Practical Considerations for Sandwich Variance Estimation in 2-Stage Regression Settings

In this paper, we present a practical approach for computing the sandwich variance estimator in 2-stage regression model settings. As a motivating example for 2-stage regression, we consider regression calibration, a popular approach for addressing covariate measurement error. The sandwich variance approach has rarely been applied in regression calibration, despite its requiring less computation time than popular resampling approaches for variance estimation, specifically the bootstrap. This is probably because it requires specialized statistical coding. Here we first outline the steps needed to compute the sandwich variance estimator. We then develop a convenient method of computation in R for sandwich variance estimation, which leverages standard regression model outputs and existing R functions and can be applied in the case of a simple random sample or complex survey design. We use a simulation study to compare the sandwich estimator to a resampling variance approach for both settings. Finally, we further compare these 2 variance estimation approaches in data examples from the Women's Health Initiative (1993-2005) and the Hispanic Community Health Study/Study of Latinos (2008-2011). In our simulations, the sandwich variance estimator typically had good numerical performance, but simple Wald bootstrap confidence intervals were unstable or overcovered in certain settings, particularly when there was high correlation between covariates or large measurement error.

A Randomized Trial of Choice Architecture and Mailed Colorectal Cancer Screening Outreach in a Community Health Setting

BACKGROUND & AIMS

Mailed outreach for colorectal cancer (CRC) screening increases uptake but it is unclear how to offer the choice of testing. We evaluated if the active choice between colonoscopy and fecal immunochemical test (FIT), or FIT alone, increased response compared with colonoscopy alone.

METHODS

This pragmatic, randomized, controlled trial at a community health center included patients between ages 50 and 74 who were not up to date with CRC screening. Patients were randomized 1:1:1 to the following: (1) colonoscopy only, (2) active choice of colonoscopy or FIT, or (3) FIT only. Patients received an outreach letter with instructions for testing (colonoscopy referral and/or an enclosed FIT kit), a reminder letter at 2 months, and another reminder at 3 to 5 months via text message or automated voice recording. The primary outcome was CRC screening completion within 6 months.

RESULTS

Among 738 patients in the final analysis, the mean age was 58.7 years (SD, 6.2 y); 48.6% were insured by Medicaid and 24.3% were insured by Medicare; and 71.7% were White, 16.9% were Black, and 7.3% were Hispanic/Latino. At 6 months, 5.6% (95% CI, 2.8-8.5) completed screening in the colonoscopy-only arm, 12.8% (95% CI, 8.6-17.0) in the active-choice arm, and 11.3% (95% CI, 7.4-15.3) in the FIT-only arm. Compared with colonoscopy only, there was a significant increase in screening in active choice (absolute difference, 7.1%; 95% CI, 2.0-12.2; P = .006) and FIT only (absolute difference, 5.7%; 95% CI, 0.8-10.6; P = .02).

CONCLUSIONS

Both choice of testing and FIT alone increased response and may align with patient preferences.

TRIAL REGISTRATION

clinicaltrials.gov NCT04711473.

smdi: an R package to perform structural missing data investigations on partially observed confounders in real-world evidence studies

Objectives

Partially observed confounder data pose a major challenge in statistical analyses aimed to inform causal inference using electronic health records (EHRs). While analytic approaches such as imputation are available, assumptions on underlying missingness patterns and mechanisms must be verified. We aimed to develop a toolkit to streamline missing data diagnostics to guide choice of analytic approaches based on meeting necessary assumptions.

Materials and methods

We developed the smdi (structural missing data investigations) R package based on results of a previous simulation study which considered structural assumptions of common missing data mechanisms in EHR.

Results

smdi enables users to run principled missing data investigations on partially observed confounders and implement functions to visualize, describe, and infer potential missingness patterns and mechanisms based on observed data.

Conclusions

The smdi R package is freely available on CRAN and can provide valuable insights into underlying missingness patterns and mechanisms and thereby help improve the robustness of real-world evidence studies.

Testing the missing at random assumption in generalized linear models in the presence of instrumental variables

Practical problems with missing data are common, and many methods have been developed concerning the validity and/or efficiency of statistical procedures. On a central focus, there have been longstanding interests on the mechanism governing data missingness, and correctly deciding the appropriate mechanism is crucially relevant for conducting proper practical investigations. In this paper, we present a new hypothesis testing approach for deciding between the conventional notions of missing at random and missing not at random in generalized linear models in the presence of instrumental variables. The foundational idea is to develop appropriate discrepancy measures between estimators whose properties significantly differ only when missing at random does not hold. We show that our testing approach achieves an objective data-oriented choice between missing at random or not. We demonstrate the feasibility, validity, and efficacy of the new test by theoretical analysis, simulation studies, and a real data analysis.

Associations Between Gestational Weight Gain, Gestational Diabetes, and Childhood Obesity Incidence

INTRODUCTION

Excessive maternal gestational weight gain (GWG) is strongly correlated with childhood obesity, yet how excess maternal weight gain and gestational diabetes mellitus (GDM) interact to affect early childhood obesity is poorly understood. The purpose of this study was to investigate whether overall and trimester-specific maternal GWG and GDM were associated with obesity in offspring by age 6 years.

METHODS

A cohort of 10,335 maternal-child dyads was established from electronic health records. Maternal weights at conception and delivery were estimated from weight trajectory fits using functional principal components analysis. Kaplan-Meier curves and Cox regression, together with generalized raking, examined time-to-childhood-obesity.

RESULTS

Obesity diagnosed prior to age 6 years was estimated at 19.7% (95% CI: 18.3, 21.1). Maternal weight gain during pregnancy was a strong predictor of early childhood obesity (p < 0.0001). The occurrence of early childhood obesity was lower among mothers with GDM compared with those without diabetes (adjusted hazard ratio = 0.58, p = 0.014). There was no interaction between maternal weight gain and GDM (p = 0.55). Higher weight gain during the first trimester was associated with lower risk of early childhood obesity (p = 0.0002) whereas higher weight gain during the second and third trimesters was associated with higher risk (p < 0.0001).

DISCUSSION

Results indicated total and trimester-specific maternal weight gain was a strong predictor of early childhood obesity, though obesity risk by age 6 was lower for children of mothers with GDM. Additional research is needed to elucidate underlying mechanisms directly related to trimester-specific weight gain and GDM that impede or protect against obesity prevalence during early childhood.

Three-phase generalized raking and multiple imputation estimators to address error-prone data

Validation studies are often used to obtain more reliable information in settings with error-prone data. Validated data on a subsample of subjects can be used together with error-prone data on all subjects to improve estimation. In practice, more than one round of data validation may be required, and direct application of standard approaches for combining validation data into analyses may lead to inefficient estimators since the information available from intermediate validation steps is only partially considered or even completely ignored. In this paper, we present two novel extensions of multiple imputation and generalized raking estimators that make full use of all available data. We show through simulations that incorporating information from intermediate steps can lead to substantial gains in efficiency. This work is motivated by and illustrated in a study of contraceptive effectiveness among 83 671 women living with HIV, whose data were originally extracted from electronic medical records, of whom 4732 had their charts reviewed, and a subsequent 1210 also had a telephone interview to validate key study variables.

Effects of Aircraft Noise on Sleep: Federal Aviation Administration National Sleep Study Protocol

Aircraft noise can disrupt sleep and impair recuperation. The last U.S. investigation into the effects of aircraft noise on sleep dates back more than 20 years. Since then, traffic patterns and the noise levels produced by single aircraft have changed substantially. It is therefore important to acquire current data on sleep disturbance relative to varying degrees of aircraft noise exposure in the U.S. that can be used to check and potentially update the existing noise policy. This manuscript describes the design, procedures, and analytical approaches of the FAA's National Sleep Study. Seventy-seven U.S. airports with relevant nighttime air traffic from 39 states are included in the sampling frame. Based on simulation-based power calculations, the field study aims to recruit 400 participants from four noise strata and record an electrocardiogram (ECG), body movement, and sound pressure levels in the bedroom for five consecutive nights. The primary outcome of the study is an exposure-response function between the instantaneous, maximum A-weighted sound pressure levels (dBA) of individual aircraft measured in the bedroom and awakening probability inferred from changes in heart rate and body movement. Self-reported sleep disturbance due to aircraft noise is the secondary outcome that will be associated with long-term average noise exposure metrics such as the Day-Night Average Sound Level (DNL) and the Nighttime Equivalent Sound Level (Lnight). The effect of aircraft noise on several other physiological and self-report outcomes will also be investigated. This study will provide key insights into the effects of aircraft noise on objectively and subjectively assessed sleep disturbance.

