目的：探讨基于病人临床信息的护理风险预警模型的建立方法，用于早期识别高风险患者，及时采取干预措施。方法筛选采集某院HIS系统中2013年9月1日～2014年9月1日出院或死亡的120例患者资料，用Logistic回归筛选出与病人病情转归有明显相关性的临床指标，并建立回归方程，计算死亡概率P值。然后选择2015年6月1日～2015年12月30日出院和死亡病人各25例，统计其出院或死亡前24 h内每小时的各指标值，计算死亡概率P，绘制每位病人的P值散点图，以验证模型的有效性。结果通过建立回归方程计算得知当P≥0.6567时，提示病人具有较高的死亡风险。模型验证结果显示，当P＞0.5时，病情恶化甚至死亡的可能性较大，该模型的预测结果与实际情况基本一致。结论应用Logistic回归分析，建立基于HIS系统的预警模型，能较好预测病人病情转归，且获取数据容易，具有较高的时效性。%Objective To introduce the establishment method of a nursing risk pre-warning model based on patient clinical information so as to early recognize high-risk patients in the early stage and take intervention measures promptly.Methods The clinical information of 120 patients who had been discharged from a certain hospital or dead between September 1st 2013 and September 1st 2014 were selected from HIS (Hospital Information Systems). The indictors which were obviously related with disease development were selected by using Logistic regression analysis to establish the regression equation and calculateP values. Then, 25 patients discharged from the hospital and another 25 patients who had been dead were selected respectively. Their indicators were collected hourly within 24 hours before their discharging from the hospital or death to calculate and mark the probability of death asP. The scatter plot for every patient was made in order to validate the effectiveness of the model.Results The establishment of regression equation was used to calculateP values. P≥0.6567 suggested that the patient had a higher risk of death. According to the model validation results, P>0.5 indicated that the disease was likely to become worse or even more likely to die. The predicted results of the model were basically consistent with the actual situation.Conclusion Application of logistic regression analysis to establish pre-warning model based on HIS can make predictions of patients’ disease development and obtain data easily with high timeliness.
展开▼
