OP26 Development and validation of prediction models for moderate-to-severe chronic postsurgical pain in older patients: a multicentre prospective cohort study

<h3>Background and Aims</h3> As the global population ages, chronic postsurgical pain in older patients has become a mounting concern, yet predictive tools specific to this population remain underdeveloped. This multicentre, prospective cohort study aimed to identify risk factors for moderate-to-severe chronic postsurgical pain and develop predictive models to facilitate early clinical decision-making in older patients. <h3>Methods</h3> We enrolled patients aged 65 years and older who underwent non-cardiac and non-neurosurgical procedures from April 2020 to April 2022. Moderate-to-severe chronic postsurgical pain, the primary outcome, was defined as a numeric rating scale score ≥ 4 three months post-surgery. Preoperative, intraoperative, and postoperative prediction models were constructed using logistic regression analysis and externally validated. <h3>Results</h3> A total of 3,580 cases were analysed, with a 15.2% incidence of moderate-to-severe chronic postsurgical pain. Significant predictors included female sex, lack of preoperative pain education, surgery for malignant tumours, specific surgical specialties, open surgical technique, surgery duration ≥ 120 min, worst pain score within 1–7 days post-surgery, and pain-related anxiety during the same timeframe. The models demonstrated areas under the curve of 0.707, 0.721, and 0.739 in the training set and improved performance in the validation set. <h3>Conclusions</h3> The prediction models, characterised by good calibration and relatively high accuracy and specificity, provide a valuable tool for identifying elderly patients at high risk of moderate-to-severe chronic postsurgical pain. These findings offer a foundation for multidisciplinary risk assessment and tailored clinical decision-making by anaesthesiologists and surgical teams.