Quantifying and correcting bias due to outcome dependent self-reported weights in longitudinal study of weight loss interventions

In response to the escalating global obesity crisis and its associated health and financial burdens, this paper presents a novel methodology for analyzing longitudinal weight loss data and assessing the effectiveness of financial incentives. Drawing from the Keep It Off trial-a three-arm randomized controlled study with 189 participants-we examined the potential impact of financial incentives on weight loss maintenance. Given that some participants choose not to weigh themselves because of small weight change or weight gains, which is a common phenomenon in many weight-loss studies, traditional methods, for example, the Generalized Estimating Equations (GEE) method tends to overestimate the effect size due to the assumption that data are missing completely at random. To address this challenge, we proposed a framework which can identify evidence of missing not at random and conduct bias correction using the estimating equation derived from pairwise composite likelihood. By analyzing the Keep It Off data, we found that the data in this trial are most likely characterized by non-random missingness. Notably, we also found that the enrollment time (i.e., duration time) would be positively associated with the weight loss maintenance after adjusting for the baseline participant characteristics (e.g., age, sex). Moreover, the lottery-based intervention was found to be more effective in weight loss maintenance compared with the direct payment intervention, though the difference was non-statistically significant. This framework's significance extends beyond weight loss research, offering a semi-parametric approach to assess missing data mechanisms and robustly explore associations between exposures (e.g., financial incentives) and key outcomes (e.g., weight loss maintenance). In essence, the proposed methodology provides a powerful toolkit for analyzing real-world longitudinal data, particularly in scenarios with data missing not at random, enriching comprehension of intricate dataset dynamics.

Multiwave validation sampling for error-prone electronic health records

Electronic health record (EHR) data are increasingly used for biomedical research, but these data have recognized data quality challenges. Data validation is necessary to use EHR data with confidence, but limited resources typically make complete data validation impossible. Using EHR data, we illustrate prospective, multiwave, two-phase validation sampling to estimate the association between maternal weight gain during pregnancy and the risks of her child developing obesity or asthma. The optimal validation sampling design depends on the unknown efficient influence functions of regression coefficients of interest. In the first wave of our multiwave validation design, we estimate the influence function using the unvalidated (phase 1) data to determine our validation sample; then in subsequent waves, we re-estimate the influence function using validated (phase 2) data and update our sampling. For efficiency, estimation combines obesity and asthma sampling frames while calibrating sampling weights using generalized raking. We validated 996 of 10,335 mother-child EHR dyads in six sampling waves. Estimated associations between childhood obesity/asthma and maternal weight gain, as well as other covariates, are compared to naïve estimates that only use unvalidated data. In some cases, estimates markedly differ, underscoring the importance of efficient validation sampling to obtain accurate estimates incorporating validated data.

A Randomized Controlled Pilot Study of Just4Us, a Counseling and Navigation Intervention to Promote Oral HIV Prophylaxis Uptake Among PrEP-Eligible Cisgender Women

HIV pre-exposure prophylaxis (PrEP) uptake among cisgender women in the United States is low. Just4Us, a theory-based counseling and navigation intervention, was evaluated in a pilot randomized controlled trial among PrEP-eligible women (n = 83). The comparison arm was a brief information session. Women completed surveys at baseline, post-intervention, and at three months. In this sample, 79% were Black, and 26% were Latina. This report presents results on preliminary efficacy. At 3 months follow-up, 45% made an appointment to see a provider about PrEP; only 13% received a PrEP prescription. There were no differences in PrEP initiation by study arm (9% Info vs. 11% Just4Us). PrEP knowledge was significantly higher in the Just4Us group at post-intervention. Analysis revealed high PrEP interest with many personal and structural barriers along the PrEP continuum. Just4Us is a promising PrEP uptake intervention for cisgender women. Further research is needed to tailor intervention strategies to multilevel barriers.Clinicaltrials.gov registration NCT03699722: A Women-Focused PrEP Intervention (Just4Us).

Nutritional Blood Concentration Biomarkers in the Hispanic Community Health Study/Study of Latinos: Measurement Characteristics and Power

Measurement error is a major issue in self-reported diet that can distort diet-disease relationships. Use of blood concentration biomarkers has the potential to mitigate the subjective bias inherent in self-reporting. As part of the Hispanic Community Health Study/Study of Latinos (HCHS/SOL) baseline visit (2008-2011), self-reported information on diet was collected from all participants (n = 16,415). The HCHS/SOL also included annual telephone follow-up, as well as a second (2014-2017) and third (2020-2023) clinic visit. Blood concentration biomarkers for carotenoids, tocopherols, retinol, vitamin B12, and folate were measured in a subset of participants (n = 476) as part of the Study of Latinos: Nutrition and Physical Activity Assessment Study (SOLNAS) (2010-2012). We examined the relationships among biomarker levels, self-reported intake, Hispanic/Latino background (Central American, Cuban, Dominican, Mexican, Puerto Rican, or South American), and other participant characteristics in this diverse cohort. We built regression calibration-based prediction equations for 10 nutritional biomarkers and used a simulation to study the power of detecting a diet-disease association in a multivariable Cox model using a predicted concentration level. Good statistical power was observed for some nutrients with high prediction model R2 values, but further research is needed to understand how best to realize the potential of these dietary biomarkers. This study provides a comprehensive examination of several nutritional biomarkers within the HCHS/SOL, characterizing their associations with subject characteristics and the influence of the measurement characteristics on the power to detect associations with health outcomes.

Issues in Implementing Regression Calibration Analyses

Regression calibration is a popular approach for correcting biases in estimated regression parameters when exposure variables are measured with error. This approach involves building a calibration equation to estimate the value of the unknown true exposure given the error-prone measurement and other covariates. The estimated, or calibrated, exposure is then substituted for the unknown true exposure in the health outcome regression model. When used properly, regression calibration can greatly reduce the bias induced by exposure measurement error. Here, we first provide an overview of the statistical framework for regression calibration, specifically discussing how a special type of error, called Berkson error, arises in the estimated exposure. We then present practical issues to consider when applying regression calibration, including: 1) how to develop the calibration equation and which covariates to include; 2) valid ways to calculate standard errors of estimated regression coefficients; and 3) problems arising if one of the covariates in the calibration model is a mediator of the relationship between the exposure and outcome. Throughout, we provide illustrative examples using data from the Hispanic Community Health Study/Study of Latinos (United States, 2008-2011) and simulations. We conclude with recommendations for how to perform regression calibration.

An augmented likelihood approach for the Cox proportional hazards model with interval-censored auxiliary and validated outcome data-with application to the Hispanic Community Health Study/Study of Latinos

In large epidemiologic studies, it is typical for an inexpensive, non-invasive procedure to be used to record disease status during regular follow-up visits, with less frequent assessment by a gold standard test. Inexpensive outcome measures like self-reported disease status are practical to obtain, but can be error-prone. Association analysis reliant on error-prone outcomes may lead to biased results; however, restricting analyses to only data from the less frequently observed error-free outcome could be inefficient. We have developed an augmented likelihood that incorporates data from both error-prone outcomes and a gold standard assessment. We conduct a numerical study to show how we can improve statistical efficiency by using the proposed method over standard approaches for interval-censored survival data that do not leverage auxiliary data. We extend this method for the complex survey design setting so that it can be applied in our motivating data example. Our method is applied to data from the Hispanic Community Health Study/Study of Latinos to assess the association between energy and protein intake and the risk of incident diabetes. In our application, we demonstrate how our method can be used in combination with regression calibration to additionally address the covariate measurement error in self-reported diet.

Analysis of the 24-h activity cycle: An illustration examining the association with cognitive function in the Adult Changes in Thought study

The 24-h activity cycle (24HAC) is a new paradigm for studying activity behaviors in relation to health outcomes. This approach inherently captures the interrelatedness of the daily time spent in physical activity (PA), sedentary behavior (SB), and sleep. We describe three popular approaches for modeling outcome associations with the 24HAC exposure. We apply these approaches to assess an association with a cognitive outcome in a cohort of older adults, discuss statistical challenges, and provide guidance on interpretation and selecting an appropriate approach. We compare the use of the isotemporal substitution model (ISM), compositional data analysis (CoDA), and latent profile analysis (LPA) to analyze 24HAC. We illustrate each method by exploring cross-sectional associations with cognition in 1,034 older adults (Mean age = 77; Age range = 65-100; 55.8% female; 90% White) who were part of the Adult Changes in Thought (ACT) Activity Monitoring (ACT-AM) sub-study. PA and SB were assessed with thigh-worn activPAL accelerometers for 7-days. For each method, we fit a multivariable regression model to examine the cross-sectional association between the 24HAC and Cognitive Abilities Screening Instrument item response theory (CASI-IRT) score, adjusting for baseline characteristics. We highlight differences in assumptions and the scientific questions addressable by each approach. ISM is easiest to apply and interpret; however, the typical ISM assumes a linear association. CoDA uses an isometric log-ratio transformation to directly model the compositional exposure but can be more challenging to apply and interpret. LPA can serve as an exploratory analysis tool to classify individuals into groups with similar time-use patterns. Inference on associations of latent profiles with health outcomes need to account for the uncertainty of the LPA classifications, which is often ignored. Analyses using the three methods did not suggest that less time spent on SB and more in PA was associated with better cognitive function. The three standard analytical approaches for 24HAC each have advantages and limitations, and selection of the most appropriate method should be guided by the scientific questions of interest and applicability of each model's assumptions. Further research is needed into the health implications of the distinct 24HAC patterns identified in this cohort.

