Development of intervention package to promote healthy lifestyle behaviors in women with polycystic ovary syndrome: a factorial randomized trial

Abstract Background Healthy lifestyle behaviors are recommended as a first-line management strategy for women with polycystic ovary syndrome (PCOS), yet most patients fail to meet the dietary and physical activity recommendations. Therefore, this study aimed to identify effective behavioral intervention components for inclusion in a practical lifestyle intervention package to improve healthy lifestyle behaviors among women with PCOS. Methods We recruited women with PCOS aged 18 years or older from gynecology clinics at a tertiary grade A general hospital in Hunan Province, Central South China. Guided by the Behavior Change Wheel (BCW) and the Multiphase Optimization Strategy (MOST) framework, we identified a core intervention component (health education) and three supplementary intervention components (goal-setting, feedback, and reminders [information reminders and environmental cues]). To examine the individual and combined effects of these intervention components, we designed a 2×2×2 full-factorial randomized trial with a 16-week intervention period. The primary outcomes were dietary behavior and physical activity behavior, measured using the Health-Promoting Lifestyle Scale (HPLS) at baseline (T0), post-intervention (T1), and the 3-month follow-up (T2). Effects were estimated using regression coefficients (β) and corresponding P values from the linear mixed-effects models, with statistical significance for the primary outcomes defined as P < 0.025 after Bonferroni correction. From the implementer perspective, incremental cost-effectiveness ratios (ICERs) were calculated to evaluate the cost-effectiveness of each intervention component (i.e., comparing the presence versus absence of each component), expressed as the incremental cost per one-point increase in the corresponding behavioral outcome score. Results A total of 160 participants (mean age: 28 years) were randomized at baseline. Of these, 150 (93.8%) completed the intervention, and 145 (90.6%) completed the 3-month follow-up visit. Goal-setting was associated with significantly greater improvements in dietary behavior at both time points (T1: β = 5.2, P < 0.001; T2: β = 4.2, P < 0.001), as was feedback (T1: β = 3.7, P < 0.001; T2: β = 4.6, P < 0.001). However, reminders were not associated with statistically significant improvements in dietary behavior at either time point (T1: β = 2.6, P = 0.04; T2: β = 2.1, P = 0.07). For physical activity, feedback was associated with significantly greater improvements at both time points (T1: β = 7.3, P < 0.001; T2: β = 8.7, P < 0.001), whereas goal-setting was associated with a statistically significant improvement at post-intervention (β = 5.6, P = 0.01), but not at follow-up (β = 4.9, P = 0.03). The reminder component was not associated with a statistically significant improvement in physical activity at either time point (T1: β = 3.8, P = 0.09; T2: β = 2.8, P = 0.21). Most two-way and three-way interaction effects were not statistically significant for either outcome. Regarding economic outcomes, for dietary behavior (score range: 7–35), the ICERs for goal-setting, feedback, and reminder components were CNY 58.2, 84.8, and 21.4 per one-point increase, respectively. For physical activity behavior (score range: 10–50), the corresponding ICERs were CNY 68.5, 33.4, and 11.4 per one-point increase. Conclusions Our findings suggest that health education combined with goal-setting constitutes the most cost-effective intervention package for long-term dietary management, while health education paired with feedback represents the most cost-effective intervention package for long-term physical activity management.