西南三省（川滇藏）细菌性痢疾（简称菌痢）持续流行，是威胁当地居民健康的首要甲乙类肠道传染病。然而该地区相关的流行病学研究很少，人们对菌痢持续流行的特征和机制认识存在不足。由于西南独特的自然环境和社会经济环境，其它地区的研究结论和防控经验无法直接套用。针对这一现状，本文基于2005-2014年间西南县级菌痢报告数据，首先通过描述性分析、圆分布分析，多角度揭示了菌痢的三间分布特征；通过系统回顾和荟萃分析建立了菌群数据库和突发疫情数据库，得到了菌群流行特征和突发疫情特征。随后，本文围绕空间格局、季节趋势和短期气象事件（暴雨洪涝）三条主线深入探讨，构建空间风险模型和时间序列模型探索了西南菌痢独特的时空流行特征及其机制，定量评估了暴雨洪涝对菌痢传播的影响，得出以下主要结论：1. 菌痢热点地区的空间位置和范围基本稳定，主要位于青藏高原东部边缘和西藏东南部的山区，这与社会经济发展水平相对落后、医疗卫生资源较少、较高环境温度、中低相对湿度、居民地距农田或水体更近、山区地形分布和藏缅族分布相关。2. 本文构建的空间风险模型以地理格网单元代替行政单元，对菌痢的发病风险进行了精细尺度空间化预测，更精准地发现了医疗卫生资源需求热点和一些可能被疾病监测忽视的区域。3. 四川南部和云南大部分地区菌痢发病高峰期比我国其它地区早1-2个月，这与当地独特的季节性干旱有关。然而干旱主要影响当地农村，对基础设施较完善的城区的影响有限。城区与高峰期相对较晚的西南地区有相似的菌痢季节趋势，主要受温度影响且发病高峰期较晚。4. 菌痢发病数与温度或蒸发量呈线性或近似线性正相关，然而部分地区气象因素的影响存在阈值效应。四川菌痢季节趋势相对较弱便是受温度阈值效应的影响，这可能与当地高温季防控得力有关。5. 暴雨洪涝是菌痢流行的危险因素。就健康风险而言，重度暴雨洪涝高于轻度，农村高于城区，山区高于非山区。即便一些基础设施较完善的城区，暴雨洪涝过后发病数也会明显上升，提示存在防控薄弱环节。不同地区疫情对暴露的响应快慢不尽相同，这可能与各地各异的自然环境和人造环境导致暴露-响应的具体过程不同有关。本文研究成果是对西南三省菌痢流行特征和机制认识的有效补充和完善，可为当地菌痢防控策略的制定、医疗卫生资源的优化调配和更好地应对未来气候变化提供参考和科学依据。
Bacillary dysentery is a primary intestinal infectious disease and has been epidemic in southwest China （Sichuan, Yunnan and Tibet）, which places a threat to the health of local residents. However, few studies have been carried out in this region and knowledge gaps still exist in the epidemic characteristics and the underlying mechanism of bacillary dysentery. Moreover, no research conclusion or experience on disease prevention and control in other regions can be directly applied to southwest China, which is mainly due to its unique natural environment and socio-economic environment. This study was designed to fill the above knowledge gaps. Based on county-level report data of bacillary dysentery in southwest China, descriptive analysis, circular distribution analysis were firstly applied to reveal the epidemic characteristics of bacillary dysentery from various perspectives, including its spatial pattern, seasonality, susceptible population. Database of Shigella serogroups and database of outbreak cases were established as achievements of systematic review works. Characteristics of etiology and outbreaks were acquired through meta-analyses. Then, this study focused on the spatial pattern, seasonality and short-term meteorological event （rainstorm-induced flood）. Spatial risk model and time series model were built to explore the spatial and seasonal characteristics of bacillary dysentery in southwest China and the underlying mechanism. Impact of rainstorm-induced flood on bacillary dysentery was estimated. The main conclusions of the study are as follows:1. The spatial location and range of hot spots of bacillary dysentery are stable during 2005-2014, which mainly locate in the eastern margin of the Tibetan Plateau and southeastern Tibet. The spatial pattern was found associated with underdeveloped socio-economy, less medical and health resources, higher environmental temperature, low and medium-low relative humidity, proximity to farmland or water bodies, the distribution of mountain lands and the distribution of the Tibeto-Burman ethnicity. 2. The spatial risk model built is able to predict the risk of bacillary dysentery with finer spatial resolution, using geographical grid units instead of administrative units. The downscaling risk prediction can help indicate where exactly medical and health resources are urgently needed and places that may be neglected by passive disease monitoring, and therefore provide scientific basis for the optimal allocation of medical resources. 3. The seasonal peak of bacillary dysentery comes 1-2 months earlier in most of Yunnan and southern Sichuan, as compared with other areas in China. The earlier peak was found associated with seasonal drought in that specific region. Drought mainly affects the countryside, but impact on built-up area with improved infrastructure is limited. The seasonal trend of bacillary dysentery in the built-up area is related to the variation of temperature and peak comes late, which is similar with other regions of southwestern China where peak late. 4. A linear or approximately linear positive correlation was found between the monthly cases of bacillary dysentery and meteorological factor （temperature or evaporation）, but the threshold effect of meteorological factor was also detected in part of the study area. The relatively weak seasonality of bacillary dysentery in Sichuan is affected by the temperature threshold effect, which may be due to effective disease prevention and control during high temperature seasons. 5. Rainstorm-induced flood is a risk factor of bacillary dysentery. Exposure to rainstorm-induced floods is conducive to the spread of bacillary dysentery. In terms of health risks associated with flood, severe flood is higher than mild one, rural area is higher than urban area, and mountain area is higher than non-mountain area. The number of bacillary dysentery cases was also found increase significantly after flood in some urban districts, which suggesting the existence of weak links of disease prevention and control even in places with improved infrastructure. The response speed of dysentery to exposure vary between different areas, which is probably due to the diverse natural and man-made environments leading to different specific exposure-response processes involved in various places.The findings of this study are effective supplements and improvements to human’s knowledge on the epidemic characteristics and the underlying mechanism of bacillary dysentery in southwest China （Sichuan, Yunnan and Tibet）. The findings are significant to public health management and can provide scientific basis for the formulation of dysentery prevention and control strategy, optimal allocation of medical and health resources, and better adaptation to climate change in the future.