Carbon Isotope Ratios of Plasma and RBC Fatty Acids Identify Meat Consumers in a 12-Week Inpatient Feeding Study of 32 Men

BACKGROUND

Molecular stable isotope ratios are a novel type of dietary biomarker with high sensitivity and specificity for certain foods. Among these, fatty acid carbon isotope ratios (CIRs) have strong potential but have not been investigated as dietary biomarkers.

OBJECTIVES

We evaluated whether fatty acid CIRs and mass proportions were associated with meat, fish, and sugar-sweetened beverage (SSB) intake.

METHODS

Thirty-two men [aged 46.2 ± 10.5 y; BMI (kg/m2): 27.2 ± 4.0] underwent a 12-wk inpatient dietary intervention at the National Institute of Diabetes and Digestive and Kidney Diseases in Phoenix, Arizona. Men were randomly assigned to 1 of 8 dietary treatments varying the presence/absence of dietary meat, fish, and SSBs in all combinations. Fatty acid CIRs and mass proportions were measured in fasting blood samples and adipose tissue biopsies that were collected pre- and postintervention. Dietary effects were analyzed using multivariable regression and receiver operating characteristic AUCs were calculated using logistic regression.

RESULTS

CIRs of the several abundant SFAs, MUFAs and arachidonic acid (20:4n-6) in plasma were strongly associated with meat, as were a subset of these fatty acids in RBCs. Effect sizes in plasma ranged from 1.01‰ to 1.93‰ and were similar but attenuated in RBCs. Mass proportions of those fatty acids were not associated with diet. CIRs of plasma dihomo-γ-linolenic acid (20:3n-6) and adipose palmitic acid (16:0) were weakly associated with SSBs. Mass proportions of plasma odd-chain fatty acids were associated with meat, and mass proportions of plasma EPA and DHA (20:5n-3 and 22:6n-3) were associated with fish.

CONCLUSIONS

CIRs of plasma and RBC fatty acids show promise as sensitive and specific measures of dietary meat. These provide different information from that provided by fatty acid mass proportions, and are informative where mass proportion is not. This trial is registered at www.clinicaltrials.gov as NCT01237093.

Multicultural Healthy Diet to Reduce Cognitive Decline & Alzheimer's Disease Risk: Study protocol for a pilot randomized controlled trial

BACKGROUND

Emerging evidence indicates that healthy dietary patterns are associated with higher cognitive status; however, few clinical trials have explored this association in diverse middle-aged adults before the onset of cognitive decline. We use novel ambulatory methods to assess cognition in natural settings in tandem with diet recording.

AIMS

We investigate whether the Multicultural Healthy Diet Study to Reduce Cognitive Decline & Alzheimer's Disease Risk, a pilot randomized controlled trial of an anti-inflammatory dietary pattern compared to usual diet, can mitigate cognitive decline and Alzheimer's Disease risk in a diverse population of 40-65 year old adults in Bronx, New York.

METHODS

Primary cognitive outcomes assessed at nine months are collected in an ecological momentary assessment "measurement burst" design, over the course of participants' daily lives. These ultra-brief, ambulatory cognitive assessments examine processing speed, visuospatial working memory, short-term associative memory binding, long-term associative memory, and working memory capacity. Key secondary outcomes relate to comparing dietary intake between study arms with respect to cognitive outcomes. We assess diet with food records using the National Cancer Institute's Automated Self-Administered 24-h record and serum biomarkers. We further investigate the association of self-reported diet and dietary biomarkers with inflammatory-based biomarkers.

CONCLUSION

This randomized controlled trial of diet and cognition for the first time combines novel measures of ambulatory cognitive assessment with web-based assessment of dietary intake recording. This new approach enabled the study to continue in the midst of the COVID-19 pandemic in remote format.

HIV Self-testing and Risk Behaviors Among Men Who Have Sex With Men in 23 US Cities, 2017

IMPORTANCE

HIV self-testing (HIVST) is a promising strategy to expand the HIV care continuum, particularly among priority populations at high risk of HIV infection. However, little is known about HIVST uptake among men who have sex with men (MSM) outside of clinical trial settings.

OBJECTIVE

To evaluate HIVST use among urban MSM in the US who reported testing within the past 12 months.

DESIGN, SETTING, AND PARTICIPANTS

A cross-sectional study of adult MSM in the 2017 National HIV Behavioral Surveillance system, which used venue-based sampling methods to collect data related to HIV testing, receipt of prevention services, and risk factors for HIV, was conducted at 588 venues in 23 urban areas in the contiguous US and Puerto Rico. All participants were offered HIV testing. Adult cisgender MSM who reported HIV-negative or unknown HIV status and obtained HIV testing in the past 12 months were included. Data for this study were collected between June 4, 2017, and December 22, 2017, and analyzed between October 23, 2020, and August 20, 2021.

MAIN OUTCOMES AND MEASURES

Self-reported HIVST in the past year. Adjusted prevalence ratios (aPRs) using survey weights were calculated to assess factors associated with HIVST.

RESULTS

A total of 6563 MSM in 23 urban areas met inclusion criteria, of whom 506 (7.7%) individuals reported HIVST in the past year. The median age of self-testers was 29 (IQR, 25-35) years, 52.8% had completed college, and 37.9% reported non-Hispanic White race. One self-tester reported seroconverting in the prior 12 months, and an additional 10 self-testers were diagnosed with HIV during the survey. HIVST was associated with sexual orientation disclosure (aPR, 10.27; 95% CI, 3.45-30.60; P < .001), perceived discrimination against people with HIV (aPR, 1.53; 95% CI, 1.09-2.03; P = .01), younger age (aPR, 0.74; 95% CI, 0.66-0.84; P < .001), higher educational level (aPR, 1.20; 95% CI, 1.04-1.37; P = .01), and higher income levels (aPR, 1.18; 95% CI, 1.04-1.32; P = .009). No association was noted with condomless anal sex (aPR, 0.96; 95% CI, 0.88-1.06, P = .88), sexually transmitted infections (aPR, 0.96; 95% CI, 0.70-1.30; P = .77), or preexposure prophylaxis use (aPR, 0.99; 95% CI, 0.75-1.30; P = .92).

CONCLUSIONS AND RELEVANCE

In this study, HIVST was relatively uncommon in this sample of urban MSM. HIVST may not be reaching those with lower socioeconomic status or who have not disclosed their sexual identity. The findings of this study suggest that efforts to increase HIVST should focus on engaging underserved and vulnerable subgroups of MSM.

Efficient odds ratio estimation under two-phase sampling using error-prone data from a multi-national HIV research cohort

Persons living with HIV engage in routine clinical care, generating large amounts of data in observational HIV cohorts. These data are often error-prone, and directly using them in biomedical research could bias estimation and give misleading results. A cost-effective solution is the two-phase design, under which the error-prone variables are observed for all patients during Phase I, and that information is used to select patients for data auditing during Phase II. For example, the Caribbean, Central, and South America network for HIV epidemiology (CCASAnet) selected a random sample from each site for data auditing. Herein, we consider efficient odds ratio estimation with partially audited, error-prone data. We propose a semiparametric approach that uses all information from both phases and accommodates a number of error mechanisms. We allow both the outcome and covariates to be error-prone and these errors to be correlated, and selection of the Phase II sample can depend on Phase I data in an arbitrary manner. We devise a computationally efficient, numerically stable EM algorithm to obtain estimators that are consistent, asymptotically normal, and asymptotically efficient. We demonstrate the advantages of the proposed methods over existing ones through extensive simulations. Finally, we provide applications to the CCASAnet cohort.

Nucleocapsid-specific antibody function is associated with therapeutic benefits from COVID-19 convalescent plasma therapy

Coronavirus disease 2019 (COVID-19) convalescent plasma (CCP), a passive polyclonal antibody therapeutic agent, has had mixed clinical results. Although antibody neutralization is the predominant approach to benchmarking CCP efficacy, CCP may also influence the evolution of the endogenous antibody response. Using systems serology to comprehensively profile severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) functional antibodies of hospitalized people with COVID-19 enrolled in a randomized controlled trial of CCP (ClinicalTrials.gov: NCT04397757), we find that the clinical benefits of CCP are associated with a shift toward reduced inflammatory Spike (S) responses and enhanced nucleocapsid (N) humoral responses. We find that CCP has the greatest clinical benefit in participants with low pre-existing anti-SARS-CoV-2 antibody function and that CCP-induced immunomodulatory Fc glycan profiles and N immunodominant profiles persist for at least 2 months. We highlight a potential mechanism of action of CCP associated with durable immunomodulation, outline optimal patient characteristics for CCP treatment, and provide guidance for development of a different class of COVID-19 hyperinflammation-targeting antibody therapeutic agents.

