Research On The Mathematical Model Of Medical Dispute Prediction ——Taking A Tertiary Hospital In Gansu Province As An Example

Objective: Through the in-depth analysis of the medical dispute data of a tertiary hospital in Gansu Province in recent years,a mathematical model for the prediction of medical disputes is established,which provides a reference for medical institutions to effectively prevent medical disputes.Method: 250 cases recorded by the medical department of a tertiary hospital in Gansu Province from 2018 to 2020 in which the patients themselves or/and their family members had medical disputes with the hospital were selected as the dispute group.The main discharge diagnosis of the patients in the dispute group was the same,According to the principle of the closest admission date,496 discharged cases without medical disputes were selected in the HIS system as the control group.The influencing factors of medical disputes were screened by Delphi method.Univariate analysis between the two groups of cases was carried out by using the chi-square test,independent sample T test and other methods to screen the influencing factors of medical disputes,and the factors with statistical differences in the occurrence of medical disputes were screened.Binary Logistic regression method was used for multivariate analysis,independent influencing factors affecting medical disputes were screened,and Logistic regression model was established;artificial neural network radial basis function was used to establish the model.According to the area AUC value,sensitivity,specificity and other indicators under the Receiver Operating Characteristic Curve(ROC)of the two models,the effectiveness of the two models in predicting medical disputes was compared.Results: 1.Determine the candidate list of 34 possible influencing factors of medical disputes through literature research.Through two rounds of Delphi method consultation,delete the factors with an average importance score less than 3.5 or a coefficient of variation greater than0.25,and finally retain 12 influencing factors of medical disputes.Respectively,age,gender,ethnicity,marriage,occupation,type of payment,number of admissions,average daily hospital charges,doctor's title,notification of critical illness,volume of blood transfusion,and discharge method.2.Univariate analysis was carried out on 12 influencing factors of the extracted 250 cases of dispute medical records and 496 control group medical records: age,gender,ethnicity,marriage,occupation,number of hospital admissions and other factors were not significantly related to the occurrence of disputes(P ? 0.05),six factors,such as type of payment,average daily hospitalization expenses,doctor's professional title,notification of critical illness,blood transfusion volume,and way of leaving the hospital,were significant between the dispute group and the control group(P<0.05).3.Logistic regression analysis was performed on the 6 statistically significant influencing factors obtained by univariate analysis,with P<0.10 as the criterion,a total of 5 factors entered the model,and the degree of influence on the occurrence of disputes was as follows: Discharge method,doctor's title,blood transfusion volume,payment type,and average daily hospitalization expenses;the positive prediction rate of the logistic regression model was finally 24.80%,the negative prediction rate was 94.76%,and the overall accuracy was 71.31%.4.Taking the six statistically significant factors influencing disputes as the input source and whether there is a dispute as the output source,through the artificial neural network model analysis,it is concluded that the order of importance of the influencing factors of medical disputes is as follows: Hospital method(100.0%),doctor's title(91.9%),average daily hospitalization expenses(81.7%),total blood transfusion(66.5%),critical illness(28.7%),payment type(20.8%);artificial neural network model prediction The positive predictive value was 58.49%,the negative predictive value was 92.23%,and the overall accuracy was 80.77%5.The area under the curve(AUC)value of the logistic regression model is 0.700;the AUC value of the artificial neural network model is 0.832.It is comprehensively judged that the artificial neural network model has better prediction effect on medical disputes.Conclusion:1.Give full play to and combine the advantages of Delphi method and data statistical analysis method,which can effectively screen out the influencing factors of medical disputes,and lay a foundation for further research on medical disputes.2.The artificial neural network model in this study is slightly better than other studies in the prediction performance of medical disputes,and the logistic regression model has a general prediction effect,which is more suitable for the prediction of medical disputes of a single disease type.The prediction ability of the two is compared and analyzed,and the best prediction model is obtained.The research method is reliable and innovative in dispute prediction.3.Among the 12 influencing factors of medical disputes included in this study,in addition to the occupation of the patient,whether the other 11 influencing factors have an impact on the occurrence of disputes is basically consistent with most studies.4.The research model can be combined with the hospital information system to collect the risk factors affecting medical quality and patient safety that occur during the diagnosis and treatment of each patient,and proactively predict the risk.Hospital managers and clinical medical staff can Early intervention of risk factors for patients can effectively reduce the occurrence of medical disputes and maintain a good clinical diagnosis and treatment environment.