基于数据挖掘的医院综合评价量化建模方法研究

为解决患者就医寻找合适医院难的问题,提出依据医院诊疗数据,建立基于死亡率与综合实力的医院评价模型。在基于死亡率评价模型中,筛选出5个影响治疗效果的病人自身因素并将其量化。以成功治愈样本为基础,计算出各因素对于治愈天数的影响权重,以此计算出所评价医院每一个死亡样本治愈的预估天数。若大于治愈天数的最大阈值,则确定该样本死亡不可避免。否则为可避免,由此得到该医院关于该病症的"不当死亡率",以此评判该医院针对该疾病的诊疗水平。在基于综合实力的评价模型中,对医院综合实力分11个指标进行评价并量化。采用主成分分析法确定各指标在整套评价方案中所占权重,将该指标量化加权后的结果与所有医院相同指标量化加权均值计算相对差,作为该医院该指标得分,加和该医院所有指标得分作为评价分值。采用4家样本医院实际诊疗数据对模型进行了检验,模型评价的结果符合样本医院的实际水平。

QUANTITATIVE MODELING OF HOSPITAL COMPREHENSIVE EVALUATION BASED ON DATA MINING

To solve the problem that it is difficult for patients to find a suitable hospital, this paper established a hospital evaluation model based on mortality and comprehensive strength according to the data of hospital diagnosis and treatment. In the mortality-based evaluation model, five factors affecting the outcome of treatment were selected and quantified. On the basis of successful healing samples, the influence weights of each factor on the healing days were calculated, and the estimated healing days of each death sample in the evaluated hospital were calculated. If the time is longer than the maximum threshold of healing days, then the death of the sample was inevitable, otherwise it was avoidable. Thus, the improper mortality was obtained to evaluate the level of diagnosis and treatment of the disease in the hospital. In the evaluation model based on comprehensive strength, the hospital comprehensive strength was evaluated and quantified by 11 indexes, and the weight of each index in the whole evaluation scheme was determined by principal component analysis. We calculated the relative difference between the quantified weighted results of the index and the quantified weighted mean of the same index in all hospitals. It was used as the index score, and all the index scores of the hospital were added as the evaluation score. We tested the model with 4 sample hospitals actual diagnosis and treatment data. The results of model evaluation are in line with the actual level of sample hospitals.