| ObjectiveIn this study,through literature research and semi-structured interviews,the influencing factors of diabetic foot(DF)recurrence were summarized and classified,then the extraction table of influencing factors for DF recurrence were formed.Using the extraction table,the influencing factors of DF recurrence were analyzed through a case-control study,and the risk prediction model was constructed based on Logistic regression.Then,the nomogram was drawn to visualize the model and the predictive performance of the model was evaluated.MethodsThe study consists of two parts.(1)Based on the literature research,Pub Med,Embase,Cochrane Library,Web of Science,including CNKI,CBM,Wan Fang Data and VIP were searched comprehensivly and systematicly.The relevant literatures on the influencing factors of DF recurrence were screened,the relevant literatures on the influencing factors of DF recurrence were screened,the quality evaluation and data extraction of the included literature were carried out,and the influencing factors of DF recurrence were clarified.Through semi-structured interview,clinical medical staff were interviewed by purpose sampling method,and the influencing factors of DF recurrence were supplemented and improved,and the influencing factors were classified.Then,the extraction table of influencing factors of DF recurrence was formulated.(2)A case-control study was adopted.According to the inclusion and exclusion criteria,the discharged patients from January 2015 to July 2021 were retrieved from the medical medical record system as the research object.According to the readmission records,whether the patients had the first recurrence of podiatry was determined.Relevant data of the case group and the control group were collected according to the intake criteria,and the modeling group and the verification group were divided into 7:3 groups.Logistic regression was used to construct a risk prediction model for DF recurrence,and a nomogram of the model was drawn.The discrimination and calibration of the model were evaluated by receiver operating characteristic curve(ROC curve),Hosmer-Lemeshow(H-L)goodness of fit and calibration curve.Results(1)In the literature research stage,15 literatures were included,including 11 cohort studies,and 4 case-control studies;9 studies from China,and 6 studies from England.The quality evaluation results were all ≥5 points,and 27 factors influencing DF recurrence were obtained.(2)According to general demographic factors,diabetic disease factors,diabetic foot disease factors,physiological and biochemical indicators,foot specific examination,treatment and intervention factors,psychological and behavioral factors,they were divided into 7categories,and an extraction table of factors affecting DF recurrence was formed.(3)A total of 1026 DF patients were included in this study,and 209 patients were relapsed.The patients were divided into modeling group and validation group.718 patients in modeling group including 146 recurrent patients,and 308 patients in validation group including 63 recurrent patients.There was no statistical significance in each index between the two groups(P>0.05).(4)The results of univariate analysis showed that there were 16 variables with statistically significant differences between the recurrence group and the non-recurrence group(P<0.05).The variables included diabetic course(years),complicated with coronary heart disease(CHD),hypertension,diabetic nephropathy(DN),Texas grade,lesion site(plantar),glycosylated hemoglobin(Hb A1 c %),C-reactive protein(CRP mg/L),ankle-brachial index(ABI),transcutaneous oxygen pressure(Tc PO2),hyperglycemic medication,amputation,vascular intervention treatment,smoking,BI index and foot care behavior were the influencing factors of DF recurrence.(5)Logistic regression results showed that 11 variables entered the DF recurrence risk prediction model,including CHD,DN,plantar lesion,Hb A1c(%),CRP(mg/L),ABI,hyperglycemic medication,amputation,vascular intervention treatment,smoking and foot care behavior.The logistic regression equation was: Logit(P)=(-9.127)+1.095×CHD+0.605×DN+0.862×plantarlesion+2.245×Hb A1c+2.365×CRP+1.749×ABI+0.743×hyperglycemic medication(1)+1.216×hyperglycemic medication(2)+ 0.794×amputation+1.181×vascular intervention+0.883×smoking-1.017×foot care behavior.(6)The nomogram was drawn by the Logistic regression analysis results,and the optimal cutoff value of the model was 0.188.The results of discrimination evaluation showed that the area under the ROC curve(AUC)of the modeling group was 0.885(95%CI: 0.856-0.913),the sensitivity(Sen)was 82.9%,and the specificity(Spe)was 77.3%.The AUC of the validation group was 0.844(95%CI: 0.795～0.892),the Sen was 79.4%,and the Spe was 75.9%.The results of calibration evaluation showed that the H-L chi-square value of the modeling group was 4.470(P>0.05),and the H-L chi-square of the validation group was 6.443(P>0.05).And the calibration curves all showed that the predicted probability of DF recurrence was consistent with the actual probability.Conclusion(1)Through literature analysis and semi-structured interviews,46 fatcors were formed,including general demographics,diabetic disease factors,diabetic foot disease factors,physiological and biochemical indicators,foot-specific examinations,treatment and intervention factors,and psychological and behavioral factors.Then,the extraction table of influencing factors of DF recurrence was formulated.(2)The single influencing factors of DF recurrence included DM duration(years),CHD,hypertension,DN,Texas grade,plantar lesion,Hb A1c(%),CRP(mg/L),ABI,Tc PO2,hyperglycemic medicine,amputation,vascular intervention treatment,smoking,BI index and foot care behavior.(3)The predictive factors of DF recurrence risk prediction model consists of CHD,DN,plantar lesion,Hb A1c(%),CRP(mg/L),ABI,hyperglycemic medication,amputation,vascular intervention treatment,smoking and foot care behavior.The model’s discriminative ability was moderate,and the predicted probability of DF recurrence was relatively consistent with the actual probability. |
PDF Full Text Request