|Fall 2015||Return to Outlook main page »|
Optimization of Behavioral Interventions SIG Talks with Bonnie Spring about her Study of Weight Loss Program Components
Kari Kugler, PhD, Optimization of Behavioral Interventions SIG co-chair
The Society of Behavioral Medicine's (SBM) new Optimization of Behavioral Interventions Special Interest Group (OBI SIG) recently interviewed Bonnie Spring, PhD, ABPP, director of the Center for Behavior and Health at Northwestern University's Feinberg School of Medicine.
The OBI SIG talked with Dr. Spring about her ongoing Opt-In research study, which aims to determine what combination of different weight loss program components is the most effective for overall weight loss.
OBI: You are halfway through your Opt-In study, which uses the multiphase optimization strategy (MOST). What made you decide to develop an optimized intervention?
Spring: It just made tremendous sense to me to go about developing an obesity treatment package from the ground up by balancing the efficacy and cost of different candidate treatment components. It has always bothered me that we tend to offer "kitchen sink" treatment packages as behavioral interventions because, most of all, we want our treatments to produce significant improvement. The problem is that this approach often leaves us with unwieldy, burdensome, very costly treatments that are difficult to scale up to reach more of the population. And despite the habit of saying we're going to conduct dismantling studies to streamline and prune back our bundled treatment packages, that research is rarely done. Because MOST uses data to drive decision making about the treatment design, it has always impressed me as a transparent means to progress toward evidence-based practice.
OBI: How difficult was it for you to understand the fundamental differences between a randomized clinical trial (RCT) and the experimental design you needed to optimize the intervention, in this case a factorial experiment?
Spring: My PhD is in personality and developmental psychology, and my doctoral training was in experimental psychopathology. The methodological tradition in these fields actually does involve the conduct of factorial experiments. I first encountered the RCT when seguing to health psychology by way of psychopharmacology. My first trials treating smoking or weight gain with drug or placebo felt very much like factorial experiments-with treatment set to on versus off. It wasn't until I discovered that patients tended to relapse whenever I withdrew drug treatment that I realized I needed to switch to behavioral treatments to achieve greater maintenance. RCTs involving behavioral interventions felt very strange to me because the intervention seemed to bundle different treatment components together rather than independently turning them on or off. I still try to perform factorial experiments when developing a behavioral treatment. I prefer not to bundle components together into a treatment package until after I've studied them as independent factors and learned their effects. So, the MOST factorial experiment approach feels very right-minded to me.
OBI: What are some considerations when it comes to the implementation of a factorial experiment? How has it been for your research staff to manage 32 experimental conditions?
Spring: Because of the complexity of having so many conditions, the staff follow an online set of prompts for the different experimental conditions. They do an excellent job of attaining treatment fidelity. Needing to be guided by the prompts makes intervention delivery a little more scripted than most interventions we do. On the other hand, I suspect that our interventionists are better able to maintain equipoise because it's more challenging for them to become attached to any single condition.
OBI: What advice would you give someone considering using MOST?
Spring: I would urge the person to specify their theoretical or conceptual model and to use it to delineate which treatment components to study. Relatedly, I would advise them to think through carefully which interactions are likely and conceptually meaningful so that the experimental design allows these to be interpretable.
OBI: How do you see MOST helping move science forward?
Spring: By helping us learn the cost-effectiveness of different intervention components for the average person and for different kinds of people, MOST can help us establish the science base needed to optimize treatments equitably to foster the health of the population.