Associations of internet website use with weight change in a long-term weight loss maintenance program

Abstract

Background: The Weight Loss Maintenance Trial (WLM) compared two long-term weight-maintenance interventions, a personal contact arm and an Internet arm, with a no-treatment control after an initial six-month Phase I weight loss program. The Internet arm focused on use of an interactive website for support of long-term weight maintenance. There is limited information about patterns of website use and specific components of an interactive website that might help promote maintenance of weight loss.

Objective: This paper presents a secondary analysis of the subset of participants in the Internet arm and focuses on website use patterns and features associated with long-term weight maintenance.

Methods: Adults at risk for cardiovascular disease (CVD) who lost at least 4 kilograms in an initial 20-week group-based, behavioral weight-loss program were trained to use an interactive website for weight loss maintenance. Of the 348 participants, 37% were male and 38% were African American. Mean weight loss was 8.6 kilograms. Participants were encouraged to log in at least weekly and enter a current weight for the 30-month study period. The website contained features that encouraged setting short-term goals, creating action plans, and reinforcing self-management habits. The website also included motivational modules, daily tips, and tailored messages. Based on log-in and weight-entry frequency, we divided participants into three website use categories: consistent, some, and minimal.

Results: Participants in the consistent user group (n = 212) were more likely to be older (P = .002), other than African American (P = .02), and more educated (P = .01). While there was no significant difference between website use categories in the amount of Phase I change in body weight (P = .45) or income (P = .78), minimal website users (n = 75) were significantly more likely to have attended fewer Phase I sessions (P = .001) and had a higher initial body mass index (BMI) (P < .001). After adjusting for baseline characteristics including initial BMI, variables most associated with less weight regain included: number of log-ins (P = .001), minutes on the website (P < .001), number of weight entries (P = .002), number of exercise entries (P < .001), and sessions with additional use of website features after weight entry (P = .002).

Conclusion: Participants defined as consistent website users of an interactive behavioral website designed to promote maintenance of weight loss were more successful at maintaining long-term weight loss.

Trial registration: NCT00054925; http://clinicaltrials.gov/ct2/show/NCT00054925 (Archived by WebCite at http://www.webcitation.org/5rC7523ue).

Figure 1

Use of various website features for consistent (n = 212), some (n = 61), and minimal (n = 75) use categories shown as median (square) with first (diamond) and third (triangle) quartiles

Figure 2

Phase II weight regain (kg) for each website use category shown as 95% confidence interval for the mean (two-tail P value testing for differences based on linear regression analysis adjusted for race, sex, education, income and age, P = .003)

Figure 3

Proportion of initial weight loss regained for each website use category shown as 95% confidence interval for the mean (two-tail P value testing for any differences based on linear regression analysis adjusted for race, sex, education, income, and age, P = .001)

Figure 4

Proportion of participants at least 4 kg below initial weight for each website use category (two-tail P value testing for any differences based on logistic regression analysis adjusted for race, sex, education, income, and age, P = .002)