基于BP神经网络的临床路径优化

针对临床路径的复杂性和模糊不确定性,对其进行综合分析,计算出临床路径诊疗单元的平均治疗天数以及临床路径的总治疗天数,分别作为实验的样本数据和最终评价指标。在此基础上,利用BP神经网络的方法对临床路径进行优化建模。并通过实验对比分析隐层神经元数量对临床路径优化结果的影响,发现神经元数量增多时,错误率明显下降,迭代次数呈上下波动。最后选取7-9-1的BP神经网络结构,以某地区三甲医院的糖尿病加高血压临床路径为例,进行仿真实验。结果显示,训练模型的输出值与期望值之间的相对误差范围在0-0.2%之内,测试模型的输出值与期望值之间的相对误差范围在0-0.1%之内。说明该模型具有较强的自学习自适应能力,能够有效地优化临床路径,减少患者的治疗天数。

Clinical Path Optimization Based on BP Neural Network

In view of the complexity and fuzzy uncertainty of the clinical pathway,a comprehensive analysis is carried out to calculate the average treatment days of the treatment unit and the total treatment days of the clinical pathway,respectively as the experimental sample data and the final evaluation indicators.On this basis,the BP neural network is used to optimize and model the clinical pathway.The effect of the number of hidden layer neurons on the optimization of the clinical pathway is compared and analyzed by experiment.It was found that when neurons increasing,the error rate decreases significantly and the number of iteration fluctuates up and down.Finally,the simulation experiment is performed by selecting the BP neural network structure of 7-9-1,with the clinical pathway of diabetes and hypertension in the top three hospitals in a certain area as an example.The results show that the relative error range between the output value and the expected value of the training model is within 0-0.2%,while that of the test model is within 0-0.1%.It indicates that the model has strong self-learning and self-adaptive,and can effectively optimize the clinical pathway and reduce the treatment days of patients.