Eric B. Hekler, Ph.D. - US grants

Abstract Objectives: The Advanced Data Analytics Program To (ADAPT) will enhance behavioral and social sciences research, by training a diverse next generation of data scientists who will learn interdisciplinary skills needed for successful careers in behavioral and social sciences health-related data science. Rationale: San Diego is a hub for genomics, mobile technology, behavioral health research and data science in Southern California, yet no data science curriculum for behavioral scientists currently exists. The ADAPT program will fill this gap and intersect the areas of health sciences, informatics, computer science, and statistics in Southern California. Design: ADAPT will educate doctoral students in the behavioral and social sciences to build and further expand an ecosystem for big data analytics that promotes finding, accessing, interoperating, and reusing digital objects and responsibly computing with human subjects? data in cloud environments. The ADAPT program will be based at the University of California, San Diego (UCSD), with faculty collaborators from San Diego State University. It will be based on two joint doctoral programs (JDPs) at these universities (Clinical Psychology and Public Health/Behavioral Health). Dual mentoring by faculty with expertise in behavioral and social sciences and computer science, biomedical informatics, or statistics will ensure a truly interdisciplinary focus that will cover team science and responsible conduct of research. Key Activities: Trainees will gain expertise through coursework, research experience during rotations and external internships, mentoring and other activities. Existing data science courses were selected for the curriculum, which will also include a new course in cloud-based human subjects? data computing. Through individualized development plans, ADAPT trainees will work with their faculty mentors to tailor the curriculum and career paths according to students? interests and skills. Data science coursework will utilize elective course slots in the JDP curricula, will typically be completed in years 1 and 2 of the JDPs. They will provide the foundational knowledge needed for academic and industry rotations and for the start of the trainees? research phase. Projected Number of Trainees: 6 first or second year JDP students Planned Duration of Appointments: 3 years, renewed annually based on good academic standing Intended Trainee Outcomes: Metrics for success will include number and quality of publications, and rate of academic milestone completion. Trainees who complete the ADAPT program will possess the scientific knowledge needed to be a behavioral health data scientist, understand ethical and regulatory aspects of computing with protected health information, and will become critical members of scientific teams working in academia, government, for-profit and non-profit research institutions.

Background: Strong evidence indicates physical activity (PA) reduces risk of bladder, breast, colon, endometrium, esophagus, gastric, and renal cancer, and there is moderate evidence for lung cancer. Individuals aged 40+ who are inactive are at high risk of developing cancers 58,65 but only 1/3 meet guidelines for PA;5-15 thus, they are an important group to target. While effective PA interventions exist, interventions often work only for some individuals or only for a limited time,16-18 thus establishing the need for interventions that can account for dynamic, idiosyncratic PA determinants in order to support each person?s PA. In response, we developed JustWalk, a modular adaptive mobile health (mHealth) intervention that makes daily N-of-1 adjustments to support PA for each person. JustWalk is based on Social Cognitive Theory (SCT) with N-of-1 adaptation driven by a mathematical dynamical model of SCT, which we have developed and validated. JustWalk can perform N-of-1 adaptation based on our innovative use of control engineering methods, which we call a control optimization trial (COT). We have a digital platform and empirical justification for our next step: to evaluate, in a randomized controlled trial (RCT), whether using a COT approach to continuously optimize a PA intervention to each individual is superior to an intervention that is identical but lacks the COT methods. Primary purpose: Evaluate differences in minutes/week of moderate-to-vigorous intensity PA (MVPA) among the COT- optimized vs. non-COT groups at 12 months. Hypotheses: We hypothesize significantly higher minutes/week of MVPA in the intervention arm (COT) relative to control (non-COT) as measured via ActiGraph (powered for effect size of ?0.32). Methods: We will conduct this RCT with 386 adults aged 40+ who are inactive and overweight/obesity. This is a high-risk group who would benefit from a PA intervention for cancer prevention and who would benefit from an adaptive intervention because of the idiosyncratic and dynamic nature of PA that is pronounced within this group. Assessments will be conducted at baseline, 6, and 12-months using a hip-worn ActiGraph for assessing minutes/week of MVPA, as justified by guidelines. Implications: This research is highly significant because our intervention would be the first scalable PA intervention squarely grounded in SCT with N-of-1 adaptation driven by a mathematical dynamical model version of SCT. Further, favorable results would justify use of our COT methods for other complex and highly idiosyncratic and dynamic behaviors such as weight management, smoking, or substance abuse. Finally, our work should improve understanding of engagement with digital health tools. This research is highly innovative as we would be the first to conduct a COT and to empirically evaluate its utility in an RCT.

A wide range of technologies, such as smartphones, wearables (e.g., Fitbit, Apple Watch), and medical devices use alerts to inspire actions of users. Potentially useful alerts come at the cost of alert fatigue whereby individuals ignore alerts over time. For example, several physical activity interventions use alerts to inspire activity; notifications work initially but with diminished efficacy over time. Ignoring alerts is problematic in a variety of domains. For example, notification fatigue reduces the potency of interventions (e.g., notifications to inspire walking) and can be a safety risk in other areas such as in hospitals where notification fatigue can lead providers to ignore safety alerts (e.g., cross-drug interaction) provided by the electronic medical record. There is a need for novel solutions for reducing alert fatigue. Location, digital traces, and other data enable inference of states when a person would desire/need alerts, henceforth labeled just-in-time states, but more advanced analytics are needed. For example, a suggestion to walk (e.g., SMS saying, Want to go for a walk?) may only produce the desired outcome when a person's state (e.g., low stress) and context (e.g., no meetings, nice weather) align such that the person appreciates the notification (what we label receptivity) and can act on it (what we label opportunity). Estimating the likelihood that a given moment is a just-in-time state requires not only data but also an approach to manage the multivariate, dynamic, idiosyncratic, and multi-timescale nature of the problem. Returning to the walking example, stress, weather, and location change dynamically with each influencing the likelihood that a notification will inspire walking. In our work, results suggest idiosyncrasy in the factors that predict steps: some people walk more when stressed, others less, and still others are not influenced by stress. Further, just-in-time notifications cannot be viewed in a vacuum and, instead, are often part of a more long-term process, such as sustained engagement in a health behavior, thus making it a multi-timescale problem. The purpose of this work is to develop a just-in-time state estimation strategy and to stage a multi-timescale controller for walking as a concrete use-case of a control systems approach to counteract alert fatigue.