Decrease in Angiotensin-Converting Enzyme activity but not concentration in plasma/lungs in COVID-19 patients offers clues for diagnosis/treatment

Although several therapeutics are used to treat coronavirus disease 2019 (COVID-19) patients, there is still no definitive metabolic marker to evaluate disease severity and recovery or a quantitative test to end quarantine. Because severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) infects human cells via the angiotensin-converting-enzyme 2 (ACE2) receptor and COVID-19 is associated with renin-angiotensin system dysregulation, we evaluated soluble ACE2 (sACE2) activity in the plasma/saliva of 80 hospitalized COVID-19 patients and 27 non-COVID-19 volunteers, and levels of ACE2/Ang (1-7) in plasma or membrane (mACE2) in lung autopsy samples. sACE2 activity was markedly reduced (p < 0.0001) in COVID-19 plasma (n = 59) compared with controls (n = 27). Nadir sACE2 activity in early hospitalization was restored during disease recovery, irrespective of patient age, demographic variations, or comorbidity; in convalescent plasma-administered patients (n = 45), restoration was statistically higher than matched controls (n = 22, p = 0.0021). ACE2 activity was also substantially reduced in the saliva of COVID-19 patients compared with controls (p = 0.0065). There is a strong inverse correlation between sACE2 concentration and sACE2 activity and Ang (1-7) levels in participant plasmas. However, there were no difference in membrane ACE2 levels in lungs of autopsy tissues of COVID-19 (n = 800) versus other conditions (n = 300). These clinical observations suggest sACE2 activity as a potential biomarker and therapeutic target for COVID-19.

Statistical considerations for analyses of time-to-event endpoints in oncology clinical trials: Illustrations with CAR-T immunotherapy studies

Chimeric antigen receptor T-cell (CAR-T) therapy is an exciting development in the field of cancer immunology and has received a lot of interest in recent years. Many time-to-event (TTE) endpoints related to relapse, disease progression, and remission are analyzed in CAR-T studies to assess treatment efficacy. Definitions of these TTE endpoints are not always consistent, even for the same outcomes (e.g., progression-free survival), which often stems from analysis choices regarding which events to consider as part of the composite endpoint, censoring or competing risk in the analysis. Subsequent therapies such as hematopoietic stem cell transplantation are common but are not treated the same in different studies. Standard survival analysis methods are commonly applied to TTE analyses but often without full consideration of the assumptions inherent in the chosen analysis. We highlight two important issues of TTE analysis that arise in CAR-T studies, as well as in other settings in oncology: the handling of competing risks and assessing the association between a time-varying (post-infusion) exposure and the TTE outcome. We review existing analytical methods, including the cumulative incidence function and regression models for analysis of competing risks, and landmark and time-varying covariate analysis for analysis of post-infusion exposures. We clarify the scientific questions that the different analytical approaches address and illustrate how the application of an inappropriate method could lead to different results using data from multiple published CAR-T studies. Codes for implementing these methods in standard statistical software are provided.

Early Termination of Clinical Trials for Futility - Considerations for a Data and Safety Monitoring Board

Early Termination of Clinical Trials for FutilityClinical trials may be stopped for futility if there is little or no chance of demonstrating the hoped-for effect. Reasons include evidence of no treatment effect, substantial missing data that would unacceptably undermine trial conclusions, or event rates too low to support meaningful comparisons. This review examines issues faced by DSMBs in such settings.

Effect of Text Messaging and Behavioral Interventions on COVID-19 Vaccination Uptake: A Randomized Clinical Trial

IMPORTANCE

COVID-19 vaccine uptake among urban populations remains low.

OBJECTIVE

To evaluate whether text messaging with outbound or inbound scheduling and behaviorally informed content might increase COVID-19 vaccine uptake.

DESIGN, SETTING, AND PARTICIPANTS

This randomized clinical trial with a factorial design was conducted from April 29 to July 6, 2021, in an urban academic health system. The trial comprised 16 045 patients at least 18 years of age in Philadelphia, Pennsylvania, with at least 1 primary care visit in the past 5 years, or a future scheduled primary care visit within the next 3 months, who were unresponsive to prior outreach. The study was prespecified in the trial protocol, and data were obtained from the intent-to-treat population.

INTERVENTIONS

Eligible patients were randomly assigned in a 1:20:20 ratio to (1) outbound telephone call only by call center, (2) text message and outbound telephone call by call center to those who respond, or (3) text message, with patients instructed to make an inbound telephone call to a hotline. Patients in groups 2 and 3 were concurrently randomly assigned in a 1:1:1:1 ratio to receive different content: standard messaging, clinician endorsement (eg, "Dr. XXX recommends"), scarcity ("limited supply available"), or endowment framing ("We have reserved a COVID-19 vaccine appointment for you").

MAIN OUTCOMES AND MEASURES

The primary outcome was the proportion of patients who completed the first dose of the COVID-19 vaccine within 1 month, according to the electronic health record. Secondary outcomes were the completion of the first dose within 2 months and completion of the vaccination series within 2 months of initial outreach. Additional outcomes included the percentage of patients with invalid cell phone numbers (wrong number or nontextable), no response to text messaging, the percentage of patients scheduled for the vaccine, text message responses, and the number of telephone calls made by the access center. Analysis was on an intention-to-treat basis.

RESULTS

Among the 16 045 patients included, the mean (SD) age was 36.9 (11.1) years; 9418 (58.7%) were women; 12 869 (80.2%) had commercial insurance, and 2283 (14.2%) were insured by Medicaid; 8345 (52.0%) were White, 4706 (29.3%) were Black, and 967 (6.0%) were Hispanic or Latino. At 1 month, 14 of 390 patients (3.6% [95% CI, 1.7%-5.4%]) in the outbound telephone call-only group completed 1 vaccine dose, as did 243 of 7890 patients (3.1% [95% CI, 2.7%-3.5%]) in the text plus outbound call group (absolute difference, -0.5% [95% CI, -2.4% to 1.4%]; P = .57) and 253 of 7765 patients (3.3% [95% CI, 2.9%-3.7%]) in the text plus inbound call group (absolute difference, -0.3% [95% CI, -2.2% to 1.6%]; P = .72). Among the 15 655 patients receiving text messaging, 118 of 3889 patients (3.0% [95% CI, 2.5%-3.6%]) in the standard messaging group completed 1 vaccine dose, as did 135 of 3920 patients (3.4% [95% CI, 2.9%-4.0%]) in the clinician endorsement group (absolute difference, 0.4% [95% CI, -0.4% to 1.2%]; P = .31), 100 of 3911 patients (2.6% [95% CI, 2.1%-3.1%]) in the scarcity group (absolute difference, -0.5% [95% CI, -1.2% to 0.3%]; P = .20), and 143 of 3935 patients (3.6% [95% CI, 3.0%-4.2%]) in the endowment group (absolute difference, 0.6% [95% CI, -0.2% to 1.4%]; P = .14).

CONCLUSIONS AND RELEVANCE

There was no detectable increase in vaccination uptake among patients receiving text messaging compared with telephone calls only or behaviorally informed message content.

TRIAL REGISTRATION

ClinicalTrials.gov Identifier: NCT04834726.

Determinants of hospital outcomes for patients with COVID-19 in the University of Pennsylvania Health System

There is growing evidence that racial and ethnic minorities bear a disproportionate burden from COVID-19. Temporal changes in the pandemic epidemiology and diversity in the clinical course require careful study to identify determinants of poor outcomes. We analyzed 6255 hospitalized individuals with PCR-confirmed SARS-CoV-2 infection from one of 5 hospitals in the University of Pennsylvania Health System between March 2020 and March 2021, using electronic health records to assess risk factors and outcomes through 8 weeks post-admission. Discharge, readmission and mortality outcomes were analyzed in a multi-state model with multivariable Cox models for each transition. Mortality varied markedly over time, with cumulative incidence (95% CI) 30 days post-admission of 19.1% (16.9, 21.3) in March-April 2020, 5.7% (4.2, 7.5) in July-October 2020 and 10.5% (9.1,12.0) in January-March 2021; 26% of deaths occurred after discharge. Average age (SD) at admission varied from 62.7 (17.6) to 54.8 (19.9) to 60.5 (18.1); mechanical ventilation use declined from 21.3% to 9-11%. Compared to Caucasian, Black race was associated with more severe disease at admission, higher rates of co-morbidities and residing in a low-income zip code. Between-race risk differences in mortality risk diminished in multivariable models; while admitting hospital, increasing age, admission early in the pandemic, and severe disease and low blood pressure at admission were associated with increased mortality hazard. Hispanic ethnicity was associated with fewer baseline co-morbidities and lower mortality hazard (0.57, 95% CI: 0.37, .087). Multi-state modeling allows for a unified framework to analyze multiple outcomes throughout the disease course. Morbidity and mortality for hospitalized COVID-19 patients varied over time but post-discharge mortality remained non-trivial. Black race was associated with more risk factors for morbidity and with treatment at hospitals with lower mortality. Multivariable models suggest there are not between-race differences in outcomes. Future work is needed to better understand the identified between-hospital differences in mortality.

