Construction And Validation Of Risk Model For Recurrence Of Diabetic Foot Ulcer Based On Evidence-based Medicine

Objective(s): The purpose of this study is to study the risk factors of recurrence of diabetic foot ulcer by using the method of systematic review and meta-analysis in evidence-based medicine,to construct the prediction model of recurrence of type 2 diabetic foot ulcer by combining the optimized Logistic regression model,and to design the corresponding risk scoring system,so as to facilitate the clinical application,provide help for guiding the individualized treatment of diabetic foot ulcer patients and identifying the high-risk population of recurrence,and try to achieve early detection,early diagnosis and early treatment,so as to promote rehabilitation,saving medical cost resources and controlling the recurrence rate.Method(s):(1)Meta-analysis and systematic review: According to the meta-analysis specification,formulate reasonable search strategy and inclusion and exclusion criteria to obtain relevant literatures,further extract research data after reasonable screening,combine the effect size of relevant risk factors of type 2 diabetic foot ulcer recurrence,and finally obtain the combined OR value of each risk factor;(2)Construction and validation of risk model: the OR values of risk factors screened by meta-analysis were mathematically transformed and substituted into the optimized Logistic regression model to construct a risk prediction model for diabetic foot ulcer recurrence,and the performance of the model was evaluated by ROC curve;(3)Design and Verification Risk Scoring System:The DFU recurrence risk prediction model was transformed into a simple and easy-to-use risk scoring system,and the performance of the scoring system was verified by the patient information of our hospital and the risk stratification of the patient population was carried out.Result(s):(1)Meta-analysis results:A total of 15 studies at home and abroad were included,involving 16 risk factors.Finally,12 significant related factors were obtained,including gender,smoking history,Wargner classification,degree of blood glucose control,ulcer site,peripheral neuropathy,peripheral vascular disease,diabetic nephropathy,osteomyelitis,amputation history,deformity and multi-drug resistant infection.The pooled OR values were1.47,1.64,2.61,3.62,3.29,2.97,3.30,1.47,2.82,2.35,1.92,2.82 respectively.(2)Construction of DFU recurrence risk model and scoring instrument:the final model was Logit(P)= α + 0.39 × sex + 0.49 × smoking history + 0.96 ×Wagner grade + 1.29 × degree of glycemic control + 1.19 × plantar ulcer + 1.09× peripheral neuropathy + 1.19 × peripheral vascular disease + 0.39 × diabetic nephropathy + 1.04 × osteomyelitis + 0.85 × amputation history + 0.65 ×malformation + 1.04 × multidrug resistant infection.The corresponding scoring system is constructed by rounding the values obtained by multiplying the regression coefficients of the above formula by 10.(3)Verification of DFU recurrence risk model and scoring system:The exposure of risk factors of DFU patients in our hospital was substituted into the risk model and ROC curve was drawn to calculate AUC(95% CI)was0.826(0.750,0.903),using scoring system to calculate the total score of each patient and find the corresponding recurrence probability,draw ROC curve,its AUC(95% CI)was 0.828(0.753,0.904).The total score(32 points)corresponding to the maximum Youden index was taken as the optimal cut-off value.A score greater than 32 points was defined as recurrence,and a score less than or equal to 32 points was defined as non-recurrence.The AUC(95% CI)was 0.784(0.655,0.835),the sensitivity was 0.944,and the specificity was0.623 when the ROC curve was drawn.Conclusion(s): Based on systematic review and meta-analysis in evidence-based medicine,this study explored the risk factors of diabetic foot ulcer recurrence,and further established a risk prediction model for diabetic foot ulcer recurrence,which was presented in the form of risk scoring system.The clinical application was more simple and rapid,and easy to promote.Finally,the data of real patients in our hospital were retrospectively collected to verify the model,which showed that the prediction performance of the model was good.The results of this study have certain reference value for predicting the risk of recurrence of diabetic foot ulcer,and can provide new ideas for the construction of prognostic models of other diseases.