Construction And Validation Of A Prediction Model For Endpoint Events In CKD Patients Under The Intervention Of Traditional Chinese Medicin

Purpose:Through the collection,statistics and analysis of the general clinical data and various objective indicators of patients with chronic kidney disease,the risk factors of endpoint events were screened,and the nomogram prediction model was constructed and verified,so as to facilitate the patients to understand the prognosis of the disease and facilitate the clinicians to evaluate the prognosis of the patients,predicting the timing of dialysis,and explore the influence of traditional Chinese medicine intervention on the prognosis of endpoint events in patients with chronic kidney disease.Material and method:The inpatient data of all patients with CKD stage V from May 1,2018 to April 30,2021 were collected through the medical record query system of the Department of Nephrology,Affiliated Hospital of Liaoning University of Traditional Chinese Medicine as the modeling dataset,and the data from May 1,2021 to December 1,2021 were collected as the validation dataset.The endpoint event was determined as the patient’s survival and receiving and entering renal replacement therapy.The collected of patients was divided into the group without endpoint event and the group with endpoint event.The Exel table,SPSS Statistics 21,R 4.2.2 and R studio software were used for statistics,analysis,drawing of nomograms,and verification of nomogram prediction model.Then according to the results of the screening of independent risk factors for analysis,and explore the impact of Chinese medicine on its prognosis.Results:There were 157 subjects in the collected modeling dataset,of which 86 subjects(54.78%)did not have endpoint events and 71 subjects(45.22%)had endpoint events;There were 43 subjects in the validation data set,among whom 19(44.19%)had no endpoint event and 24(55.81%)had an endpoint event.Statistical analysis of the modeling dataset showed that there were statistical differences between the two groups in whether taking Chinese medicine decoction,hemoglobin,albumin,triglyceride,potassium ion,phosphorus ion and uric acid(P<0.05).Binary Logistic regression analysis showed that no taking Chinese medicine decoction,lower hemoglobin,lower albumin and higher potassium and phosphorus ions were more likely to enter renal replacement therapy.According to the results of binary Logistic regression analysis,the nomogram prediction model was drawn and verified by internal and external data.C index C=0.913>0.75(95%CI=0.802,0.915),C=0.912>0.75(95%CI=0.842,0.917),and the mean absolute error was 0.044,0.032.Conclusion:In this study,the data of hospitalized patients with CKD stage V from May 1,2018 to December 1,2021 in the Department of Nephrology,Affiliated Hospital of Liaoning University of Traditional Chinese Medicine were collected and sorted out through retrospective analysis.Through binary logistic regression analysis,the independent risk factors for renal replacement therapy events in patients were screened as not taking traditional Chinese medicine decoction,lower hemoglobin,lower albumin and higher potassium and phosphorus ions.A nomogram prediction model with good discrimination and calibration is drawn,which can be used by clinicians to predict the risk of RRT and the timing of dialysis for patients intuitively to a certain extent,and the traditional Chinese medicine prescription has certain curative effect on the prognosis of patients.