Ambulatory computer-assisted behavior therapy for obesity: An empirical model for examining behavioral correlates of treatment outcome

Abstract A behavioral treatment program for obesity was implemented in the natural environment through the use of hand-held computers. Participants were 40 mild-to-moderately overweight females, randomly assigned either to computer-assisted behavior therapy alone or to computer-assisted behavior therapy augmented by participation in a support group. Participants in the computer-assisted behavior therapy alone condition were given varying instructions regarding how many days per week to use the computer program (either 4 or 7 days per week). In addition to implementing treatment, the computer program was used to unobtrusively conduct a “behavioral audit” of each participant's frequency of utilization of each program feature, and to maintain a log of all self-report data on caloric intake and exercise. The treatment outcome data showed that participants who were instructed to use the program 7 days per week lost the most weight and that participation in a support group was not critical to treatment effectiveness. Also of interest was the fact that behavioral indicators of adherence to the treatment regimen had much higher correlations with total weight loss than did self-reported caloric intake or self-reported exercise. The behavioral indicators most highly correlated with weight loss were daily frequency of making self-reports, daily frequency of viewing motivational statements, daily frequency of viewing goal-related feedback, and total days of program use. Computer-assisted behavior therapy conducted in the natural environment is viewed as providing a powerful means for examining behavioral correlates of treatment outcome.