Dietmate - Computerized Intervention for Weight Loss
Computer-Assisted Management of Obesity: A Preliminary Investigation
Albert Jerome, Ph.D.
Lee W. Frederiksen, Ph.D.
Candace B. Frederiksen, Ph.D.
Health Innovations, Inc., Reston, VA
This research was funded, in part, by a grant from the National Institute of Diabetes, Digestive, and Kidney Diseases (R43 DK42397). Please do not quote without permission.
Objective # 1. Test and refine components of the paper and pencil version through research with overweight subjects.
We recruited subjects to test the program. The purpose of this aspect of the research was fourfold: 1) to obtain initial data on the appropriateness of the caloric adjustment and exercise algorithms; 2) to obtain data on palatability and ease of preparation of menus; 3) to obtain feedback on the initial drafts of the supplementary manuals; and 4) to obtain feedback on a wide range of design and development issues in a focus group format.
This research continued for several months and over 50 subjects participated. During this period, the program was under constant development and the opinions and data provided by subjects resulted in extensive revisions to program algorithms, menus, and written materials.
Objective # 2. Assemble an Initial Computer Prototype Using Information Gained with the Paper and Pencil Research.
Based on refinements to the program made during the paper and pencil trials, an initial computer prototype (Beta-I) was assembled and a second version of support materials drafted. One indication of the impact of the paper and pencil research on program development is that we decided to completely revise the program menus on the basis of feedback from users suggesting that the food combinations, although meeting strict nutritional criteria, were not very palatable. This is one example of how both scientific (nutritional) information and consumer feedback guided program development.
Objective # 3. Assess Usability and Reliability of the Prototype with a Small Group of Subjects.
In preparation for a self-help trial, we recruited 10 subjects who had completed the paper and pencil trials to serve as initial test subjects for the computer prototypes. Subjects used the computers for varying periods of time ranging from a few weeks to several months. Subjects met every two weeks with a research assistant to upload data from the computer and to discuss their experiences and opinions related to treatment.
From this research we were able to ascertain that the program was easy to follow and the reliability of the units was acceptable. We also gained important information related to specific program components and made many program modifications as a result.
Subjects were 24 women 2 men recruited through newspaper advertisements for a weight loss research trial. Mean age of subjects was 41.1 years (SD = 7.3) and mean number of years of education was 15.1 years (SD = 1.9). All subjects had previous dieting experience with 56% reporting having been on 10 or more diets in the past. At pretreatment assessment, subjects were a mean of 21% over ideal weight (based on guidelines provided by U.S. Department of Agriculture and U.S. Department of Health and Human Services, 1990).
Prior to treatment, a 45-minute assessment was conducted to explain the program, obtain informed consent, collect demographic data, physical measures, and ratings of confidence and body satisfaction. Subjects received their computers, instructions for use, and supporting manuals in small groups of 8 - 12 and were asked to begin treatment. Treatment was run in a self-help manner without further group meetings or therapist-assistance. At treatment end, subjects were assessed on the same measures collected prior to treatment and they were administered a detailed structured interview that explored their impressions of the program and ideas for modifications and improvements.
The computer worked by tailoring a nutritionally balanced meal plan for each individual user at a caloric level appropriate to allow for weight loss of .5 - 1.5 lbs. weekly. Subjects entered their weight daily and caloric levels were adjusted periodically if weight loss was too rapid or too slow. Users selected from among 75 meals, and the computer provided nutritionally equivalent substitution lists for foods and allowed users to change quantities of individual food items.
Individualized exercise targets were provided for a wide range of activities. Following a one-week baseline where subjects recorded their usual exercise, the computer provided exercise targets that gradually increased up to an equivalent of 300 calories expended for each of three weekly exercise sessions. The computer prompted users for mealtimes, drinking water, and exercising, and also provided a daily summary screen showing progress toward meeting daily target goals for total calories, percent calories from fat, and calories expended exercising.
Four sets of variables were collected during this study including: demographic and historical variables; physical measures such as weight, body mass index (BMI), circumference measures, blood pressure, and body fat, ratings of program ease of use and program satisfaction and confidence in success; and a structured interview at treatment end to gather opinions concerning program modifications.
Results are presented for 18 subjects who completed 12 weeks of treatment. For weight and BMI t-tests revealed significant change from pretreatment to posttreatment, t(17) = 6.38, p<.0001; and t(17) = 6.30, p <.0001, respectively (see Table 1 below).
Pretreatment and Posttreatment Weight and BMI (Body Mass Index)
Paired samples t-tests were also used to evaluate changes from pretreatment to posttreatment on percent body fat, blood pressure, and circumference measures for waist, hips, and biceps, and body satisfaction ratings. Significant changes occurred for percent body fat, t(17) = 2.41, p < .05; diastolic blood pressure t(17) = 4.05, p < .01; biceps circumference, t(17) = 2.41, p < .05; and body satisfaction ratings, t(17) = 5.2, <.0001.
Finally, mean end of treatment rating of program satisfaction was 5.6 (SD = 1.4) on a seven point scale and program ease of use was 6.0 (SD = .98) on a seven point scale. Seventy percent of subjects rated themselves as being either 'extremely satisfied' or 'very satisfied with the program, and 63% of subjects rated the program as being either 'extremely easy' or 'very easy' to use.
Overall, the results of this trial indicated that the computerized treatment program was both effective and viewed as user friendly by subjects. Significant changes from pretreatment to posttreatment were seen for weight, BMI, percent body fat, diastolic blood pressure, biceps circumference, and ratings of body satisfaction. Moreover, treatment process variables suggested that subjects complied very well with the prescribed diet. Subjects also rated the program as being easy to use and satisfaction ratings were very high.