“CATCH-IT Reports” are Critically Appraised Topics in Communication, Health Informatics, and Technology, discussing recently published ehealth research. We hope these reports will draw attention to important work published in journals, provide a platform for discussion around results and methodological issues in eHealth research, and help to develop a framework for evidence-based eHealth. CATCH-IT Reports arise from “journal club” - like sessions founded in February 2003 by Gunther Eysenbach.

Sunday, November 8, 2009

Web-based Weight Loss in Primary Care: A Randomized Controlled Trial

Bennett GG, Herring SJ, Puleo E, Stein EK, Emmons KM and Gillman MW. Web-based Weight Loss in Primary Care: A Randomized Controlled Trial. Obesity (2009) doi:10.1038/oby.2009.242

Full text: click here

Evidence is lacking regarding effective and sustainable weight loss approaches for use in the primary care setting. We conducted a 12-week randomized controlled trial to evaluate the short-term efficacy of a web-based weight loss intervention among 101 primary care patients with obesity and hypertension. Patients had access to a comprehensive website that used a moderate-intensity weight loss approach designed specifically for web-based implementation. Patients also participated in four (two in-person and two telephonic) counselling sessions with a health coach. Intent-to-treat analysis showed greater weight loss at 3 months (-2.56 kg; 95% CI -3.60, -1.53) among intervention participants (-2.28 +/- 3.21 kg), relative to usual care (0.28 +/- 1.87 kg). Similar findings were observed among intervention completers (-3.05 kg; 95% CI -4.24, -1.85). High rates of participant retention (84%) and website utilization were observed, with the greatest weight loss found among those with a high frequency of website logins (quartile 4 vs. 1: -4.16 kg; 95% CI -1.47, -6.84). The intervention's approach promoted moderate weight loss at 12 weeks, though greater weight loss was observed among those with higher levels of website utilization. Efficacious web-based weight loss interventions can be successfully offered in the primary care setting.

Additional resources to assess study:
1. CONSORT for Non Pharmacological Treatments: click here
2. STARE-HI (Talmon J et al. International Journal of Medical Informatics
Volume 78, Issue 1, January 2009, Pages 1-9)


  1. My overall question would be, what is novel about this paper?

    None of the outcomes are reported as significant. As a result, it is challenging to identify what the paper tells us that would be considered new about weight loss interventions in an online environment.

  2. I found the use of the weight loss coach and a raffle to increase website use interesting.

    Though I understand the value of the face to face meeting and phone meetings with the weight loss coach for issues of attrition, I wonder really how applicable this approach is if this program is roled out as a large scale endeavour.

    I found the raffle interesting as it gave the sense along with the weight loss coach of a study that is looking to determine the efficacy of their approach vs. the effectiveness.

  3. One issue I have with this study is the relatively small sample size. The study required a minimum of 100 participants to minimize the play of chance (statistical power of 80%). The authors managed to get 101 participants out of 390 potentials. At the end of the study, because of attrition, they were left with 85 participants. Doesn't this affect the validity of their results and reduce statistical power?

  4. It's interesting to note that more web logins equated to better health outcomes. You would think that if they are on the internet more often, the participants would be doing less exercise. This may show that the website usage actually has some sort of motivating factor to encourage people to stick to the regimen. Perhaps, even more motivating than the life coach.

  5. The randomized control trial is the strongest type of design that the authors could have used in the quantitative type studies. I think the authors covered many important points with respect to the result. However, the only critique I have is regarding the issue of contamination bias at the patient level. Since the participants for both control and intervention groups were recruited from the same physician office, the members of the control group might inadvertently exposed to the intervention by the intervention group, which this bias potentially could minimizes the difference in outcomes between the two groups.

  6. I agree with Marjan, but wouldn't the issue rather be one of 'bias towards null', in which the physicians (all from the same office) would be contaminated in how they treated the control group?

    A few more questions:

    1. How did they identify their obesogenic behaviour change goals (they just say that they are empirically supported but offered no references)

    2. The weight difference between the intervention and control arms seems quite large (101 kg s.d. 15.4 vs 97.3 kg s.d. 10.9). Could this have affected the results?

    3. The intervention is also multi-modal. The question arises, therefore, regarding the causitive agent. Was it the web-based component or the fact that the intervention group patients had coach support? (the authors do, however mention this - "we were unable to isolate the independent contribution of discrete intervention components")

  7. What self-monitoring metrics were the participants imputing onto the website?

    I wonder if there would be a significant difference if the intervention only consisted of the website without the 4 counseling sessions with a health coach.

  8. I agree with Andrew, since the intervention group consisted of not ‘just’ the website as an intervention, they were also provided with a ‘health-coach’. However the other group of participants did not receive any such service and were only provided with a copy of the “Aim for a Healthy Weight” materials.

    My concern is that the health coach may have had an ‘external’ influence on the intervention group. This raises another concern for me, because my understanding of a web-based intervention of this type would be either of the following:

    1. Same consultation service is provided in both web-based and face-to-face where the study would determine, if the web-based interaction about personal “embarrassing” details maybe easier to share.

    2. Or, the same service is provided in both scenarios and the study’s objective would be to determine whether providing guidance and materials online would make it easier for participants to achieve better results (by participating at any time they want etc…)

  9. This type of study is similar to the one I evaluated in the first presentation. Comparing this study to STARE-HI—Statement on reporting of evaluation studies in Health Informatics, I think this report was well done.

    The observation I have is that table 3 indicated a gradual reduction in week 8-12. The participant was permitted to select new obesogenic behavior change goals at week 6. Did this have an impact on the outcome or login? What caused the reduction in web login? It would be interesting to know when the raffle for the gift certificates where drawn to determine if this also impacted the web login.

  10. I liked how the authors in their discussion section compare this study with the ones previously conducted to assess a similar intervention. However, am still not convinced with the inclusion of health coach in their study design, although on pg.4 it is mentioned that the intervention which did not include coach support also produced a similar magnitutde (2.6kg at 3-month follow-up) of weight losses observed in the present study. It is not clear whether the sessions with the 'health coach' was mandatory for the intervention group. It would be interesting to see the impact of keeping the human coaching as an optional factor on efficacy and results in primary care setting.