Abstract
BACKGROUND: Individualized preoperative information can enhance patient satisfaction. However, existing studies have methodological limitations, largely adopting a healthcare "supply-side" perspective and lacking a patient-centered approach. Moreover, systematic quantitative assessments of information preferences among refractive surgery patients remain scarce. SETTINGS: Department of Ophthalmology at a tertiary general hospital in Luzhou and a tertiary ophthalmic specialized hospital in Shenzhen, China. PARTICIPANTS: A total of 119 patients in the pilot survey phase, and 567 patients in the formal survey phase. METHODS: Attributes and levels were identified through literature review, qualitative interviews, expert panel consultation and importance ranking. An orthogonal design was generated using Ngene for pilot choice sets, with a D-efficient design subsequently optimized for the main survey. Model estimation was performed in Stata 17.0, commencing with a multinomial logit (MNL) model and a random parameters logit (RPL) model to capture unobserved preference heterogeneity. A latent class logit (LCL) model was subsequently estimated to identify preference-based subgroups. Attribute interaction effects were examined to explore potential complementarities and substitutabilities. Finally, Scenario prediction analysis were conducted to predict the uptake probabilities of alternative information packages. RESULTS: A total of 10,200 choice observations (425 patients) passed the consistency check. Psychological support, mode of information interaction feedback, and duration of information provision were identified as the core driving factors, and two heterogeneous latent subgroups were identified. Joint analysis of attribute interactions showed significant overall effects [χ (2) (21) = 78.17, p < 0.001], indicating complex complementary and substitution patterns, though individual interactions were not significant after correction. The optimal combination was identified as "in-depth information" + "standardized + personalized" + "music" + "illustrated manual + video explanation + WeChat push notification" + "real-time interaction" + "one day before surgery" + "<30 min." CONCLUSION: Patients value humanized care, efficient communication, and respect for their time. Clinical education should shift from a "one-size-fits-all" approach to individualized optimization, enhancing overall utility through optimized combinations. The ideal service model constructed in this study provides evidence-based guidance for optimizing preoperative education processes.