Research On The Intervention Of Disease Component Values Based On PDCA And The Establishment Of A Prediction Model For Hospitalization Expenses In COPD Patients

Objective: Based on the background of Chengdu payment by disease component value under total amount control,and taking the real-world data of the hospital as an example,the PDCA cycle is used for medical insurance management,in order to improve the medical insurance status of the hospital,and verify the implementation effect of the PDCA cycle intervention.Taking COPD as an example,in order to provide a reference for cost control a prediction model of hospitalization expenses was established,and the influencing factors of hospitalization expenses were mined.Methods: Based on the background of payment by disease component value under total amount control,according to the cost deviation data issued by the medical insurance bureau of a hospital,the management quality of PDCA cycle is carried out with keywords such as the suitability of diagnosis and treatment,whether the admission indications are met,and whether the quality of medical services is reduced.Then,analyzing the causes of problems and the implementation effect after intervention with these keywords;For COPD patients,a hospitalization cost prediction model is established by using artificial neural network and support vector machine.After that,the model prediction performance is to evaluate,then,obtaining important factors that affect hospitalization costs.Results:（1）After adopting the PDCA cycle intervention for the high-cost part,the difference of hospital’s medical insurance payment showed a cumulative positive value of more than600,000 yuan,and the proportion of ultra-high cost deviation from 1 to 1.4 decreased by10.82%（P<0.05）.（2）After the PDCA cycle intervention for the low-cost part,the difference of hospital’s payment remained positive,and there was no negative payment difference.The low-cost deviation did not change much,and there was no statistical difference.（3）In the establishment of the hospitalization cost prediction model for COPD patients,there are 9 factors in the univariate analysis that have a statistically significant impact on the hospitalization cost.For the cost prediction part of the continuous hospitalization cost variable,the SVR-polynomial kernel function prediction model has the smallest mean absolute error.,the better the correlation,the most important variable affecting the hospitalization cost is the length of stay.For the classification of the hospitalization cost,the cost classification prediction part,the ANN model has a better prediction effect in the cost classification prediction part,and the most important factor of the predictor variable is the length of stay.Conclusion: In general,the study verifies the role of PDCA cycle management in medical insurance management under the background of patient group value payment.For the intervention effect,it has a certain intervention effect on controlling ultra-high hospitalization expenses and reducing medical insurance deductions.However,the intervention effect on low expenses is not ideal.At the same time,it also provides a new method for other medical institutions to take measures to carry out medical insurance cost control.Data mining is used to screen the hospitalization cost model of COPD patients,and the optimal model for predicting the cost of COPD patients is screened out,which has a certain reference value for medical institutions to reasonably control the hospitalization cost.