广东省气象因素对疟疾发病影响的时间序列分析

Time-series analysis on the malaria morbidity affected by meteorological factors in Guangdong province

目的评价广东省气象因素与疟疾发病风险的关系。方法采用时间序列方法，使用分布滞后非线性模型，以1980--2004年广东省气象部门提供的温度数据和广东省CDC提供的疟疾发病数据对温度、相对湿度和降雨量等主要气象因素与每月疟疾发病率进行负二项分布回归分析。对气象因素对疟疾发病影响的即时效应、滞后效应、累计效应及气象因素的交互作用进行分析。结果广东省1980--2004年间每月疟疾发病数最高达到4010例，月最高温度平均为26．3℃，每月最低平均温度为18．8℃，每月平均温度为21．9℃；月相对湿度最高可达到88．0％；月平均降雨量平均为5．6mm。月最高温度与疟疾发病的即时效应呈非线性关系，当温度达到32．3℃时，发病风险最高，RR值达到2．51（95％CI：1．99～3．16）；相对湿度为60．0％时，疟疾发病的RR值最高1．19（95％CI：0．66～2．11），随后逐渐下降，当相对湿度为86．6％时疟疾的发病风险降至最低，RR值为0．51（95％CI：0．34～0．76）；疟疾的发病风险随月平均降雨量的增加而增加，当月平均降雨量为14．5mm时，疟疾的发病风险最高，RR值为1．29（95％CI：0．87～1．93）。月最高温度在31．5℃，滞后2个月时对疟疾发病的效应最强，RR值达到1．81（95％CI：1．02～3．22）；降雨量在〉15．2mm时对疟疾发病的作用较强，但持续时间则较短。33．7℃高温天气下疟疾的发病超额危险度较大，滞后1个月时为92．2％（95％C1：30．5％-183．2％）。相对湿度较小时，疟疾发病的滞后作用持续时间长，累计作用大；相对湿度达到87．0％时，滞后3个月的超额危险度仅为-66．6％（95％CI：-86．4％～-17．7％）。降雨量为15．5mm时对疟疾发病的累计作用在滞后3个月时达到最高，超额危险度值为40．7％（95％CI：-30．0％～182．6％），此后累计效应逐渐减弱。在滞后2个月时，每月最高温度和降雨量对疟疾的发病具有明显的正向交互作用，而降雨量与相对湿度对疟疾发病具有负向交互作用。结论较高温度和较大降雨量有可能是广东省疟疾流行的危险因素，且两者间存在明显的交互作用。

Objective To evaluate the associations between malaria risk and meteorological factors. Methods A negative binomial distribution regression analysis was built between the temperature, relative humidity, rainfall capacity and the monthly incidence of malaria, based on the temperature information provided by Guangdong Meteorological Department and the malaria incidence information provided by Guangdong Center of Disease Prevention and Control during year 1980 to 2004, adopting the time-series analysis method and by distributed lag non-linear model, in order to analyze the immediate factors. Results The number of monthly malaria cases in Guangdong province reached 4010 between year 1984 and 2004, while the monthly maximal temperature, minimal temperature, average temperature, relative humidity and average rainfall capacity was separately 26. 3 ℃ , 18. 8 ℃ , 21.9 ℃ , 88.0% and 5.6 ram. The immediate effect of monthly maximal temperature on malaria incidence showed non-linear relationships. When the temperature reached 32. 3 ℃ , the risk was highest, the relative risk （RR） was 2. 51 （ 95 % CI： 1.99 -3. 16） ;when the relative humidity was 60. 0% , the relative risk of malaria was highest as 1.19（95% CI： 0. 66 -2. 11 ）and then decreased gradually; and when the relative humidity was 86. 6%, the risk of malaria was lowest at O. 51 （95% CI：0. 34 -0.76）. The risk of malaria increased while the rainfall capacity was 14.5 mm,the risk of malaria was the highest at 1.29（95% CI：0. 87 - 1.93 ）. Strongest delayed effects on malaria incidence was observed when the monthly maximal temperature reached 31.5 ℃ at lagged 2 months, with the value of RR at 1.81 （95% CI： 1.02 - 3.22） . When the monthly rainfall capacity was over 15.2 ram,the delayed effects was strong but short. When the monthly maximal temperature of 33.7 ℃, the excess risk of malaria was comparatively high, the excess risk was 92. 2% （95% CI： 30. 5% - 183.2% ） when lagging one month. When the relative humidity was low, the delayed effect of malaria lasted for a long time, and the cumulative effect was huge. When the relative humidity reached 87.0% ,the excess risk lagging 3 months was only -66.6% （95% CI： -86. 4% - - 17.7% ）. When the rainfall capacity was 15.5 ram, the cumulative effect on malaria reached the peak after 3 months, while the excess risk was 40. 7% （95% CI： -30.0%--182.6%）; afterwards the cumulative effect gradually weakened. Positive and negative interaction effects were significant between malaria risk and maximal temperature and monthly rainfall capacity,and monthly rainfall capacity and relaive humidity at lagged 2 months, repectively. Conclusion High temperature and large rainfall capacity might be the risk factors of malaria in Guangdong province,and there was an obvious interaction between the two factors.