Bibhas Chakraborty

BACKGROUND Maintaining engagement with digital apps is a key concern when developing effective digital health interventions. Previous research typically examines engagement through a single-event outcome, such as time to first disengagement. However, users may repeatedly stop using and later return to the app, experiencing multiple disengagement and re-engagement cycles. Understanding the factors associated with re-engagement is valuable and may help inform the optimization and personalization of digital intervention strategies. OBJECTIVE This study aimed to characterize re-engagement with Drink Less, an alcohol reduction app, as a recurrent-event process and to examine its associations with baseline characteristics and recent app-use history. METHODS We analyzed data from 19,233 users of the Drink Less app who had a baseline Alcohol Use Disorders Identification Test (AUDIT) score of at least 8 on the 0 to 40 scale, indicative of drinking at hazardous or harmful levels. Disengagement was defined as 7 or more consecutive days without app use and a disengagement episode began when this threshold was reached. Re-engagement was defined as the first recorded app-use session after entry into that episode. The Prentice-Williams-Peterson Gap-Time (PWP-GT) model was used to examine associations between baseline characteristics, recent app-use history, and re-engagement. Recent app-use history was represented by previous gap duration, measured as the number of days from entry into the immediately preceding disengagement episode to the corresponding re-engagement that ended it; and previous engagement duration, defined as the number of days from that re-engagement to entry into the current disengagement episode. RESULTS In the baseline characteristics model, users in the Harmful AUDIT category (16-19) were more likely to re-engage after disengagement than those in the Hazardous category (8-15; HR 1.0431, 95% CI 1.0026-1.0852; P=.04). Older age was also associated with a greater likelihood of returning to the app (HR 1.0043 per year, 95% CI 1.0029-1.0058; P<.001), whereas users in manual employment (HR 0.8437, 95% CI 0.8000-0.8899; P<.001) and other employment categories (HR 0.9436, 95% CI 0.9016-0.9876; P=.01) were less likely to re-engage than those in non-manual employment. There was no clear evidence of a difference by sex. In separate lagged variable models, longer previous gap duration was negatively associated with re-engagement (HR 0.9921 per day, 95% CI 0.9913-0.9930; P<.001), whereas longer previous engagement duration was positively associated with re-engagement (HR 1.0078 per day, 95% CI 1.0067-1.0089; P<.001). All comparisons accounted for the other covariates in the respective model and were conditional on disengagement episode order. CONCLUSIONS Re-engagement with Drink Less was significantly associated with age and employment, as well as with users’ recent app-use history. Together, these findings could inform the development and testing of targeted support both at registration and as the app learns about users’ behavior.

With the rapid emergence of personalized healthcare, adaptive interventions have gained significant traction and relevance. Contemporary research has introduced a sophisticated trial design known as the sequential multiple assignment randomized trial (SMART) to advance the development of effective adaptive interventions. A critical element of the SMART design is the determination of the sample size. The existing literature provides sample size calculation formulas for SMART designs, which encompass a variety of approaches and types of outcome data. However, these formulas have primarily been developed under the assumption of normality for the outcome data. In practice, the fields where SMART is employed exhibit a significant prevalence of non-normality in the outcomes of interest. This paper delves into a specific scenario where the outcome demonstrates skewed behavior. We develop precision-based and power-based formulas for skewed outcomes, considering the requirements of both pilot and full-scale SMARTs. We present the operating characteristics of the formulas and perform extensive simulations under various design specifications to validate their usefulness. To demonstrate practical utility, we apply our formulas to the SMART+ study, a digital intervention trial that includes a skewed outcome among its outcomes of interest, highlighting relevance in real-world planning.

Purpose: Digital mental health interventions could improve population health. However, our understanding of how to improve their effectiveness through personalization is limited. We describe the outcomes of a micro-randomized trial (MRT) of a 60-day text messaging program, delivered during the final stages of the COVID-19 pandemic. We analyze the effects of 6 different types of cognitive behavioral therapy (CBT) and dialectical behavioral therapy (DBT) messages, and examine whether socio-demographic, health-related, and contextual factors (eg, timing of sending a message) moderate effectiveness. Methods: This secondary analysis included 373 adults aged 18 years or older (mean = 38.07), mostly English-speaking female, highly educated White individuals who reported being in good health, but having high stress and struggling to get their basic needs met. Participants were recruited between December 2021 and September 2022 through social media and a recruitment service (Health Match). We used the weighted and centered least squares method to examine the proximal effects of the text-message categories on daily mood, and whether sociodemographic, health-related factors, and the timing of message delivery moderated effectiveness. Results: None of the 6 CBT/DBT-message categories significantly improved mood ratings marginally. However, their effectiveness varied based on the time of day and self-reported health. Overall, sending a message in the afternoon was more effective. However, certain message categories showed significant effects when sent in the morning among participants with poorer self-reported health, while some showed lower effectiveness when sent in the late afternoon for participants with better health. Conclusion: Our analysis underscores the importance of considering the time of day and individual health status when developing digital mental health interventions. These insights can help in designing more personalized, adaptive, and effective digital interventions.

