Presenting Author:

Sara Hoffman

Principal Investigator:

Bonnie Spring, Ph.D.

Department:

Preventive Medicine

Keywords:

weight loss, obesity, intervention, barrier, dropout

Location:

Third Floor, Feinberg Pavilion, Northwestern Memorial Hospital

PH52 - Public Health & Social Sciences

Cold feet at the threshold of weight loss treatment

A minority (10-20%) of those screened for weight loss treatment research reach randomization to a study condition. Among those eligible and consenting to participate, we examined differences between individuals who dropped out prior to and those who proceeded on to randomization. Our aim was to understand bias and limitations to generalizability arising from late-stage pre-randomization study withdrawal. Participants were Chicago-area residents with a BMI of 25-40 who indicated interest in a weight loss study, attended an orientation session, gave informed consent, and were deemed eligible. All were offered 6 months of treatment with a smartphone app and calls from a health coach. Each participant was randomized to a combination of other helpful strategies to augment treatment; there was no inactive group. These analyses compared consented eligible individuals who proceeded to be randomized and begin treatment (“Randomized”), versus those who chose not to proceed (“Non-Randomized”). Individuals were compared on age, BMI, smartphone type (Android vs. iPhone) and self-reported exercise coded from 0 (do not exercise or plan to) to 5 (exercise 4+ days a week). We hypothesized that Randomized participants would have more motivation and fewer barriers to weight loss (i.e., older, lower BMI, more likely to report regular exercise) than Non-Randomized participants. We also examined differences in smartphone ownership (iPhone vs. Android). Study candidates who attended an orientation session and consented as of December 2016 [n=931, m age=36.51 yrs, m BMI=31.88 kg/m2] were examined. Of these, 519 were Randomized and 412 were Non-Randomized. Independent-samples t-tests and Chi-square showed Non-Randomized participants were younger (M=33.96, SD=10.75) than Randomized (M=38.55, SD=10.81; p<.001). Non-Randomized (M=2.74, SD=.97) versus Randomized (M=2.96, SD=1.07) participants were less likely to exercise (p=.001). The groups did not differ on BMI (p=0.7) or smartphone type (p=.21). Treatment-seeking eligible and consenting weight loss volunteers who drop out prior to randomization tend to be younger and report less exercise than those who persist through randomization, but do not differ on BMI. Future studies should be aware of selective potential for dropout among these participants during screening and engage in participant-centered retention methods. By further exploring these characteristics, researchers may enhance generalizability of weight loss study findings.