游泳场所卫生状况综合评价方法的应用和比较

Application and Comparison of Comprehensive Evaluation Methods for Hygienic Evaluation of Swimming Places

[目的]应用综合评价方法评价游泳场所卫生状况，并对各方法的合理性进行比较，并验证其一致性。[方法]以2009年上海市长宁区30家游泳场所的卫生监督检查和水质检测两大类资料作为研究对象。对卫生监督检查资料应用量化分级管理评分法综合评价游泳场所卫生状况；对水质检测资料分别应用合格率法、综合指数法、模糊数学法和人工神经网络法构建水质综合评价模型并进行水质等级判断。最后，对五种综合评价方法的结果进行一致性检验。[结果]相比于其他评价方法，人工神经网络法有其方法学的优势，经其构建的水质综合评价模型评价的游泳场所水质平均等级为2．27级，各水质等级的游泳场所单位数的分布呈“两头少，中间多”的纺锤形结构，且显示出体育场馆内设游泳场所的等级达标率较高（100．0％）和小区会所内设游泳场所的等级达标率较低（68．2％），这些结果与历年来的游泳场所卫生监督检测结果基本相符。经综合指数法、合格率法和模糊数学法构建的水质综合评价模型评价的游泳场所水质平均等级分别为3．23级、2．87级和2．20级。经量化分级管理评分法评价的游泳场所卫生平均等级为2．27级。五种综合评价方法的结果间的Kappa系数在-0．134～0．173之间。[结论]五种综合评价方法结果的一致性较低，应用人工神经网络法构建的游泳场所水质综合评价模型相对合理。

[ Objective ] To apply several comprehensive methods to the evaluation of hygienic conditions of swimming places, compare the reasonability of these methods, and test their consistency. [ Methods ] Two kinds of data were obtained from the archives of the health supervision and water quality monitoring of thirty swimming places in Changning District, Shanghai in 2009. A quantitative and graded management method was used for health supervision data. Models based on four methods, i.e. qualification rate, comprehensive index, fuzzy mathematics, and artificial neural network, were built to rank water quality levels. Results of five comprehensive models were tested for consistency. [ Results ] Compared with the other evaluation methods, the advantage of artificial neural network method lay in methodology. The evaluation model based on artificial neural network method showed that the average level of water quality was 2.27 and that the water quality level distribution of swimming places was spindle shaped. Moreover, the proportion of swimming places at excellent and general water quality levels in gymnasiums （100.0%） was higher than those in community health clubs （68.2%）. These results were basically consistent with health supervision data over the past years. The average level of water quality indicated by models based on comprehensive index method, qualification rate method, and fuzzy mathematics method was 3.23, 2.87, and 2.20, respectively. The average level of swimming places evaluated by quantitative and graded management method was 2.27. The coefficients of Kappa ranged from -0.134 to 0.173 between the results of the five comprehensive evaluation models. [ Conclusion ]There is a low consistency between the results of the five comprehensive evaluation methods. The model based on artificial neural network method is relatively reasonable.