Optimizing Digital Interventions: The Multiphase Optimization Strategy (MOST) Way
Workshop length: Full-day
Presenters:
Linda M. Collins, Ph.D., School of Global Public Health, New York University
Inbal (Billie) Nahum-Shani, Ph.D., Institute for Social Research, University of Michigan
Description
Advances in digital technologies have created unprecedented opportunities to deliver effective and scalable behavior change interventions. Two-arm evaluation randomized controlled trials (ERCTs) provide an excellent way to determine whether a digital intervention package is effective. However, this approach is less helpful in providing empirical information that can be used to optimize the intervention to achieve intervention EASE, a strategic balance of effectiveness, affordability, scalability, and efficiency. In this workshop an innovative methodological framework for optimizing behavioral interventions, the multiphase optimization strategy (MOST), will be presented. MOST is based on ideas inspired by engineering methods, which stress both ongoing improvement of products and careful management of research and implementation resources. MOST is a comprehensive strategy that includes three phases: preparation, optimization, and evaluation. A substantial amount of time will be devoted to introducing three types of optimization RCTs (ORCTs) and explaining how they can be employed in the context of MOST to optimize digital interventions. These ORCTs are the factorial experiment, the sequential multiple assignment randomized trial (SMART), and the micro-randomized trial (MRT). A variety of case studies will be used for illustration and time will be reserved for open discussion of how the concepts presented can be applied in the research of attendees.
Learning objective 1: Understand how the multiphase optimization strategy (MOST) can be used to optimize digital interventions.
Learning objective 2: Understand the fundamentals of factorial experiment, sequential multiple-assignment randomized trials (SMARTs) and micro-randomized trials; and be able to select the appropriate ORCT when using the MOST framework to optimize digital interventions.
Designing sophisticated and interactive digital interventions, without coding and at little to no cost: The Computerized Intervention Authoring System v. 3.0
Workshop length: Full-day
Presenters:
Steven J. Ondersma, PhD, Michigan State University
Jordan M. Braciszewski, PhD, Henry Ford Health
Amy M. Loree, PhD, Henry Ford Health
Eva Noack, MD, PhD, University Medical Center Göttingen
Frank Mueller, PhD, Michigan State University and University Medical Center Göttingen
Philipp Geisler, Gesellschaft für wissenschaftliche Datenverarbeitung mbH Göttingen (GWDG)
Description
The Computerized Intervention Authoring System (CIAS; www.cias.app) is an NIH-funded, open-source, non-commercial platform that enables researchers to easily develop, edit, and share sophisticated interactive content without coding of any kind. Interventions built with CIAS deploy as cross-platform compatible web/mobile web apps of any duration, with easy personalization, optional animated narrators who speak aloud in over 40 languages, and instant translation. Researchers can either use the hosted platform to build and deploy their interventions (without worrying about the technicalities of platform maintenance) or deploy the platform on their server.
This workshop will proceed in four phases. First, we will highlight the advantages of mobile web apps and no-code Software as a Service (SaaS) platforms. Second, we will introduce attendees to CIAS 3.0, highlighting key features and capabilities, followed by independent practice while the presenters circulate and provide guidance. This phase will include a review of embedded intervention templates designed with input from top researchers using CIAS. Third, we will consolidate learning from the prior section by breaking into small groups that will collaboratively develop the beginnings of a specific intervention of their choosing. (An additional breakout group will be led by a software developer for others interested in learning more about the code and possible collaboration on new features.) Fourth and finally, the presenters will share their experience in disseminating and implementing CIAS-developed applications in healthcare settings. Workshop attendees will leave with the information and skills needed to start developing their own custom digital interventions using CIAS, including access to templates, training resources, and technical assistance.
Key learning objectives:
- Workshop participants will understand the implications of a no-code, open-source digital intervention development platform and how it can accelerate research.
- Participants will understand how to access and use CIAS to develop any kind of digital intervention, in any language, and how to use additional features like tailored SMS, aggregate data visualization, and secure live chat.
- Participants will learn how to access ongoing CIAS support and how to collaborate on intervention development with other teams.
*Registered participants are requested to bring their own laptop to be able to work on CIAS platform during the workshop.
Images below are just small examples of the CIAS dashboard and what you can create with it.
Hands-on workshop on digital biomarkers: Passive sensing in daily life
Workshop length: Full-day
Presenters:
Prof. Dr. Tobias Kowatsch, School of Medicine, University of St. Gallen, St. Gallen; Switzerland, Institute for Implementation Science in Health Care, University of Zurich, Zurich, Switzerland, Centre for Digital Health Interventions (CDHI), Department of Management, Technology, and Economics, ETH Zürich, Zurich, Switzerland
Dr. Marcia Nißen, School of Medicine, University of St. Gallen, St. Gallen, Switzerland
Dr. Mia Jovanova, School of Medicine, University of St. Gallen, St. Gallen, Switzerland
Marc-Robin Gruener, School of Medicine, University of St. Gallen, St. Gallen, Switzerland
Akshaye Shenoi, Singapore-ETH Centre, Singapore
Description
This workshop offers participants a focused exploration of digital biomarkers—quantifiable physiological or behavioral data collected through digital devices. Digital biomarkers are often collected passively, continuously, and non-invasively and offer new ways to explain and predict health outcomes outside the lab. These data sources are rapidly transforming health research by enabling remote monitoring and personalized, just-in-time adaptive interventions (e.g., digital therapeutics).
To this end, this workshop will: (a) overview guidelines on designing research that incorporates new digital biomarkers; (b) examine applications of digital biomarkers across three cases studies in metabolic health, female health, and mental health; (c) discuss lessons learned and opportunities for interventions; and (d) offer participants an exciting opportunity to collect and analyze data from wearables, i.e., Fitbits and continuous glucose monitors (CGM devices), during the workshop. Participants will also have an opportunity to apply new learnings from the workshop into an ongoing joint research project, thereby fostering collaboration within the community and inspiring new research on digital biomarkers.
Learning objectives:
- Understand guidelines on designing research that incorporates digital biomarkers.
- Learn new applications of digital biomarker research across three case study contexts: metabolic health, female health, and mental health.
- Understand opportunities and potential for digital health interventions, as well as challenges and lessons learned from past digital biomarker studies.
- Gain hands-on experience on digital biomarker data collection and analysis (i.e., from Fitbits and/or continuous glucose monitors (CGM devices).
Machine Learning for Mental Health
Workshop length: Half-day
Presenters
Darragh Glavin, Doctoral Student, Department of Electronic & Computer Engineering, University of Limerick, Ireland
Eoin Martino Grua, PhD, Assistant Professor, Department of Electronic & Computer Engineering, University of Limerick, Ireland
Description
This workshop will provide an in-depth discussion of how machine learning (ML) techniques, being part of the domain of artificial intelligence, can be applied to the domain of mental health. First, the presenters will discuss ML techniques and then focus on applications for predictive modeling (both predicting therapeutic outcome as well as short term developments for patients) and personalization of therapies. They will then use real-life case studies to illustrate these techniques that feature hands-on activities to provide attendees with a deeper understanding of ML.
Learning objectives
- Learn about multiple types of machine learning techniques.
- Learn to apply these machine learning techniques to the Health domain via the programming language Python.