Amino Acid Nitrogen Isotope Ratios Respond to Fish and Meat Intake in a 12-Week Inpatient Feeding Study of Men

BACKGROUND

The natural abundance nitrogen stable isotope ratio (NIR) of whole tissue correlates with animal protein intakes, including meat and fish. Amino acid (AA) NIRs (NIRAAs) are more variable than the whole-tissue NIRs and may thus better differentiate among foods.

OBJECTIVES

We evaluated whether NIRAAs were associated with intakes of fish and meat and whether these dietary associations were larger than with whole-tissue NIRs.

METHODS

Men were recruited at the National Institute of Diabetes and Digestive and Kidney Diseases in Phoenix, Arizona, and randomly assigned to one of eight 12-wk inpatient dietary interventions, which varied the presence/absence of fish, meat, and sugar-sweetened beverages (SSBs) in all possible combinations. Fasting blood was drawn pre- and postintervention and plasma and RBC NIRAAs (free and protein-bound) were measured as secondary outcomes in 32 participants. Multivariable regression was used to determine responses of postintervention NIRAAs to dietary variables, and logistic regression was used to calculate receiver operating characteristic AUCs.

RESULTS

Most plasma NIRAAs increased with fish and meat intakes, but to a greater extent with fish intake. The largest increase in response to fish intake was plasma NIRLeucine (β = 2.19, SE = 0.26). The NIRThreonine decreased with both fish and meat intakes. Fewer RBC NIRAAs increased with fish intake, and only RBC NIRProline increased with meat intake. No plasma or RBC NIRAA responded to SSB intake. We identified fish intake with a high degree of accuracy using plasma NIRLeucine (corrected AUC, cAUC = 0.96) and NIRGlutamic acid/glutamine (cAUC = 0.93), and meat intake to a lower degree using plasma NIRProline (cAUC = 0.80) and RBC NIRProline (cAUC = 0.85).

CONCLUSIONS

Plasma and RBC NIRAAs were associated with fish and meat intakes but were not superior to whole-tissue stable isotope biomarkers in identifying these intakes in a US diet. This trial is registered at www.clinicaltrials.gov as NCT01237093.

A retrospective study of HIV pre-exposure prophylaxis counselling among non-Hispanic Black youth diagnosed with bacterial sexually transmitted infections in the United States, 2014-2019

INTRODUCTION

Youth account for a disproportionate number of new HIV infections; however, pre-exposure prophylaxis (PrEP) use is limited. We evaluated PrEP counselling rates among non-Hispanic Black youth in the United States after a bacterial sexually transmitted infection (STI) diagnosis.

METHODS

We conducted a retrospective cohort study of Black youth receiving care at two academically affiliated clinics in Philadelphia between June 2014 and June 2019. We compared PrEP counselling for youth who received primary care services versus those who did not receive primary care services, all of whom met PrEP eligibility criteria due to STI diagnosis per U.S. Centers for Disease Control and Prevention clinical practice guidelines. Two logistic regression models for receipt of PrEP counselling were fit: Model 1 focused on sexual and gender minority (SGM) status and Model 2 on rectal STIs with both models adjusted for patient- and healthcare-level factors.

RESULTS

Four hundred and sixteen patients met PrEP eligibility criteria due to STI based on sex assigned at birth and sexual partners. Thirty patients (7%) had documentation of PrEP counselling. Receipt of primary care services was not significantly associated with receipt of PrEP counselling in either Model 1 (adjusted OR (aOR) 0.10 [95% CI 0.01, 0.99]) or Model 2 (aOR 0.52 [95% CI 0.10, 2.77]). Receipt of PrEP counselling was significantly associated with later calendar years of STI diagnosis (aOR 6.80 [95% CI 1.64, 29.3]), assigned male sex at birth (aOR 26.2 [95% CI 3.46, 198]) and SGM identity (aOR 317 [95% CI 39.9, 2521]) in Model 1 and later calendar years of diagnosis (aOR 3.46 [95% CI 1.25, 9.58]), assigned male sex at birth (aOR 18.6 [95% CI 3.88, 89.3]) and rectal STI diagnosis (aOR 28.0 [95% CI 8.07, 97.5]) in Model 2. Fourteen patients (3%) started PrEP during the observation period; 12/14 (86%) were SGM primary care patients assigned male sex at birth.

CONCLUSIONS

PrEP counselling and uptake among U.S. non-Hispanic Black youth remain disproportionately low despite recent STI diagnosis. These findings support the need for robust investment in PrEP-inclusive sexual health services that are widely implemented and culturally tailored to Black youth, particularly cisgender heterosexual females.

Combining multiple imputation with raking of weights: An efficient and robust approach in the setting of nearly true models

Multiple imputation (MI) provides us with efficient estimators in model-based methods for handling missing data under the true model. It is also well-understood that design-based estimators are robust methods that do not require accurately modeling the missing data; however, they can be inefficient. In any applied setting, it is difficult to know whether a missing data model may be good enough to win the bias-efficiency trade-off. Raking of weights is one approach that relies on constructing an auxiliary variable from data observed on the full cohort, which is then used to adjust the weights for the usual Horvitz-Thompson estimator. Computing the optimally efficient raking estimator requires evaluating the expectation of the efficient score given the full cohort data, which is generally infeasible. We demonstrate MI as a practical method to compute a raking estimator that will be optimal. We compare this estimator to common parametric and semi-parametric estimators, including standard MI. We show that while estimators, such as the semi-parametric maximum likelihood and MI estimator, obtain optimal performance under the true model, the proposed raking estimator utilizing MI maintains a better robustness-efficiency trade-off even under mild model misspecification. We also show that the standard raking estimator, without MI, is often competitive with the optimal raking estimator. We demonstrate these properties through several numerical examples and provide a theoretical discussion of conditions for asymptotically superior relative efficiency of the proposed raking estimator.

A randomized controlled study of convalescent plasma for individuals hospitalized with COVID-19 pneumonia

BackgroundAntibody-based strategies for COVID-19 have shown promise in prevention and treatment of early disease. COVID-19 convalescent plasma (CCP) has been widely used but results from randomized trials supporting its benefit in hospitalized patients with pneumonia are limited. Here, we assess the efficacy of CCP in severely ill, hospitalized adults with COVID-19 pneumonia.MethodsWe performed a randomized control trial (PennCCP2), with 80 adults hospitalized with COVID-19 pneumonia, comparing up to 2 units of locally sourced CCP plus standard care versus standard care alone. The primary efficacy endpoint was comparison of a clinical severity score. Key secondary outcomes include 14- and 28-day mortality, 14- and 28-day maximum 8-point WHO ordinal score (WHO8) score, duration of supplemental oxygenation or mechanical ventilation, respiratory SARS-CoV-2 RNA, and anti-SARS-CoV-2 antibodies.ResultsEighty hospitalized adults with confirmed COVID-19 pneumonia were enrolled at median day 6 of symptoms and day 1 of hospitalization; 60% were anti-SARS-CoV-2 antibody seronegative. Participants had a median of 3 comorbidities, including risk factors for severe COVID-19 and immunosuppression. CCP treatment was safe and conferred significant benefit by clinical severity score (median [MED] and interquartile range [IQR] 10 [5.5-30] vs. 7 [2.75-12.25], P = 0.037) and 28-day mortality (n = 10, 26% vs. n = 2, 5%; P = 0.013). All other prespecified outcome measures showed weak evidence toward benefit of CCP.ConclusionTwo units of locally sourced CCP administered early in hospitalization to majority seronegative participants conferred a significant benefit in clinical severity score and 28-day mortality. Results suggest CCP may benefit select populations, especially those with comorbidities who are treated early.Trial RegistrationClinicalTrials.gov NCT04397757.FundingUniversity of Pennsylvania.

Leveraging machine learning predictive biomarkers to augment the statistical power of clinical trials with baseline magnetic resonance imaging

A key factor in designing randomized clinical trials is the sample size required to achieve a particular level of power to detect the benefit of a treatment. Sample size calculations depend upon the expected benefits of a treatment (effect size), the accuracy of measurement of the primary outcome, and the level of power specified by the investigators. In this study, we show that radiomic models, which leverage complex brain MRI patterns and machine learning, can be utilized in clinical trials with protocols that incorporate baseline MR imaging to significantly increase statistical power to detect treatment effects. Akin to the historical control paradigm, we propose to utilize a radiomic prediction model to generate a pseudo-control sample for each individual in the trial of interest. Because the variability of expected outcome across patients can mask our ability to detect treatment effects, we can increase the power to detect a treatment effect in a clinical trial by reducing that variability through using radiomic predictors as surrogates. We illustrate this method with simulations based on data from two cohorts in different neurologic diseases, Alzheimer's disease and glioblastoma multiforme. We present sample size requirements across a range of effect sizes using conventional analysis and models that include a radiomic predictor. For our Alzheimer's disease cohort, at an effect size of 0.35, total sample size requirements for 80% power declined from 246 to 212 for the endpoint cognitive decline. For our glioblastoma multiforme cohort, at an effect size of 1.65 with the endpoint survival time, total sample size requirements declined from 128 to 74. This methodology can decrease the required sample sizes by as much as 50%, depending on the strength of the radiomic predictor. The power of this method grows with increased accuracy of radiomic prediction, and furthermore, this method is most helpful when treatment effect sizes are small. Neuroimaging biomarkers are a powerful and increasingly common suite of tools that are, in many cases, highly predictive of disease outcomes. Here, we explore the possibility of using MRI-based radiomic biomarkers for the purpose of improving statistical power in clinical trials in the contexts of brain cancer and prodromal Alzheimer's disease. These methods can be applied to a broad range of neurologic diseases using a broad range of predictors of outcome to make clinical trials more efficient.

