Computerized Feedback Effective for Short-Term Weight Loss

CHAPEL HILL, N.C., Aug. 14 -- Patients can lose weight over the short term by following the dictates of their terminals, with automatically generated computer feedback reinforcing adherence to a healthy diet and activity regimen.

For three months an automatic computer-generated feedback on weight loss behavior matched e-mailed behavioral advice from human counselors, according to Deborah F. Tate, Ph.D., of the School or Public Health at the University of North Carolina in Chapel Hill.

But then the human-generated weight loss counseling edged ahead of the machines, Dr. Tate and colleagues reported in the Aug. 14-28 issue of the Archives of Internal Medicine.

"From a public health perspective this is a promising direction to explore," Dr. Tate and colleagues wrote.

In the study, computer-generated feedback and human e-mail counseling were both better than no counseling at all. At three months, automated feedback yielded an average weight loss of 5.3 kg compared with 6.1 kg with human counseling and 2.8 kg with no counseling (P<0.001).

At six months, both behavioral counseling groups continued to help individuals lose weight, but only those who received human-generated advice lost significantly more weight than the control group (P<0.001). The computer-generated advice group lost 4.9 kg while the human counseling group dropped 7.2 kg and the control group lost 2.6 kg.

Likewise, both feedback methods promoted clinically significant weight loss (at least 5% of initial body weight) with 52% of the human counseling group, 34% of the computerized counseling group and 27% of the control group.

The 192 women and men, randomized to the three groups, who completed the study were on average 49 years old with a body mass index of about 32. The participants initially attended a small group meeting where they received recommendations for diet, exercise and behavioral changes to promote weight loss and were oriented to a Web site where they could access further information.

The counseling groups also completed a weekly electronic diary on the Web site regarding their weight and adherence to the program. Feedback was either generated immediately by computer from preprogrammed responses tailored to information entered in the diary or generated by a human weight loss counselor the participants had never met.

All three groups reported significant reductions in caloric intake during the study and initially increased physical activity that then declined after the first three months. However, the human counseling group tended to consume less fat calories than the other groups (P=0.004).

The difference in results between the two feedback methods may have arisen over time because automated messages could have repeated and the structure of the messages did not change whereas e-mails from the human counselors had no predefined structure or content, the authors said.

Also, the computer feedback group filled out the online diary less often and therefore received less feedback than those in the human counseling group, who received e-mails from the counselor checking up on progress in weeks they skipped.

The authors cautioned that the participants were fairly homogenous (mostly female and predominantly Caucasian), had an average of about four years of Internet experience, and fairly well educated.

Other behavioral counseling programs in which individuals meet in-person with a counselor typically produce about 9 kg of weight loss over six months. However, several well-conducted weight loss trials have found significant health benefits with just 5 to 6 kg of weight loss.

The study was sponsored by the SlimFast Nutrition Institute.