Optimized Interventions for Prevention and Treatment in Behavioral Health
Understanding how control engineering principles can be used to optimize behavioral interventions is the dominant research theme in our laboratory. We are pursuing this topic in collaboration with Professors Linda Collins at Penn State University and Susan Murphy from the University of Michigan, with funding provided through grants K25DA021173 and R21DA024266. Specifically, our focus is on exploring the application of engineering process control concepts to the analysis, design, and implementation of adaptive, time-varying interventions in the field of behavioral health. Adaptive, time-varying interventions represent a novel and promising approach to prevention and treatment of chronic, relapsing disorders, such as alcoholism, drug abuse, and cigarette smoking. In an adaptive intervention, different dosages of prevention or treatment components are assigned to different individuals across time, with dosage varying in response to the needs of the individual. Conceptually, adaptive time-varying interventions represent closed-loop control systems where intervention dosages (i.e., manipulated variables) are determined by decision rules (i.e., control laws) based on the values of a participant’s key characteristics (also referred to as tailoring variables, or in process control terminology, controlled variables). Effective time-varying adaptive interventions share similar goals to well-designed process control systems in that they both seek to:
- Reduce negative effects
- Increase compliance and intervention potency
- Reduce waste
Significant challenges exist in both dynamic modeling and the application of control system design to the problem, which we are seeking to investigate at this time. Topical areas that we are exploring include:
- Robust decision frameworks for delivering time-varying adaptive behavioral interventions for prevention and treatment based on Model Predictive Control,
- Dynamic modeling of behavioral interventions for weight loss and body composition change. This includes activities on modeling an intervention to prevent excessive gestational weight gain (with D. Downs, Kinesiology, Penn State; D. Thomas, Mathematical Sciences, Montclair State University; K. Hall and C. Chow, National Institute of Diabetes and Kidney Diseases),
- Dynamic modeling and optimization of a low-dose naltrexone intervention for fibromyalgia (with J. Younger, Systems Neuroscience and Pain Lab, Stanford),
- Modeling smoking activity and cessation as a closed-loop dynamical system (with T. Walls, Psychology, U of Rhode Island; M. Fiore and T. Baker, Center for Tobacco Research and Intervention, University of Wisconsin).
As a result of the training component of the K25 grant there has been significant interaction with faculty in the Department of Psychology and the Prevention Research Center at Arizona State University; this includes Professors Irwin Sandler, Laurie Chassin, Steve West, Felipe Gonzalez Castro, David MacKinnon, and Roger Millsap.
Weigh-IT Plus, an interactive tool for simulating weight change as a result of diet and exercise, can be downloaded here.
The work of our laboratory in this area has been profiled in a May 2011 ASU Research Stories article entitled "Fighting addiction with engineering algorithms." An earlier ASU news article written in December 2007 describes the two NIH grants funding this work.
A handout (available in two-slide-per-page and four-slide-per-page formats) is available for download for the presentation "Optimized behavioral interventions: what does control systems engineering have to offer" given at the joint Prevention and Methodology seminar at Penn State University on October 6, 2010.
Handouts for talks at the 2011 Annual Meeting of the Society of Behavioral Medicine:
"Optimized behavioral interventions: what does dynamical systems and control engineering have to offer? pre-conference workshop: From Discovery to Public Health Impact: New Approaches to Developing, Testing, and Optimizing Behavioral Interventions (Tuesday, April 26, 2011).
"Robust optimal decision policies for adaptive time-varying interventions using Model Predictive Control" invited symposium: "Drawing from ideas in engineering and computer science to build better behavioral interventions." (Friday, April 29, 2011).
Handout for our talk at the 19th Annual Meeting of the Society for Prevention Research (SPR):
"A Novel Dynamical Systems Approach to Statistical Mediation Analysis" (co-authored with K. Timms, J. Trail, J.E. Navarro-Barrientos, M.E. Piper, and L. Collins), oral symposium entitled, "Idiographic Methods: Important Alternative Research Methods for the Behavioral Sciences." (Friday, June 3, 2011).
Lecture handout for the Fall 2011 BDE 598 class: 2011_BDE598_lecture.pdf
Handout for 2012 SRNT Pre-conference workshop on New Methods for Advancing Research on Tobacco Dependence: "Dynamical Systems Modeling using EMA Data: An Illustration from Smoking Cessation," (co-authored with K.P. Timms, J.B. Trail, M.E. Piper and L.M. Collins).
Handouts for the July 19, 2013 NCI - BRP Science of Research and Technology Branch Crosstalk webinar (2 slides/page and 4 slides/page) are available for download. A reference list is downloadable as well.
A list of recent publications and presentations regarding this topic can be found here.