An approximate quasi-likelihood approach for error-prone failure time outcomes and exposures

Measurement error arises commonly in clinical research settings that rely on data from electronic health records or large observational cohorts. In particular, self-reported outcomes are typical in cohort studies for chronic diseases such as diabetes in order to avoid the burden of expensive diagnostic tests. Dietary intake, which is also commonly collected by self-report and subject to measurement error, is a major factor linked to diabetes and other chronic diseases. These errors can bias exposure-disease associations that ultimately can mislead clinical decision-making. We have extended an existing semiparametric likelihood-based method for handling error-prone, discrete failure time outcomes to also address covariate error. We conduct an extensive numerical study to compare the proposed method to the naive approach that ignores measurement error in terms of bias and efficiency in the estimation of the regression parameter of interest. In all settings considered, the proposed method showed minimal bias and maintained coverage probability, thus outperforming the naive analysis which showed extreme bias and low coverage. This method is applied to data from the Women's Health Initiative to assess the association between energy and protein intake and the risk of incident diabetes mellitus. Our results show that correcting for errors in both the self-reported outcome and dietary exposures leads to considerably different hazard ratio estimates than those from analyses that ignore measurement error, which demonstrates the importance of correcting for both outcome and covariate error.

Higher Neighborhood Population Density Is Associated with Lower Potassium Intake in the Hispanic Community Health Study/Study of Latinos (HCHS/SOL)

Current U.S. dietary guidelines recommend a daily potassium intake of 3400 mg/day for men and 2600 mg/day for women. Sub-optimal access to nutrient-rich foods may limit potassium intake and increase cardiometabolic risk. We examined the association of neighborhood characteristics related to food availability with potassium intake in the Hispanic Community Health Study/Study of Latinos (HCHS/SOL). 13,835 participants completed a 24-h dietary recall assessment and had complete covariates. Self-reported potassium intake was calibrated with an objective 24-h urinary potassium biomarker, using equations developed in the SOL Nutrition & Physical Activity Assessment Study (SOLNAS, N = 440). Neighborhood population density, median household income, Hispanic/Latino diversity, and a retail food environment index by census tract were obtained. Linear regression assessed associations with 24-h potassium intake, adjusting for individual-level and neighborhood confounders. Mean 24-h potassium was 2629 mg/day based on the SOLNAS biomarker and 2702 mg/day using multiple imputation and HCHS/SOL biomarker calibration. Compared with the lowest quartile of neighborhood population density, living in the highest quartile was associated with a 26% lower potassium intake in SOLNAS (adjusted fold-change 0.74, 95% CI 0.59–0.94) and a 39% lower intake in HCHS/SOL (adjusted fold-change 0.61 95% CI 0.45–0.84). Results were only partially explained by the retail food environment. The mechanisms by which population density affects potassium intake should be further studied.

Two-Phase Sampling Designs for Data Validation in Settings with Covariate Measurement Error and Continuous Outcome

Measurement errors are present in many data collection procedures and can harm analyses by biasing estimates. To correct for measurement error, researchers often validate a subsample of records and then incorporate the information learned from this validation sample into estimation. In practice, the validation sample is often selected using simple random sampling (SRS). However, SRS leads to inefficient estimates because it ignores information on the error-prone variables, which can be highly correlated to the unknown truth. Applying and extending ideas from the two-phase sampling literature, we propose optimal and nearly-optimal designs for selecting the validation sample in the classical measurement-error framework. We target designs to improve the efficiency of model-based and design-based estimators, and show how the resulting designs compare to each other. Our results suggest that sampling schemes that extract more information from the error-prone data are substantially more efficient than SRS, for both design- and model-based estimators. The optimal procedure, however, depends on the analysis method, and can differ substantially. This is supported by theory and simulations. We illustrate the various designs using data from an HIV cohort study.

Humanized CD19-Targeted Chimeric Antigen Receptor (CAR) T Cells in CAR-Naive and CAR-Exposed Children and Young Adults With Relapsed or Refractory Acute Lymphoblastic Leukemia

PURPOSE

CD19-targeted chimeric antigen receptor (CAR)-modified T cells demonstrate unprecedented responses in B-cell acute lymphoblastic leukemia (B-ALL); however, relapse remains a substantial challenge. Short CAR T-cell persistence contributes to this risk; therefore, strategies to improve persistence are needed.

METHODS

We conducted a pilot clinical trial of a humanized CD19 CAR T-cell product (huCART19) in children and young adults with relapsed or refractory B-ALL (n = 72) or B-lymphoblastic lymphoma (n = 2), treated in two cohorts: with (retreatment, n = 33) or without (CAR-naive, n = 41) prior CAR exposure. Patients were monitored for toxicity, response, and persistence of huCART19.

RESULTS

Seventy-four patients 1-29 years of age received huCART19. Cytokine release syndrome developed in 62 (84%) patients and was grade 4 in five (6.8%). Neurologic toxicities were reported in 29 (39%), three (4%) grade 3 or 4, and fully resolved in all cases. The overall response rate at 1 month after infusion was 98% (100% in B-ALL) in the CAR-naive cohort and 64% in the retreatment cohort. At 6 months, the probability of losing huCART19 persistence was 27% (95% CI, 14 to 41) for CAR-naive and 48% (95% CI, 30 to 64) for retreatment patients, whereas the incidence of B-cell recovery was 15% (95% CI, 6 to 28) and 58% (95% CI, 33 to 77), respectively. Relapse-free survival at 12 and 24 months, respectively, was 84% (95% CI, 72 to 97) and 74% (95% CI, 60 to 90) in CAR-naive and 74% (95% CI, 56 to 97) and 58% (95% CI, 37 to 90) in retreatment cohorts.

CONCLUSION

HuCART19 achieved durable remissions with long-term persistence in children and young adults with relapsed or refractory B-ALL, including after failure of prior CAR T-cell therapy.

Effect of Financial Incentives and Environmental Strategies on Weight Loss in the Healthy Weigh Study: A Randomized Clinical Trial

IMPORTANCE

Modest weight loss can lead to meaningful risk reduction in adults with obesity. Although both behavioral economic incentives and environmental change strategies have shown promise for initial weight loss, to date they have not been combined, or compared, in a randomized clinical trial.

OBJECTIVE

To test the relative effectiveness of financial incentives and environmental strategies, alone and in combination, on initial weight loss and maintenance of weight loss in adults with obesity.

DESIGN, SETTING, AND PARTICIPANTS

This randomized clinical trial was conducted from 2015 to 2019 at 3 large employers in Philadelphia, Pennsylvania. A 2-by-2 factorial design was used to compare the effects of lottery-based financial incentives, environmental strategies, and their combination vs usual care on weight loss and maintenance. Interventions were delivered via website, text messages, and social media. Participants included adult employees with a body mass index (BMI; weight in kilograms divided by height in meters squared) of 30 to 55 and at least 1 other cardiovascular risk factor. Data analysis was performed from June to July 2021.

INTERVENTIONS

Interventions included lottery-based financial incentives based on meeting weight loss goals, environmental change strategies tailored for individuals and delivered by text messages and social media, and combined incentives and environmental strategies.

MAIN OUTCOME AND MEASURES

The primary outcome was weight change from baseline to 18 months, measured in person.

RESULTS

A total of 344 participants were enrolled, with 86 participants each randomized to the financial incentives group, environmental strategies group, combined financial incentives and environmental strategies group, and usual care (control) group. Participants had a mean (SD) age of 45.6 (10.5) years and a mean (SD) BMI of 36.5 (7.1); 247 participants (71.8%) were women, 172 (50.0%) were Black, and 138 (40.1%) were White. At the primary end point of 18 months, participants in the incentives group lost a mean of 5.4 lb (95% CI, -11.3 to 0.5 lb [mean, 2.45 kg; 95% CI, -5.09 to 0.23 kg]), those in the environmental strategies group lost a mean of a 2.2 lb (95% CI, -7.7 to 3.3 lb [mean, 1.00 kg; 95% CI, -3.47 to 1.49 kg]), and the combination group lost a mean of 2.4 lb (95% CI, -8.2 to 3.3 lb [mean, 1.09 kg; 95% CI, -3.69 to 1.49 kg]) more than participants in the usual care group. Financial incentives, environmental change strategies, and their combination were not significantly more effective than usual care. At 24 months, after 6 months without an intervention, the difference in the change from baseline was similar to the 18-month results, with no significant differences among groups.

