基于贝叶斯层次二元logistic回归探究主动发现策略对新疆维吾尔自治区肺结核病例遗漏风险的影响

梁智超; 王森路; 胥婉婷; 刘广超; 王小敏; 王新旗; 曹明芹

doi:10.3784/jbjc.202308020393

基于贝叶斯层次二元logistic回归探究主动发现策略对新疆维吾尔自治区肺结核病例遗漏风险的影响

Effect of proactive detection strategy on risk for underreporting of tuberculosis in Xinjiang Uygur Autonomous Region based on Bayesian hierarchical binary logistic regression modeling

摘要

摘要:
目的采用2014—2020年新疆维吾尔自治区（新疆）南北疆14个地州肺结核报告数据结合影响肺结核报告发病率的相关社会、经济、环境因素，构建贝叶斯层次二元logistic回归，探究主动发现策略对新疆肺结核病例遗漏风险趋势变化的影响，为后续的结核病的精准防控决策提供有力的证据支持。
方法收集影响肺结核发病风险的相关因素，使用地理探测器法中的因子探测器筛选重要协变量并将其纳入模型，构建新疆肺结核病例遗漏风险遗漏的贝叶斯logistic模型。应用模型对新疆肺结核期望发病率进行相应估计，探究主动发现策略实施后肺结核病例遗漏风险的变化趋势。
结果因子探测器结果显示，人均国内生产总值与对肺结核病例遗漏风险的贡献率最大（q值为0.668）；贝叶斯logistic模型结果显示，肺结核标化遗漏风险与人口密度、全年平均气温、全年平均日照时数总体呈现出正相关性，与农村人均年收入、农村居民参加基本医疗保险人数、城市居民参加基本医疗保险人数、城市人均年收入总体呈现负相关性；2014—2020年新疆肺结核遗漏风险比呈先上升后下降趋势，2018年新疆肺结核遗漏风险比上升至顶点；结果显示主动发现策略实施后（2018年），新疆肺结核登记率、遗漏风险比均呈现下降趋势，南疆肺结核登记率、遗漏风险比均高于北疆。
结论 2018年后，新疆肺结核登记率、遗漏风险比均呈现显著下降趋势，主动发现策略取得了一定的成效。南疆的肺结核病例遗漏风险较北疆高，南疆地区实施肺结核防控策略可为进一步研究肺结核防治策略，控制肺结核病疫情提供理论依据。

Abstract:
Objective To construct a Bayesian model based on reporting data of tuberculosis (TB) from 14 prefectures in northern and southern Xinjiang Uyghur Autonomous Region(Xinjiang) from 2014 to 2020 in combination with relevant social, economic and environmental factors affecting the report of TB incidence to analyze the impact of proactive detection strategy on the trend of TB underreporting risk in Xinjiang and provide a strong evidence-based support for the subsequent decision making for the precision prevention and control of TB.
Methods The risk factors affecting the TB incidence were collected, and important covariates were screened by using the factor detector in geo-detector method to construct a Bayesian logistic model for the estimation of true incidence of TB and evaluate the changes in the trend of TB underreporting after the implementation of proactive detection strategy in Xinjiang.
Results The factor detector showed that gross domestic product (GDP) per capita was associated with the greatest contribution to the risk for TB underreporting (q=0.668); Bayesian logistic model results showed that the standardized underreporting risk of TB was positively correlated with population density, average annual temperature and average annual sunshine hours, but negatively correlated with annual per capita incomes, the numbers of residents participating in basic medical insurance in both rural area and urban areas. The risk ratio of TB underreporting in Xinjiang showed increase first and then decrease from 2014 to 2020 and the peak was in 2018. The results showed that after the implementation of the proactive discovery strategy from 2018, the registration rate and risk ratio of TB underreporting showed decreasing trends in Xinjiang, and the registration rate and risk ratio of TB underreporting were higher in southern Xinjiang than in northern Xinjiang.
Conclusion After the implementation of the proactive detection strategy, the registration rate and risk ratio of TB underreporting in Xinjiang all showed decreasing trends, indicating the achievement in this field. The risk for TB underreporting was higher in southern Xinjiang than in northern Xinjiang, and it is recommended to focus southern Xinjiang in TB prevention and control, and the implementation of TB prevention and control strategy in southern Xinjiang can provide theoretical basis for further research and practice of TB prevention and control.

HTML全文

参考文献(21)

施引文献

资源附件(0)