This is the first study to directly compare weight loss outcomes from an in-person behavioral weight control intervention with one delivered exclusively online. While weight losses of the in-person participants were superior, individuals enrolled in the Internet treatment programs lost clinically meaningful amounts of weight as well. Specifically, over half of Internet subjects lost at least 5% of their baseline weight. This amount of weight loss has been associated with reduced chronic disease risk (National Institutes of Health/National Heart, Lung, and Blood Institute, 1998
). It is important to note that although the weight losses achieved in the current study were smaller than those in the in-person program, they are among the highest achieved by an online weight control program reported to date (Gold et al., 2007
; Harvey-Berino et al., 2004
; Micco et al., 2007
). Further, this is the first study to examine internet delivered weight management that has included a significant proportion of minority participants, a group that has demonstrated modest response to available behavioral weight management technologies (Kumanyika, 2008
; West et al., 2007
) but for whom the burden of obesity is high (Ogden, 2009
). Finally, the weight losses achieved in the InPerson condition in the current study were substantial, comparable to the average 1-year losses reported in the Diabetes Prevention Program (Diabetes Prevention Program Research Group, 2002
) and the Look AHEAD trial (Wadden et al., 2006
). Thus, the weight losses against which the online condition was compared were among some of the best reported in the literature.
The addition of minimal in-person support did not significantly improve weight loss over and above online delivered program alone. Recent research by Micco and colleagues (2007)
also suggests that little advantage is conferred by including in-person contact within an online intervention. While other research has evaluated computer technology as an adjunct to face-to-face communication (Glasgow et al., 1999
; Gustafson et al., 1999
), the present investigation is only the second to evaluate this hybrid model for weight loss. Given the complexity of managing two forms of intervention delivery, and the burden of travel to periodic in-person meetings, coupled with the apparent absence of improved weight loss outcomes, it is harder to advocate for the future potential of this approach. However, it is reasonable to consider that while the in-person meetings added little to the weight loss outcomes in this study, periodic in-person support could be a valuable addition to a clinical intervention as they could allow for other types of medical monitoring.
The treatment conditions resembled each other in several ways. All groups reported similar changes in diet and physical activity behaviors, which may not be surprising given the well-established measurement error associated with these self-report measures (Scagliusi et al., 2003
). Also, attendance at group meetings and self monitoring did not differ between groups, although both self-monitoring and attendance tend to predict weight loss success in other studies (Acharya et al., 2009
; Boutelle & Kirschenbaum, 1998
). However, it is important to note that the lack of observed differences in self-monitoring and attendance may be related to the fairly high levels of compliance with both in all conditions. Subjects turned in 73% of self-monitoring journals and attended an average of 69% of their group meetings.
In contrast, the conditions did differ in perceived social support; social support is another aspect of behavioral obesity treatment that has been associated with better weight loss outcomes (Wing & Jeffery, 1999
). Perceived group support was significantly higher when the intervention was delivered in-person, suggesting that social support may be one possible mechanism by which the in-person format produced the superior weight losses observed. There were no significant differences reported for alliance with the weight loss counselor across conditions, indicating that online treatment does not attenuate the therapeutic relationship. Differences were only seen in relationship to perceived group support. Online participants attended weekly chat sessions that included all group members and a consistent counselor. This is identical to the format used to deliver in-person treatment. The obvious difference is the lack of face-to-face communication. It is important to point out however, that the addition of monthly in-person contacts did not significantly increase perceived group support for the Hybrid group. In fact, participants reported lower alliance scores with their in-person counselor than their on-line counselor. Having separate on-line and in-person counselors allowed for an evaluation of a public health model of internet weight loss delivery where participants could connect virtually across a national or even international network but have periodic local, in-person meetings run by a physician's practice, community or public health organization. However, based on the results of this study, evaluating strategies to facilitate and enhance a sense of group cohesion online is warranted, and adding some face-to-face contact is not likely to be a promising avenue. However, other studies have demonstrated the beneficial impact of “engineering social support” on weight loss outcomes (Kumanyika et al., 2009
; Wing & Jeffery, 1999
Study Strengths and Limitations
This study has several strengths, including a randomized design and a protocol-driven intervention that offered treatment goals and behavioral strategies to achieve these goals that were comparable across all study conditions, allowing the difference in delivery channel to be the only distinguishing feature between the on-line and in-person programs. Additional strengths included that counselors were similarly trained and supervised, objective measures of weight were obtained, and a large, racially-diverse subject pool was included. The study did not include a no-treatment control group, outcome data are based on a six month intervention and generalization to the broader population of overweight and obese individuals may be hampered by the high education level of participants and the small number of men included in the sample.