Abstract
In a secondary analysis, we assessed the ability of dietary and physical activity surveys to explain variability in weight loss within a worksite-adapted Diabetes Prevention Program. The program involved 58 overweight/obese female employees (average age = 46 ± 11 years SD; average body mass index = 34.7 ± 7.0 kg/m2 SD) of four long-term care facilities who survey-reported liking and frequency of dietary and physical activity behaviors. Data were analyzed using a latent variable approach, analysis of covariance, and nested regression analysis to predict percent weight change from baseline to intervention end at week 16 (average loss = 3.0%; range-6% gain to 17% loss), and follow-up at week 28 (average loss = 2.0%; range-8% gain to 16% loss). Using baseline responses, restrained eaters (reporting liking but low intakes of high fat/sweets) achieved greater weight loss at 28 weeks than those reporting high liking/high intake (average loss = 3.5 ± 0.9% versus 1.0 ± 0.8% S.E., respectively). Examining the dietary surveys separately, only improvements in liking for a healthy diet were associated significantly with weight loss (predicting 44% of total variance, p < 0.001). By contrasting liking versus intake changes, women reporting concurrent healthier diet liking and healthier intake lost the most weight (average loss = 5.4 ± 1.1% S.E.); those reporting eating healthier but not healthier diet liking (possible misreporting) gained weight (average gain = 0.3 ± 1.4% S.E.). Change in liking and frequency of physical activity were highly correlated but neither predicted weight loss independently. These pilot data support surveying dietary likes/dislikes as a useful measure to capture dietary behaviors associated with weight loss in worksite-based programs. Comparing dietary likes and intake may identify behaviors consistent (appropriate dietary restraint) or inconsistent (misreporting) with weight loss success.
