly diagnosed active TB. Data were obtained from the Georgian national programs of hepatitis C and TB. The exposures of interest were untreated and treated HCV infection. A Cox proportional hazards model was used to calculate adjusted hazard ratios (aHRs). RESULTS: A total of 1 828 808 adults were included (median follow-up time: 26 months; IQR: 13-39 months). Active TB was diagnosed in 3163 (0.17%) individuals after a median of 6 months follow-up (IQR: 1-18 months). The incidence rate per 100 000 person-years was 296 among persons with untreated HCV infection, 109 among those with treated HCV infection, and 65 among HCV-negative persons. In multivariable analysis, both untreated (aHR = 2.9; 95% CI: 2.4-3.4) and treated (aHR = 1.6; 95% CI: 1.4-2.0) HCV infections were associated with a higher hazard of active TB, compared with HCV-negative persons. CONCLUSIONS: Adults with HCV infection, particularly untreated individuals, were at higher risk of developing active TB disease. Screening for latent TB infection and active TB disease should be part of clinical evaluation of people with HCV infection, especially in high-TB-burden areas.

Hepatitis C care cascade among patients with and without tuberculosis: Nationwide observational cohort study in the country of Georgia, 2015-2020.

BACKGROUND: The Eastern European country of Georgia initiated a nationwide hepatitis C virus (HCV) elimination program in 2015 to address a high burden of infection. Screening for HCV infection through antibody testing was integrated into multiple existing programs, including the National Tuberculosis Program (NTP). We sought to compare the hepatitis C care cascade among patients with and without tuberculosis (TB) diagnosis in Georgia between 2015 and 2019 and to identify factors associated with loss to follow-up (LTFU) in hepatitis C care among patients with TB. METHODS AND FINDINGS: Using national ID numbers, we merged databases of the HCV elimination program, NTP, and national death registry from January 1, 2015 to September 30, 2020. The study population included 11,985 adults (aged ≥18 years) diagnosed with active TB from January 1, 2015 through December 31, 2019, and 1,849,820 adults tested for HCV antibodies between January 1, 2015 and September 30, 2020, who were not diagnosed with TB during that time. We estimated the proportion of patients with and without TB who were LTFU at each step of the HCV care cascade and explored temporal changes. Among 11,985 patients with active TB, 9,065 (76%) patients without prior hepatitis C treatment were tested for HCV antibodies, of which 1,665 (18%) had a positive result; LTFU from hepatitis C care was common, with 316 of 1,557 (20%) patients with a positive antibody test not undergoing viremia testing and 443 of 1,025 (43%) patients with viremia not starting treatment for hepatitis C. Overall, among persons with confirmed viremic HCV infection, due to LTFU at various stages of the care cascade only 28% of patients with TB had a documented cure from HCV infection, compared to 55% among patients without TB. LTFU after positive antibody testing substantially decreased in the last 3 years, from 32% among patients diagnosed with TB in 2017 to 12% among those diagnosed in 2019. After a positive HCV antibody test, patients without TB had viremia testing sooner than patients with TB (hazards ratio [HR] = 1.46, 95% confidence intervals [CI] [1.39, 1.54], p < 0.001). After a positive viremia test, patients without TB started hepatitis C treatment sooner than patients with TB (HR = 2.05, 95% CI [1.87, 2.25], p < 0.001). In the risk factor analysis adjusted for age, sex, and case definition (new versus previously treated), multidrug-resistant (MDR) TB was associated with an increased risk of LTFU after a positive HCV antibody test (adjusted risk ratio [aRR] = 1.41, 95% CI [1.12, 1.76], p = 0.003). The main limitation of this study was that due to the reliance on existing electronic databases, we were unable to account for the impact of all confounding factors in some of the analyses. CONCLUSIONS: LTFU from hepatitis C care after a positive antibody or viremia test was high and more common among patients with TB than in those without TB. Better integration of TB and hepatitis C care systems can potentially reduce LTFU and improve patient outcomes both in Georgia and other countries that are initiating or scaling up their nationwide hepatitis C control efforts and striving to provide personalized TB treatment.

Analysis of the HIV Vaccine Trials Network 702 Phase 2b-3 HIV-1 Vaccine Trial in South Africa Assessing RV144 Antibody and T-Cell Correlates of HIV-1 Acquisition Risk.