CONCLUSIONS AND RELEVANCE

In this randomized clinical trial, across all study groups, participants lost a modest amount of weight but those who received financial incentives, environmental change, or the combined intervention did not lose significantly more weight than those in the usual care group. Employees with obesity may benefit from more intensive individualized weight loss strategies.

TRIAL REGISTRATION

ClinicalTrials.gov Identifier: NCT02878343.

Lack of Atorvastatin Effect on Monocyte Gene Expression and Inflammatory Markers in HIV-1-infected ART-suppressed Individuals at Risk of non-AIDS Comorbidities

BACKGROUND

Many people living with HIV have persistent monocyte activation despite viral suppression by antiretroviral therapy (ART), which contributes to non-AIDS complications including neurocognitive and other disorders. Statins have immunomodulatory properties that might be beneficial by reducing monocyte activation.

METHODS

We previously characterized monocyte gene expression and inflammatory markers in 11 HIV-positive individuals on long-term ART (HIV/ART) at risk for non-AIDS complications because of low nadir CD4+ counts (median 129 cells/uL) and elevated hsCRP. Here, these individuals participated in a double-blind, randomized, placebo-controlled crossover study of 12 weeks of atorvastatin treatment. Monocyte surface markers were assessed by flow cytometry, plasma mediators by ELISA and Luminex, and monocyte gene expression by microarray analysis.

RESULTS

Among primary outcome measures, 12 weeks of atorvastatin treatment led to an unexpected increase in CCR2+ monocytes (P=0.04), but did not affect CD16+ or CD163+ monocytes, nor levels in plasma of CCL2/MCP-1 or sCD14. Among secondary outcomes, atorvastatin treatment was associated with decreased plasma hsCRP (P=0.035) and IL-2R (P=0.012). Treatment was also associated with increased total CD14+ monocytes (P=0.015), and increased plasma CXCL9 (P=0.003) and IL-12 (P<0.001). Comparable results were seen in a subgroup that had inflammatory marker elevations at baseline. Atorvastatin treatment did not significantly alter monocyte gene expression or normalize aberrant baseline transcriptional patterns.

CONCLUSIONS

In this study of aviremic HIV+ individuals at high risk of non-AIDS events, 12 weeks of atorvastatin did not normalize monocyte gene expression patterns nor lead to significant changes in monocyte surface markers or plasma mediators linked to non-AIDS comorbidities.

On the Use of Regression Calibration in a Complex Sampling Design With Application to the Hispanic Community Health Study/Study of Latinos

Regression calibration is the most widely used method to adjust regression parameter estimates for covariate measurement error. Yet its application in the context of a complex sampling design, for which the common bootstrap variance estimator can be less straightforward, has been less studied. We propose 2 variance estimators for a multistage probability-based sampling design, a parametric and a resampling-based multiple imputation approach, where a latent mean exposure needed for regression calibration is the target of imputation. This work was motivated by the Hispanic Community Health Study/Study of Latinos (HCHS/SOL) data from 2008 to 2011, for which relationships between several outcomes and diet, an error-prone self-reported exposure, are of interest. We assessed the relative performance of these variance estimation strategies in an extensive simulation study built on the HCHS/SOL data. We further illustrate the proposed estimators with an analysis of the cross-sectional association of dietary sodium intake with hypertension-related outcomes in a subsample of the HCHS/SOL cohort. We have provided guidelines for the application of regression models with regression-calibrated exposures. Practical considerations for implementation of these 2 variance estimators in the setting of a large multicenter study are also discussed. Code to replicate the presented results is available online.

Improved generalized raking estimators to address dependent covariate and failure-time outcome error

Biomedical studies that use electronic health records (EHR) data for inference are often subject to bias due to measurement error. The measurement error present in EHR data is typically complex, consisting of errors of unknown functional form in covariates and the outcome, which can be dependent. To address the bias resulting from such errors, generalized raking has recently been proposed as a robust method that yields consistent estimates without the need to model the error structure. We provide rationale for why these previously proposed raking estimators can be expected to be inefficient in failure-time outcome settings involving misclassification of the event indicator. We propose raking estimators that utilize multiple imputation, to impute either the target variables or auxiliary variables, to improve the efficiency. We also consider outcome-dependent sampling designs and investigate their impact on the efficiency of the raking estimators, either with or without multiple imputation. We present an extensive numerical study to examine the performance of the proposed estimators across various measurement error settings. We then apply the proposed methods to our motivating setting, in which we seek to analyze HIV outcomes in an observational cohort with EHR data from the Vanderbilt Comprehensive Care Clinic.

Behavioral interventions to improve population health outreach for hepatitis C screening: randomized clinical trial

OBJECTIVE

To evaluate whether opt out framing, messaging incorporating behavioral science concepts, or electronic communication increases the uptake of hepatitis C virus (HCV) screening in patients born between 1945 and 1965.

DESIGN

Pragmatic randomized controlled trial.

SETTING

43 primary care practices from one academic health system (Philadelphia, PA, USA) between April 2019 and May 2020.

PARTICIPANTS

Patients born between 1945 and 1965 with no history of screening and at least two primary care visits in the two years before the enrollment period.

INTERVENTIONS

This multilevel trial was divided into two studies. Substudy A included 1656 eligible patients of 17 primary care clinicians who were randomized in a 1:1 ratio to a mailed letter about HCV screening (letter only), or a similar letter with a laboratory order for HCV screening (letter+order). Substudy B included the remaining 19 837 eligible patients followed by 417 clinicians. Active electronic patient portal users were randomized 1:5 to receive a mailed letter about HCV screening (letter), or an electronic patient portal message with similar content (patient portal); inactive patient portal users were mailed a letter. In a factorial design, patients in substudy B were also randomized 1:1 to receive standard content (usual care), or content based on principles of social norming, anticipated regret, reciprocity, and commitment (behavioral content).

MAIN OUTCOME MEASURES

Proportion of patients who completed HCV testing within four months.

RESULTS

21 303 patients were included in the intention-to-treat analysis. Among the 1642 patients in substudy A, 19.2% (95% confidence interval 16.5% to 21.9%) completed screening in the letter only arm and 43.1% (39.7% to 46.4%) in the letter+order arm (P<0.001). Among the 19 661 patients in substudy B, 14.6% (13.9% to 15.3%) completed screening with usual care content and 13.6% (13.0% to 14.3%) with behavioral science content (P=0.06). Among active patient portal users, 17.8% (16.0% to 19.5%) completed screening after receiving a letter and 13.8% (13.1% to 14.5%) after receiving a patient portal message (P<0.001).

CONCLUSIONS

Opt out framing and effort reduction by including a signed laboratory order with outreach increased screening for HCV. Behavioral science messaging content did not increase uptake, and mailed letters achieved a greater response rate than patient portal messages.

TRIAL REGISTRATION

ClinicalTrials.gov NCT03712553.

CCR5-edited CD4+ T cells augment HIV-specific immunity to enable post-rebound control of HIV replication

BackgroundWe conducted a phase I clinical trial that infused CCR5 gene-edited CD4+ T cells to determine how these T cells can better enable HIV cure strategies.MethodsThe aim of trial was to develop RNA-based approaches to deliver zinc finger nuclease (ZFN), evaluate the effect of CCR5 gene-edited CD4+ T cells on the HIV-specific T cell response, test the ability of infused CCR5 gene-edited T cells to delay viral rebound during analytical treatment interruption, and determine whether individuals heterozygous for CCR5 Δ32 preferentially benefit. We enrolled 14 individuals living with HIV whose viral load was well controlled by antiretroviral therapy (ART). We measured the time to viral rebound after ART withdrawal, the persistence of CCR5-edited CD4+ T cells, and whether infusion of 10 billion CCR5-edited CD4+ T cells augmented the HIV-specific immune response.ResultsInfusion of the CD4+ T cells was well tolerated, with no serious adverse events. We observed a modest delay in the time to viral rebound relative to historical controls; however, 3 of the 14 individuals, 2 of whom were heterozygous for CCR5 Δ32, showed post-viral rebound control of viremia, before ultimately losing control of viral replication. Interestingly, only these individuals had substantial restoration of HIV-specific CD8+ T cell responses. We observed immune escape for 1 of these reinvigorated responses at viral recrudescence, illustrating a direct link between viral control and enhanced CD8+ T cell responses.ConclusionThese findings demonstrate how CCR5 gene-edited CD4+ T cell infusion could aid HIV cure strategies by augmenting preexisting HIV-specific immune responses.REGISTRATIONClinicalTrials.gov NCT02388594.FundingNIH funding (R01AI104400, UM1AI126620, U19AI149680, T32AI007632) was provided by the National Institute of Allergy and Infectious Diseases (NIAID), the National Institute on Drug Abuse (NIDA), the National Institute of Mental Health (NIMH), and the National Institute of Neurological Disorders and Stroke (NINDS). Sangamo Therapeutics also provided funding for these studies.

