成人术后急性中重度疼痛预测模型的系统评价

Prediction models for postoperative acute moderate-to-severe pain in adults: a systematic review

摘要:

目的 对已有成人术后急性中重度疼痛预测模型进行系统评价,为相关预测模型的构建、应用及优化提供参考。方法 全面检索Cochrane Library、PubMed、Embase、CINAHL、Web of Science、中国知网、维普、万方和中国生物医学文献数据库中发表的与成人术后急性中重度疼痛预测模型相关的文献,检索时限为建库至2023年9月1日。由2名研究人员严格按照纳入与排除标准独立筛选文献并提取资料,使用预测模型研究的偏倚风险评估工具PROBAST分析纳入研究的偏倚风险和适用性。结果 共纳入9篇文献,包括19个术后急性中重度疼痛预测模型；受试者工作特征曲线下面积为0.607～0.900； 3项研究采用机器学习法建模,2项进行了外部验证,9项研究偏倚风险均高；纳入预测模型前5的预测因子包括年龄、手术类型、术前疼痛、性别、术前使用阿片类镇痛药。结论 目前开发的综合性成人术后急性中重度疼痛预测模型的整体偏倚风险较高,模型预测性能有待进一步提升,未来研究应遵循相关规范进行模型建立及报告。此外,医疗服务机构应积极组建急性疼痛服务小组,根据临床实际选择适当的预测模型对高危患者进行个体化防控与诊疗,促进建立疼痛综合管理医院。

Abstract:

Objective: To systematically analyze and evaluate the prediction models of acute moderate to severe postoperative pain in adults, in order to provide references for the construction, application and optimization of relevant prediction models. Methods: A comprehensive search was conducted for original articles related to prediction models of acute moderate to severe postoperative pain in adults published in Cochrane Library, PubMed, Embase, CINAHL, Web of Science, CNKI, VIP, Wanfang and CBM. The search period was from the database establishment to August 1, 2023. Two researchers independently screened articles and extracted data in strict accordance with inclusion and exclusion criteria. PROBAST, a bias risk assessment tool for predictive model studies, was used to analyze the bias risk and applicability of included studies. Results: A total of 9 articles were included to construct prediction models for acute moderate to severe postoperative pain in adults, involving a total of 19 models. The area under the receiver operating characteristic curve ranging from 0.607 to 0.900; 3 studies used machine learning methods to build models, 2 studies were externally validated, and the risk of bias was high in all 9 studies. The top 5 predictors were age, type of operation, preoperative pain, gender, and preoperative use of opioid analgesics. Conclusion: The comprehensive prediction model of acute moderate to severe postoperative pain in adults developed at present has a high risk of overall bias, and its prediction performance needs to be further improved. Future research should follow the relevant norms to establish and report the model. In addition, medical service institutions should actively set up acute pain service teams, select appropriate predictive models according to clinical practice for individualized prevention and treatment of high-risk patients and promote the establishment of integrated pain management hospitals.