Abstract
Background Gastric bypass (RYGB) can reverse type 2 diabetes (T2DM); however, a lack of randomized controlled trials (RCTs) comparing RYGB against medical therapy has limited widespread use. Barriers to conducting RCTs include: challenges recruiting informed patients willing to be randomized to surgical or non-surgical treatment who are not coerced by financial incentives or access to care, selection biases associated with standard recruitment methods, and high cost of RYGB. We assessed the feasibility of identifying, recruiting and randomizing a well-informed, population-based sample using a Shared Decision Making (SDM) approach. Based on prior research, we hypothesized a 100:1 ratio of contacted members to enrolled participants.
Methods Group Health members with T2DM and BMI of 30–40 kg/m2 are identified from electronic databases. We use a phased process to carefully screen participants: Call 1 assesses interest in obesity treatment and invites participation in SDM, which includes reviewing a DVD decision aid and booklet on treatment for obesity and diabetes. Subsequently, a nurse conducts a 45 minute SDM call to determine preferences regarding treatment. People remaining undecided about treatment after SDM (at equipoise) are invited to participate in the RCT. Written informed consent is obtained and 40 participants will be randomized to RYGB or behavioral intervention. Participants are followed for 1 year.
Results Recruitment started May 2011 and is ongoing. As of November 2011, 1220 members have been mailed, and 84% contacted. Of these, 21% completed Call 1 and agreed to SDM. Among 130 (11%) who’ve completed SDM, 64% choose diet/exercise; 3% drug treatment, 21% choose surgery, and 12% undecided. Among those who remained at equipoise, 47 (4%) were interested in the RCT, and 20 (1.6%) have been randomized. The randomized cohort has a mean age of 55.4, 60% female, 25% BMI 30–34.9, and 75% BMI 35–40.
Conclusions Using novel methods, we’ve demonstrated the feasibility of conducting a RCT of RYGB vs. behavioral intervention in an HMO setting, including enrolling participants with a BMI <35 (below current NIH standards for RYGB). We confirmed the ratio of screened to enrolled is ~50–100:1 and therefore large populations are needed to apply this approach to recruit a population in genuine equipoise without significant coercion.