BACKGROUND: The ALVAC/gp120 + MF59 vaccines in the HIV Vaccine Trials Network (HVTN) 702 efficacy trial did not prevent human immunodeficiency virus-1 (HIV-1) acquisition. Vaccine-matched immunological endpoints that were correlates of HIV-1 acquisition risk in RV144 were measured in HVTN 702 and evaluated as correlates of HIV-1 acquisition. METHODS: Among 1893 HVTN 702 female vaccinees, 60 HIV-1-seropositive cases and 60 matched seronegative noncases were sampled. HIV-specific CD4+ T-cell and binding antibody responses were measured 2 weeks after fourth and fifth immunizations. Cox proportional hazards models assessed prespecified responses as predictors of HIV-1 acquisition. RESULTS: The HVTN 702 Env-specific CD4+ T-cell response rate was significantly higher than in RV144 (63% vs 40%, P = .03) with significantly lower IgG binding antibody response rate and magnitude to 1086.C V1V2 (67% vs 100%, P < .001; Pmag < .001). Although no significant univariate associations were observed between any T-cell or binding antibody response and HIV-1 acquisition, significant interactions were observed (multiplicity-adjusted P ≤.03). Among vaccinees with high IgG A244 V1V2 binding antibody responses, vaccine-matched CD4+ T-cell endpoints associated with decreased HIV-1 acquisition (estimated hazard ratios = 0.40-0.49 per 1-SD increase in CD4+ T-cell endpoint). CONCLUSIONS: HVTN 702 and RV144 had distinct immunogenicity profiles. However, both identified significant correlations (univariate or interaction) for IgG V1V2 and polyfunctional CD4+ T cells with HIV-1 acquisition. Clinical Trials Registration . NCT02968849.

Accounting for motion in resting-state fMRI: What part of the spectrum are we characterizing in autism spectrum disorder?

The exclusion of high-motion participants can reduce the impact of motion in functional Magnetic Resonance Imaging (fMRI) data. However, the exclusion of high-motion participants may change the distribution of clinically relevant variables in the study sample, and the resulting sample may not be representative of the population. Our goals are two-fold: 1) to document the biases introduced by common motion exclusion practices in functional connectivity research and 2) to introduce a framework to address these biases by treating excluded scans as a missing data problem. We use a study of autism spectrum disorder in children without an intellectual disability to illustrate the problem and the potential solution. We aggregated data from 545 children (8-13 years old) who participated in resting-state fMRI studies at Kennedy Krieger Institute (173 autistic and 372 typically developing) between 2007 and 2020. We found that autistic children were more likely to be excluded than typically developing children, with 28.5% and 16.1% of autistic and typically developing children excluded, respectively, using a lenient criterion and 81.0% and 60.1% with a stricter criterion. The resulting sample of autistic children with usable data tended to be older, have milder social deficits, better motor control, and higher intellectual ability than the original sample. These measures were also related to functional connectivity strength among children with usable data. This suggests that the generalizability of previous studies reporting naïve analyses (i.e., based only on participants with usable data) may be limited by the selection of older children with less severe clinical profiles because these children are better able to remain still during an rs-fMRI scan. We adapt doubly robust targeted minimum loss based estimation with an ensemble of machine learning algorithms to address these data losses and the resulting biases. The proposed approach selects more edges that differ in functional connectivity between autistic and typically developing children than the naïve approach, supporting this as a promising solution to improve the study of heterogeneous populations in which motion is common.

Estimation of the Average Causal Effect in Longitudinal Data With Time-Varying Exposures: The Challenge of Nonpositivity and the Impact of Model Flexibility.

There are important challenges to the estimation and identification of average causal effects in longitudinal data with time-varying exposures. Here, we discuss the difficulty in meeting the positivity condition. Our motivating example is the per-protocol analysis of the Effects of Aspirin in Gestation and Reproduction (EAGeR) Trial. We estimated the average causal effect comparing the incidence of pregnancy by 26 weeks that would have occurred if all women had been assigned to aspirin and complied versus the incidence if all women had been assigned to placebo and complied. Using flexible targeted minimum loss-based estimation, we estimated a risk difference of 1.27% (95% CI: -9.83, 12.38). Using a less flexible inverse probability weighting approach, the risk difference was 5.77% (95% CI: -1.13, 13.05). However, the cumulative probability of compliance conditional on covariates approached 0 as follow-up accrued, indicating a practical violation of the positivity assumption, which limited our ability to make causal interpretations. The effects of nonpositivity were more apparent when using a more flexible estimator, as indicated by the greater imprecision. When faced with nonpositivity, one can use a flexible approach and be transparent about the uncertainty, use a parametric approach and smooth over gaps in the data, or target a different estimand that will be less vulnerable to positivity violations.

