Influencing factors and prediction algorithms to estimate public preference for painless medical procedures in China: A national cross-sectional study

Objectives To improve the medical treatment process, the concept of comfortable medical care has been proposed. This concept has been fully implemented in developed countries, but there are still many restrictions on its promotion and implementation in developing countries. This study investigates public preferences about comfortable medical care and identifies the factors influencing these in China. Design A cross-sectional survey using an online questionnaire. We used multivariate logistic regression analysis to identify the factors influencing preferences for comfortable medical care and developed a prediction model of preferences for four types of this care among the general population. Setting Peking University Third Hospital, China. Participants Members of the general public with no medical or biomedical background. Primary and secondary outcome measures The measures included demographic information and attitudes towards four types of comfortable medical services. The primary outcome was attitude towards general anaesthesia. Results Overall, 5330 people participated in the survey. Nearly 80% would choose painless gastroenteroscopy, labour analgesia and postoperative analgesia. Interestingly, just 43.9% preferred general anaesthesia over regional anaesthesia. The impact of general anaesthesia on learning and memory (69.8%) was the most worrying issue. Younger people with higher education levels and incomes, and living in cities, preferred painless medical services. In the prediction model for general anaesthesia, the areas under the receiver operating characteristic curves were 0.790 (95% CI 0.776 to 0.804) in the development group and 0.777 (95% CI 0.754 to 0.799) in the internal validation group. Conclusion The Chinese public has a good level of acceptance of comfortable medical care, although there are still many obstacles to its popularisation and promotion. Our prediction model could be used to screen the population for targeted popularisation work to expand the implementation of this approach.