To determine the degree of dietary adherence or change in eating patterns, and demographic, psychosocial and study characteristics associated with adherence, in the Comparing Healthy Options in Cooking and Eating (CHOICE) Study.

Randomized controlled trial where women were randomized to one of three eating patterns: (i) Whole Foods, plant-based, macrobiotic-style (n 22); and Moderate Fat with (ii), and without (iii), 10 g of ground flaxseed added daily, which were combined (n 49).

A year-long intervention based on social cognitive theory, consisting of twenty-four class sessions involving hands-on cooking classes and behavioural sessions. Monthly 24 h food recalls were obtained and a psychosocial questionnaire was administered at baseline, 6 and 12 months.

Healthy, free-living, postmenopausal women.

A non-adherence score measuring all food servings out-of-compliance with eating pattern recommendations was specifically designed for the present study. Non-adherence scores decreased significantly (P < 0·05) in both groups to about 65 % during the adoption phase (first 4 months) and remained so during the 8-month maintenance period. Class attendance of the Moderate Fat group showed a trend towards significance as a predictor of adherence (P = 0·063). None of the other predictors (e.g. demographic and psychosocial factors) in a longitudinal regression model were significant.

Postmenopausal women were able to adopt and maintain significant changes in their eating patterns, including those on a demanding, near-vegetarian eating plan, suggesting that behavioural interventions with a healthy free-living population can be effective. The non-adherence score developed for the study provides an example of a means for evaluating eating pattern adherence to a dietary intervention.