目的探讨广西壮族自治区(广西)医疗机构法定传染病报告质量影响因素,并比较log-binomial回归与logistic回归模型估计关联强度的差异。方法采用多阶段分层抽样方法抽取广西县级以上医疗机构为调查对象,开展现场调查收集信息,在R v3.3.3中拟合log-binomial回归和logistic回归模型。结果共抽查法定报告传染病2 458例,平均报告率为95.08%,及时报告率为97.74%,报告卡填写完整率为77.60%,准确率为61.24%,网络直报录入信息一致率为95.27%,身份证填报完整率为75.59%。多变量log-binomial回归分析结果表明认真开展自查工作(PR=1.03,95%CI:1.01~1.05)和按要求开展培训(PR=1.08,95%CI:1.02~1.15)能有效促进法定传染病报告率的提高;设置项目齐全的门诊日志(PR=1.21,95%CI:1.07~1.37)、认真开展自查工作(PR=1.09,95%CI:1.03~1.14)、建立奖惩制度(PR=2.03,95%CI:1.49~2.78)和按要求开展培训(PR=1.18,95%CI:1.02~1.37)均能有效促进报告卡完整率的提高。在定性判别影响因素对结局事件发生概率影响时,logistic回归和log-binomial回归结果基本一致,但结局发生频率和其在比较组间差值每增加1.00%,logistic回归估计值OR相较于PR分别增加高0.65%(95%CI:0.34%~0.95%)和1.31%(95%CI:0.20%~2.41%)。结论广西县级以上医疗机构法定传染病报告质量仍有待提高,进一步改进院内自查方法,加强培训工作,落实奖惩制度,规范设置诊疗日志,加强医务人员传染病报告法律意识,是提高报告质量的重中之重。此外,log-binomial回归应被推广应用于横断面或队列研究中定量估计暴露与结局变量的关联强度。 Objective To investigate the influencing factors on the quality of notifiable infectious diseases reported by medical institutions in Guangxi Zhuang Autonomous Region (Guangxi), and to compare the differences in the correlation between log-binomial regression and logistic regression models. Methods A multistage stratified sampling method was used to select medical institutions above the county level in Guangxi as the survey subjects. Field surveys were conducted to collect information and fitted log-binomial regression and logistic regression models in R v3.3.3. Results A total of 2 458 cases of statutory infectious diseases were randomly selected. The average reporting rate was 95.08%, the timely reporting rate was 97.74%, the report card complete rate was 77.60% and the accuracy rate was 61.24% , ID card complete rate of 75.59%. Multivariate log-binomial regression analysis showed that serious self-examination (PR = 1.03, 95% CI: 1.01-1.05) and training required (PR = 1.08, 95% CI: 1.02-1.15) (PR = 1.21, 95% CI: 1.07-1.37). The self-examination work was carried out conscientiously (PR = 1.09,95% CI: 1.03-1.14) and the reward and punishment system = 2.03, 95% CI: 1.49-2.78) and training on demand (PR = 1.18, 95% CI: 1.02-1.37) could all contribute to the improvement of the completeness of the report card. Logistic regression and log-binomial regression results were basically the same when qualitatively determining the impact of influencing factors on the probability of outcome, but the frequency of outcomes and its increase of 1.00% Increased by 0.65% (95% CI: 0.34% -0.95%) and 1.31% (95% CI: 0.20% -2.41%) respectively. Conclusion The reporting quality of legal infectious diseases in medical institutions above the county level in Guangxi still needs to be improved, the method of hospital self-examination is further improved, the training work, the system of reward and punishment, the setting of medical logbooks, and the legal awareness of infectious diseases reporting by medical staffs are to be improved. Top priority. In addition, log-binomial regression should be applied to quantitatively estimate the association between exposure and outcome variables in cross-sectional or cohort studies.