Super LeArner Prediction of NAb Panels (SLAPNAP): a containerized tool for predicting combination monoclonal broadly neutralizing antibody sensitivity.

MOTIVATION: A single monoclonal broadly neutralizing antibody (bnAb) regimen was recently evaluated in two randomized trials for prevention efficacy against HIV-1 infection. Subsequent trials will evaluate combination bnAb regimens (e.g. cocktails, multi-specific antibodies), which demonstrate higher potency and breadth in vitro compared to single bnAbs. Given the large number of potential regimens, methods for down-selecting these regimens into efficacy trials are of great interest. RESULTS: We developed Super LeArner Prediction of NAb Panels (SLAPNAP), a software tool for training and evaluating machine learning models that predict in vitro neutralization sensitivity of HIV Envelope (Env) pseudoviruses to a given single or combination bnAb regimen, based on Env amino acid sequence features. SLAPNAP also provides measures of variable importance of sequence features. By predicting bnAb coverage of circulating sequences, SLAPNAP can improve ranking of bnAb regimens by their potential prevention efficacy. In addition, SLAPNAP can improve sieve analysis by defining sequence features that impact bnAb prevention efficacy. AVAILABILITYAND IMPLEMENTATION: SLAPNAP is a freely available docker image that can be downloaded from DockerHub (https://hub.docker.com/r/slapnap/slapnap). Source code and documentation are available at GitHub (https://github.com/benkeser/slapnap and https://benkeser.github.io/slapnap/).

Sexual Mixing by HIV Status and Pre-exposure Prophylaxis Use Among Men Who Have Sex With Men: Addressing Information Bias.

BACKGROUND: Population-level estimates of sexual network mixing for parameterizing prediction models of pre-exposure prophylaxis (PrEP) effectiveness are needed to inform prevention of HIV transmission among men who have sex with men (MSM). Estimates obtained by egocentric sampling are vulnerable to information bias due to incomplete respondent knowledge. METHODS: We estimated patterns of serosorting and PrEP sorting among MSM in the United States using data from a 2017-2019 egocentric sexual network study. Respondents served as proxies to report the HIV status and PrEP use of recent sexual partners. We contrasted results from a complete-case analysis (unknown HIV and PrEP excluded) versus a bias analysis with respondent-reported data stochastically reclassified to simulate unobserved self-reported data from sexual partners. RESULTS: We found strong evidence of preferential partnering across analytical approaches. The bias analysis showed concordance between sexual partners of HIV diagnosis and PrEP use statuses for MSM with diagnosed HIV (39%; 95% simulation interval: 31, 46), MSM who used PrEP (32%; 21, 37), and MSM who did not use PrEP (83%; 79, 87). The fraction of partners with diagnosed HIV was higher among MSM who used PrEP (11%; 9, 14) compared with MSM who did not use PrEP (4%; 3, 5). Comparatively, across all strata of respondents, the complete-case analysis overestimated the fractions of partners with diagnosed HIV or PrEP use. CONCLUSIONS: We found evidence consistent with HIV and PrEP sorting among MSM, which may decrease the population-level effectiveness of PrEP. Bias analyses can improve mixing estimates for parameterization of transmission models.

A Cluster-based Method to Quantify Individual Heterogeneity in Tuberculosis Transmission.