Background: The current landscape of emergency care (EC) is marked by high demand, leading to issues such as emergency department boarding, overcrowding, and subsequent delays that impact the quality and safety of patient care. Integrating data science into EC can enhance decision-making with predictive, preventative, personalized, and participatory approaches. However, gaps in adherence to fairness, accountability, interpretability, and responsibility are evident, particularly due to barriers to data-sharing, which often result in a lack of transparency and robust oversight in these applications. Objective: The FAIR-EC (Fair, Accountable, Interpretable, and Responsible-Emergency Care) collaboration adapts the existing Fair, Accountable, Interpretable, and Responsible principles to address emerging challenges as data science integrates with EC. This initiative aims to transform EC by establishing ethical artificial intelligence standards specifically tailored for this integration. By bridging the gap between EC professionals, data scientists, and other stakeholders, the collaboration promotes international cooperation that leverages advanced data science techniques to enhance EC outcomes across different care settings. Methods: We propose a federated research design to analyze extensive datasets from various global institutions without compromising patient privacy. This approach transforms epidemiological research with advanced data science techniques, emphasizing the harmonization of data for comprehensive analyses across different health care systems. Results: The FAIR-EC initiative has facilitated the identification and harmonization of datasets from diverse geographical regions, enabling the examination of regional variations in EC practices. As of paper submission, participating sites have identified retrospective EC datasets totaling >2 million records (eg, Duke Health >400,000 and Singapore General Hospital >1.7 million records). Initial projects have demonstrated feasibility and operational readiness, including implementation of federated workflows and ongoing development of a federated scoring system, cross-site evaluation, and adaptation of association studies and predictive models across various regions. Cross-site harmonization and pilot analyses are underway (with local ethics approvals in progress), and first multisite results are expected to be submitted in mid-late 2026, with additional project-level publications anticipated in 2027. These efforts highlight the feasibility of leveraging advanced data science techniques to address the complexities of EC while preserving patient privacy without centralizing individual-level data. This project was funded from September 1, 2022, to August 31, 2023. Conclusions: FAIR-EC integrates data science ethically and effectively into EC, addressing challenges such as fragmented data, real-time handoffs, and public health crises. Its federated design harmonizes diverse data streams while preserving privacy, and its emphasis on ethical artificial intelligence aligns with the dynamic nature of EC. Despite challenges in data variability and system complexity, FAIR-EC establishes a strong foundation for innovation in global EC.

Objectives: Electronic health record (EHR) phenotyping often relies on noisy proxy labels, which undermine the reliability of downstream risk prediction. Active learning can reduce annotation costs, but typical heuristics do not directly optimize downstream prediction. Our goal was to develop a framework that directly uses downstream prediction performance as feedback to guide phenotype correction and sample selection under constrained labeling budgets. Materials and Methods: We propose reinforcement-enhanced label-efficient active phenotyping (RELEAP), a reinforcement learning-based active learning framework. Reinforcement-enhanced label-efficient adaptively integrates multiple querying strategies and, unlike prior methods, updates its policy based on feedback from downstream models. We evaluated RELEAP on a de-identified Duke University Health System (DUHS) cohort (2014-2024) for incident lung cancer risk prediction, using logistic regression and penalized Cox survival models. Performance was benchmarked against noisy-label baselines and single-strategy active learning. Results: Reinforcement-enhanced label-efficient improved over the proxy-only baseline and approached oracle performance under the same budget. Logistic AUC increased from 0.774 to 0.807. Survival concordance index increased from 0.715 to 0.749. Gains were stable across iterations using downstream feedback. These trends were consistent in sex-stratified subgroup analyses (female vs male). Discussion: By linking phenotype refinement to prediction outcomes, RELEAP learns which samples most improve downstream discrimination and calibration, offering a more principled alternative to fixed active learning rules. Conclusion: Reinforcement-enhanced label-efficient optimizes phenotype correction through downstream feedback, offering a scalable, label-efficient paradigm that reduces manual chart review and enhances the reliability of EHR-based risk prediction.

Strong statistical voice is defined as the ability to advocate and negotiate for good and ethical statistical practices, including integrating and resolving differing scientific approaches. This skill is crucial for biostatisticians who work on biomedical research teams, as it ensures the integrity and accuracy of statistical analyses and fosters productive collaborations with non-statisticians. Despite its importance, new graduates often lack targeted training opportunities. This manuscript presents a scalable training approach through the development of online videos. Preliminary didactic materials focused on two key applications: providing written comments on manuscripts and engaging in study design discussions. To evaluate this training approach, a survey was conducted among biostatistics staff in the Duke Biostatistics, Epidemiology, and Research Design Core. The survey results indicated that all respondents strongly agreed on the importance of strong statistical voice in biostatistics practice. The clarity of the training materials and examples received positive feedback, though suggestions for improvement included enhancing video engagement and providing more hands-on training. This information will guide the development of formal training videos embedded within a mentored training program that aims to teach biostatisticians and other quantitative scientists how to effectively work on teams in biomedical research.