2020年武汉市MSM“互联网+HIV自检服务平台” HIV感染个体风险预测

Individual risk prediction of HIV infection among men who have sex with men on the internet-based HIV self-testing platform in Wuhan, 2020

目的 建立适合武汉市MSM人群线上使用的风险评估模型,为后续开发电子化风险评估工具提供科学基础。方法 通过武汉市MSM人群“互联网+HIV自检服务平台”进行问卷调查。2020年6月1日—12月31日期间获得合格调查问卷2 843份。使用列线图和专家评估法建立风险评估模型,根据受试者工作特征曲线和曲线下面积对模型进行选择,并用C指数对模型预测能力进行评估。结果 参与互联网线上自检的MSM人群HIV检出率为1.8%。逐步logistic回归构建的模型预测能力较好且预测变量较少,最终筛选出的危险因素是近6个月HIV阳性性伴情况、近6个月有性病感染症状或被诊断为性病、近6个月性角色、文化程度和月经济收入,根据筛选出的危险因素建立的列线图C-指数为0.790。结论 本研究构建的MSM人群HIV感染个体风险评估模型具有较好的区分能力和校准度,可为后续有针对性地干预提供科学依据。

Objective To establish a risk assessment model suitable for online use among men who have sex with men(MSM), and to provide a scientific basis for the subsequent development of electronic risk assessment tools. Methods A questionnaire survey was conducted through the Wuhan MSM internet-based HIV testing and intervention platform.A total of 2 843 valid questionnaires were obtained between June 1 and December 31, 2020.We created risk assessment models by using Nomogram and Delphi method.Suitable models were selected based on receiver operating characteristic curves and area under the curve, and the predictive ability of the models was evaluated using C index. Results The HIV detection rate of MSM who participated in online self-testing was 1.8%.The stepwise logistic regression model was chosen since it had higher predictive power and fewer predictive variable.The final predictive risk factors were the partner status of HIV positive MSM, symptoms of sexually transmitted diseases(STD) or confirmed STD, sexual roles in the most recent 6 months as well as education level and monthly income.The C-index of Nomogram graph was 0.790. Conclusions The individual risk assessment model of HIV infection in MSM established in the study has decent discriminatory power and calibration, which could provide scientific basis for subsequent targeted interventions.