BACKGROUND: Recent evidence suggests transmission of Mycobacterium tuberculosis (Mtb) may be characterized by extreme individual heterogeneity in secondary cases (i.e., few cases account for the majority of transmission). Such heterogeneity implies outbreaks are rarer but more extensive and has profound implications in infectious disease control. However, discrete person-to-person transmission events in tuberculosis (TB) are often unobserved, precluding our ability to directly quantify individual heterogeneity in TB epidemiology. METHODS: We used a modified negative binomial branching process model to quantify the extent of individual heterogeneity using only observed transmission cluster size distribution data (i.e., the simple sum of all cases in a transmission chain) without knowledge of individual-level transmission events. The negative binomial parameter k quantifies the extent of individual heterogeneity (generally, indicates extensive heterogeneity, and as transmission becomes more homogenous). We validated the robustness of the inference procedure considering common limitations affecting cluster size data. Finally, we demonstrate the epidemiologic utility of this method by applying it to aggregate US molecular surveillance data from the US Centers for Disease Control and Prevention. RESULTS: The cluster-based method reliably inferred k using TB transmission cluster data despite a high degree of bias introduced into the model. We found that the TB transmission in the United States was characterized by a high propensity for extensive outbreaks (; 95% confidence interval = 0.09, 0.10). CONCLUSIONS: The proposed method can accurately quantify critical parameters that govern TB transmission using simple, more easily obtainable cluster data to improve our understanding of TB epidemiology.

Causal inference methods for vaccine sieve analysis with effect modification.

The protective effects of vaccines may vary depending on individual characteristics, such as age. Traditionally, such effect modification has been examined with subgroup analyses or inclusion of cross-product terms in regression frameworks. However, in many vaccine settings, effect modification may also depend on the infecting pathogen's characteristics, which are measured postrandomization. Sieve analysis examines whether such effects are present by combining pathogen genetic sequence information with individual-level data and can generate new hypotheses on the pathways whereby vaccines provide protection. In this article, we develop a causal framework for evaluating effect modification in the context of sieve analysis. Our approach can be used to assess the magnitude of sieve effects and, in particular, whether these effects are modified by individual-level characteristics. Our method accounts for difficulties occurring in real-world data analysis, such as competing risks, nonrandomized treatments, and differential dropout. Our approach also integrates modern machine learning techniques. We demonstrate the validity and efficiency of our approach in simulation studies and apply the methodology to a malaria vaccine study.

Clinical Endpoints for Evaluating Efficacy in COVID-19 Vaccine Trials.

Several vaccine candidates to protect against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection or coronavirus disease 2019 (COVID-19) have entered or will soon enter large-scale, phase 3, placebo-controlled randomized clinical trials. To facilitate harmonized evaluation and comparison of the efficacy of these vaccines, a general set of clinical endpoints is proposed, along with considerations to guide the selection of the primary endpoints on the basis of clinical and statistical reasoning. The plausibility that vaccine protection against symptomatic COVID-19 could be accompanied by a shift toward more SARS-CoV-2 infections that are asymptomatic is highlighted, as well as the potential implications of such a shift.

A Deferred-Vaccination Design to Assess Durability of COVID-19 Vaccine Effect After the Placebo Group Is Vaccinated.

Multiple candidate vaccines to prevent COVID-19 have entered large-scale phase 3 placebo-controlled randomized clinical trials, and several have demonstrated substantial short-term efficacy. At some point after demonstration of substantial efficacy, placebo recipients should be offered the efficacious vaccine from their trial, which will occur before longer-term efficacy and safety are known. The absence of a placebo group could compromise assessment of longer-term vaccine effects. However, by continuing follow-up after vaccination of the placebo group, this study shows that placebo-controlled vaccine efficacy can be mathematically derived by assuming that the benefit of vaccination over time has the same profile for the original vaccine recipients and the original placebo recipients after their vaccination. Although this derivation provides less precise estimates than would be obtained by a standard trial where the placebo group remains unvaccinated, this proposed approach allows estimation of longer-term effect, including durability of vaccine efficacy and whether the vaccine eventually becomes harmful for some. Deferred vaccination, if done open-label, may lead to riskier behavior in the unblinded original vaccine group, confounding estimates of long-term vaccine efficacy. Hence, deferred vaccination via blinded crossover, where the vaccine group receives placebo and vice versa, would be the preferred way to assess vaccine durability and potential delayed harm. Deferred vaccination allows placebo recipients timely access to the vaccine when it would no longer be proper to maintain them on placebo, yet still allows important insights about immunologic and clinical effectiveness over time.

Improving precision and power in randomized trials for COVID-19 treatments using covariate adjustment, for binary, ordinal, and time-to-event outcomes.

