常用艾滋病疫情估计和预测模型研究进展

目的 艾滋病疫情的准确估计和预测，能为艾滋病防控决策和资源配置提供重要的依据。本研究分析和比较了常用的艾滋病疫情估计和预测模型。方法 采用文献综述的方法分析和比较了Workbook模型、亚洲疫情模型（asian epidemic model，AEM）、Spectrum/估计与预测软件包（estimation and projection package, EPP）和传染病动力学模型（susceptible infected recovered model, SIR）对艾滋病疫情估计和预测的特点。结果 通过对归纳和总结Workbook模型、AEM、Spectrum/EPP模型和SIR模型的基本情况、模型原理、适用条件、所需指标、结果产出和劣势分析等，为模型的选择和应用提供了科学依据。结果发现当模型指标要求单一的模型，不能很好地处理各类人群之间的交叉导致的重复估计，导致预测不准确；而指标量要求过多和过全时，数据难以收集和估计，部分数据缺失或质量不佳都会严重影响结果准确性，应根据实际需求进行模型选择。结论 定期有效的对高危人群和普通人群开展艾滋病流行病学监测，获取有效的监测数据，才能够准确的估计和预测艾滋病疫情。

Advances on estimation and prediction models of AIDS epidemic

Objective To provide an important basis for AIDS prevention and control decision-making and resource allocation. Methods By literature review, this study analyzed and compared the commonly used AIDS epidemic estimation and prediction models, including Workbook model, AEM, Spectrum/EPP model and SIR model in estimating and predicting AIDS epidemic. Results By summarizing the principles, applicable conditions, required indicators, outcomes and disadvantages of these models, it provided a scientific basis for the selection of AIDS epidemic estimation and prediction models. The results showed that when the model index required a single model, it could not deal with the repeated estimates caused by the crossover among different groups, which led to inaccurate prediction. But when the index quantity was too much and too full, it was difficult to collect and estimate the data, and the accuracy of the results would be seriously affected by the missing or poor quality of some data. So model selection should be based on actual demand. Conclusion Regular and effective epidemiological surveillance of AIDS among high-risk groups and general population to obtain effective surveillance data can accurately estimate and predict the AIDS epidemic.