The carbon isotope ratios of nonessential amino acids identify sugar-sweetened beverage (SSB) consumers in a 12-wk inpatient feeding study of 32 men with varying SSB and meat exposures

BACKGROUND

The carbon isotope ratios (CIRs) of individual amino acids (AAs) may provide more sensitive and specific biomarkers of sugar-sweetened beverages (SSBs) than total tissue CIR. Because CIRs turn over slowly, long-term controlled-feeding studies are needed in their evaluation.

OBJECTIVE

We assessed the responses of plasma and RBC CIRAA's to SSB and meat intake in a 12-wk inpatient feeding study.

METHODS

Thirty-two men (aged 46.2 ± 10.5 y) completed the feeding study at the National Institute of Diabetes and Digestive and Kidney Diseases in Phoenix, Arizona. The effects of SSB, meat, and fish intake on plasma and RBC CIRAA's were evaluated in a balanced factorial design with each dietary variable either present or absent in a common weight-maintaining, macronutrient-balanced diet. Fasting blood samples were collected biweekly from baseline. Dietary effects on the postfeeding CIR of 5 nonessential AAs (CIRNEAA's) and 4 essential AAs (CIREAA's) were analyzed using multivariable regression.

RESULTS

In plasma, 4 of 5 CIRNEAA's increased with SSB intake. Of these, the CIRAla was the most sensitive (β = 2.81, SE = 0.38) to SSB intake and was not affected by meat or fish intake. In RBCs, all 5 CIRNEAA's increased with SSBs but had smaller effect sizes than in plasma. All plasma CIREAA's increased with meat intake (but not SSB or fish intake), and the CIRLeu was the most sensitive (β = 1.26, SE = 0.23). CIRs of leucine and valine also increased with meat intake in RBCs. Estimates of turnover suggest that CIRAA's in plasma, but not RBCs, were in equilibrium with the diets by the end of the study.

CONCLUSIONS

The results of this study in men support CIRNEAA's as potential biomarkers of SSB intake and suggest CIREAA's as potential biomarkers of meat intake in US diets. This trial was registered at clinicaltrials.gov/ct2/show/NCT01237093 as NCT01237093.

Efficient semiparametric inference for two-phase studies with outcome and covariate measurement errors

In modern observational studies using electronic health records or other routinely collected data, both the outcome and covariates of interest can be error-prone and their errors often correlated. A cost-effective solution is the two-phase design, under which the error-prone outcome and covariates are observed for all subjects during the first phase and that information is used to select a validation subsample for accurate measurements of these variables in the second phase. Previous research on two-phase measurement error problems largely focused on scenarios where there are errors in covariates only or the validation sample is a simple random sample of study subjects. Herein, we propose a semiparametric approach to general two-phase measurement error problems with a quantitative outcome, allowing for correlated errors in the outcome and covariates and arbitrary second-phase selection. We devise a computationally efficient and numerically stable expectation-maximization algorithm to maximize the nonparametric likelihood function. The resulting estimators possess desired statistical properties. We demonstrate the superiority of the proposed methods over existing approaches through extensive simulation studies, and we illustrate their use in an observational HIV study.

Raking and regression calibration: Methods to address bias from correlated covariate and time-to-event error

Medical studies that depend on electronic health records (EHR) data are often subject to measurement error, as the data are not collected to support research questions under study. These data errors, if not accounted for in study analyses, can obscure or cause spurious associations between patient exposures and disease risk. Methodology to address covariate measurement error has been well developed; however, time-to-event error has also been shown to cause significant bias, but methods to address it are relatively underdeveloped. More generally, it is possible to observe errors in both the covariate and the time-to-event outcome that are correlated. We propose regression calibration (RC) estimators to simultaneously address correlated error in the covariates and the censored event time. Although RC can perform well in many settings with covariate measurement error, it is biased for nonlinear regression models, such as the Cox model. Thus, we additionally propose raking estimators which are consistent estimators of the parameter defined by the population estimating equation. Raking can improve upon RC in certain settings with failure-time data, require no explicit modeling of the error structure, and can be utilized under outcome-dependent sampling designs. We discuss features of the underlying estimation problem that affect the degree of improvement the raking estimator has over the RC approach. Detailed simulation studies are presented to examine the performance of the proposed estimators under varying levels of signal, error, and censoring. The methodology is illustrated on observational EHR data on HIV outcomes from the Vanderbilt Comprehensive Care Clinic.

Regression calibration to correct correlated errors in outcome and exposure

Measurement error arises through a variety of mechanisms. A rich literature exists on the bias introduced by covariate measurement error and on methods of analysis to address this bias. By comparison, less attention has been given to errors in outcome assessment and nonclassical covariate measurement error. We consider an extension of the regression calibration method to settings with errors in a continuous outcome, where the errors may be correlated with prognostic covariates or with covariate measurement error. This method adjusts for the measurement error in the data and can be applied with either a validation subset, on which the true data are also observed (eg, a study audit), or a reliability subset, where a second observation of error prone measurements are available. For each case, we provide conditions under which the proposed method is identifiable and leads to consistent estimates of the regression parameter. When the second measurement on the reliability subset has no error or classical unbiased measurement error, the proposed method is consistent even when the primary outcome and exposures of interest are subject to both systematic and random error. We examine the performance of the method with simulations for a variety of measurement error scenarios and sizes of the reliability subset. We illustrate the method's application using data from the Women's Health Initiative Dietary Modification Trial.

D614G Spike Mutation Increases SARS CoV-2 Susceptibility to Neutralization

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) spike protein acquired a D614G mutation early in the pandemic that confers greater infectivity and is now the globally dominant form. To determine whether D614G might also mediate neutralization escape that could compromise vaccine efficacy, sera from spike-immunized mice, nonhuman primates, and humans were evaluated for neutralization of pseudoviruses bearing either D614 or G614 spike. In all cases, the G614 pseudovirus was moderately more susceptible to neutralization. The G614 pseudovirus also was more susceptible to neutralization by receptor-binding domain (RBD) monoclonal antibodies and convalescent sera from people infected with either form of the virus. Negative stain electron microscopy revealed a higher percentage of the 1-RBD "up" conformation in the G614 spike, suggesting increased epitope exposure as a mechanism of enhanced vulnerability to neutralization. Based on these findings, the D614G mutation is not expected to be an obstacle for current vaccine development.

Persistent Disparities in Smoking Rates Among PLWH Compared to the General Population in Philadelphia, 2009-2014

Despite reductions in smoking rates in the general population, little is known about recent smoking trends among people living with HIV (PLWH). We compared the risk for smoking and temporal trends in smoking among PLWH and the general population in the Philadelphia metropolitan area between 2009 and 2014. We used weighted logistic regression to assess the relation between HIV and smoking, and examined temporal smoking trends. The adjusted odds ratio (OR) for smoking comparing PLWH to the general population was 1.80 (95% CI 1.55-2.09) after adjusting for socio-economic, demographic, and mental health diagnosis variables. Smoking prevalence decreased in both the PLWH and general populations during the study period, and we did not observe a significant difference in rates of decline between groups (P = 0.54). Despite overall progress in smoking cessation, a disparity persisted in smoking rates between PLWH and the general population, with and without adjustment for socio-economic, demographic, and mental health variables. Further research is needed to understand the mechanisms linking HIV and tobacco use in order to inform public health efforts to reduce smoking among PLWH.

Two-phase analysis and study design for survival models with error-prone exposures

Increasingly, medical research is dependent on data collected for non-research purposes, such as electronic health records data. Health records data and other large databases can be prone to measurement error in key exposures, and unadjusted analyses of error-prone data can bias study results. Validating a subset of records is a cost-effective way of gaining information on the error structure, which in turn can be used to adjust analyses for this error and improve inference. We extend the mean score method for the two-phase analysis of discrete-time survival models, which uses the unvalidated covariates as auxiliary variables that act as surrogates for the unobserved true exposures. This method relies on a two-phase sampling design and an estimation approach that preserves the consistency of complete case regression parameter estimates in the validated subset, with increased precision leveraged from the auxiliary data. Furthermore, we develop optimal sampling strategies which minimize the variance of the mean score estimator for a target exposure under a fixed cost constraint. We consider the setting where an internal pilot is necessary for the optimal design so that the phase two sample is split into a pilot and an adaptive optimal sample. Through simulations and data example, we evaluate efficiency gains of the mean score estimator using the derived optimal validation design compared to balanced and simple random sampling for the phase two sample. We also empirically explore efficiency gains that the proposed discrete optimal design can provide for the Cox proportional hazards model in the setting of a continuous-time survival outcome.