Time is of the essence in evaluating potential drugs and biologics for the treatment and prevention of COVID-19. There are currently 876 randomized clinical trials (phase 2 and 3) of treatments for COVID-19 registered on clinicaltrials.gov. Covariate adjustment is a statistical analysis method with potential to improve precision and reduce the required sample size for a substantial number of these trials. Though covariate adjustment is recommended by the U.S. Food and Drug Administration and the European Medicines Agency, it is underutilized, especially for the types of outcomes (binary, ordinal, and time-to-event) that are common in COVID-19 trials. To demonstrate the potential value added by covariate adjustment in this context, we simulated two-arm, randomized trials comparing a hypothetical COVID-19 treatment versus standard of care, where the primary outcome is binary, ordinal, or time-to-event. Our simulated distributions are derived from two sources: longitudinal data on over 500 patients hospitalized at Weill Cornell Medicine New York Presbyterian Hospital and a Centers for Disease Control and Prevention preliminary description of 2449 cases. In simulated trials with sample sizes ranging from 100 to 1000 participants, we found substantial precision gains from using covariate adjustment-equivalent to 4-18% reductions in the required sample size to achieve a desired power. This was the case for a variety of estimands (targets of inference). From these simulations, we conclude that covariate adjustment is a low-risk, high-reward approach to streamlining COVID-19 treatment trials. We provide an R package and practical recommendations for implementation.

Efficient nonparametric inference on the effects of stochastic interventions under two-phase sampling, with applications to vaccine efficacy trials.

The advent and subsequent widespread availability of preventive vaccines has altered the course of public health over the past century. Despite this success, effective vaccines to prevent many high-burden diseases, including human immunodeficiency virus (HIV), have been slow to develop. Vaccine development can be aided by the identification of immune response markers that serve as effective surrogates for clinically significant infection or disease endpoints. However, measuring immune response marker activity is often costly, which has motivated the usage of two-phase sampling for immune response evaluation in clinical trials of preventive vaccines. In such trials, the measurement of immunological markers is performed on a subset of trial participants, where enrollment in this second phase is potentially contingent on the observed study outcome and other participant-level information. We propose nonparametric methodology for efficiently estimating a counterfactual parameter that quantifies the impact of a given immune response marker on the subsequent probability of infection. Along the way, we fill in theoretical gaps pertaining to the asymptotic behavior of nonparametric efficient estimators in the context of two-phase sampling, including a multiple robustness property enjoyed by our estimators. Techniques for constructing confidence intervals and hypothesis tests are presented, and an open source software implementation of the methodology, the txshift R package, is introduced. We illustrate the proposed techniques using data from a recent preventive HIV vaccine efficacy trial.

Viral genetic diversity and protective efficacy of a tetravalent dengue vaccine in two phase 3 trials.

Two phase 3 placebo-controlled trials of the CYD-TDV vaccine, evaluated in children aged 2-14 y (CYD14) and 9-16 y (CYD15), demonstrated vaccine efficacy (VE) of 56.5% and 60.8%, respectively, against symptomatic virologically confirmed dengue (VCD). Sieve analyses were conducted to evaluate whether and how VE varied with amino acid sequence features of dengue viruses (DENVs). DENV premembrane/envelope amino acid sequences from VCD endpoint cases were aligned with the vaccine insert sequences, and extensions of the proportional hazards model were applied to assess variation in VE with amino acid mismatch proportion distances from vaccine strains, individual amino acid residues, and phylogenetic genotypes. In CYD14, VE against VCD of any serotype (DENV-Any) decreased significantly with increasing amino acid distance from the vaccine, whereas in CYD15, VE against DENV-Any was distance-invariant. Restricting to the common age range and amino acid distance range between the trials and accounting for differential VE by serotype, however, showed no evidence of VE variation with distance in either trial. In serotype-specific analyses, VE against DENV4 decreased significantly with increasing amino acid distance from the DENV4 vaccine insert and was significantly greater against residue-matched DENV4 at eight signature positions. These effects were restricted to 2- to 8-y-olds, potentially because greater seropositivity of older children at baseline might facilitate a broader protective immune response. The relevance of an antigenic match between vaccine strains and circulating DENVs was also supported by greater estimated VE against serotypes and genotypes for which the circulating DENVs had shorter amino acid sequence distances from the vaccine.

Prediction of VRC01 neutralization sensitivity by HIV-1 gp160 sequence features.

The broadly neutralizing antibody (bnAb) VRC01 is being evaluated for its efficacy to prevent HIV-1 infection in the Antibody Mediated Prevention (AMP) trials. A secondary objective of AMP utilizes sieve analysis to investigate how VRC01 prevention efficacy (PE) varies with HIV-1 envelope (Env) amino acid (AA) sequence features. An exhaustive analysis that tests how PE depends on every AA feature with sufficient variation would have low statistical power. To design an adequately powered primary sieve analysis for AMP, we modeled VRC01 neutralization as a function of Env AA sequence features of 611 HIV-1 gp160 pseudoviruses from the CATNAP database, with objectives: (1) to develop models that best predict the neutralization readouts; and (2) to rank AA features by their predictive importance with classification and regression methods. The dataset was split in half, and machine learning algorithms were applied to each half, each analyzed separately using cross-validation and hold-out validation. We selected Super Learner, a nonparametric ensemble-based cross-validated learning method, for advancement to the primary sieve analysis. This method predicted the dichotomous resistance outcome of whether the IC50 neutralization titer of VRC01 for a given Env pseudovirus is right-censored (indicating resistance) with an average validated AUC of 0.868 across the two hold-out datasets. Quantitative log IC50 was predicted with an average validated R2 of 0.355. Features predicting neutralization sensitivity or resistance included 26 surface-accessible residues in the VRC01 and CD4 binding footprints, the length of gp120, the length of Env, the number of cysteines in gp120, the number of cysteines in Env, and 4 potential N-linked glycosylation sites; the top features will be advanced to the primary sieve analysis. This modeling framework may also inform the study of VRC01 in the treatment of HIV-infected persons.

Robust inference on the average treatment effect using the outcome highly adaptive lasso.

Many estimators of the average effect of a treatment on an outcome require estimation of the propensity score, the outcome regression, or both. It is often beneficial to utilize flexible techniques, such as semiparametric regression or machine learning, to estimate these quantities. However, optimal estimation of these regressions does not necessarily lead to optimal estimation of the average treatment effect, particularly in settings with strong instrumental variables. A recent proposal addressed these issues via the outcome-adaptive lasso, a penalized regression technique for estimating the propensity score that seeks to minimize the impact of instrumental variables on treatment effect estimators. However, a notable limitation of this approach is that its application is restricted to parametric models. We propose a more flexible alternative that we call the outcome highly adaptive lasso. We discuss the large sample theory for this estimator and propose closed-form confidence intervals based on the proposed estimator. We show via simulation that our method offers benefits over several popular approaches.

Improved estimation of the cumulative incidence of rare outcomes.

Studying the incidence of rare events is both scientifically important and statistically challenging. When few events are observed, standard survival analysis estimators behave erratically, particularly if covariate adjustment is necessary. In these settings, it is possible to improve upon existing estimators by considering estimation in a bounded statistical model. This bounded model incorporates existing scientific knowledge about the incidence of an event in the population. Estimators that are guaranteed to agree with existing scientific knowledge on event incidence may exhibit superior behavior relative to estimators that ignore this knowledge. Focusing on the setting of competing risks, we propose estimators of cumulative incidence that are guaranteed to respect a bounded model and show that when few events are observed, the proposed estimators offer improvements over existing estimators in bias and variance. We illustrate the proposed estimators using data from a recent preventive HIV vaccine efficacy trial. Copyright © 2017 John Wiley & Sons, Ltd.

Online cross-validation-based ensemble learning.

Online estimators update a current estimate with a new incoming batch of data without having to revisit past data thereby providing streaming estimates that are scalable to big data. We develop flexible, ensemble-based online estimators of an infinite-dimensional target parameter, such as a regression function, in the setting where data are generated sequentially by a common conditional data distribution given summary measures of the past. This setting encompasses a wide range of time-series models and, as special case, models for independent and identically distributed data. Our estimator considers a large library of candidate online estimators and uses online cross-validation to identify the algorithm with the best performance. We show that by basing estimates on the cross-validation-selected algorithm, we are asymptotically guaranteed to perform as well as the true, unknown best-performing algorithm. We provide extensions of this approach including online estimation of the optimal ensemble of candidate online estimators. We illustrate excellent performance of our methods using simulations and a real data example where we make streaming predictions of infectious disease incidence using data from a large database. Copyright © 2017 John Wiley & Sons, Ltd.