Analysis Of Type2 Diabetes Of Tianjin Based On Latent Class Analysis And SF-8

Objective: Preventing or delaying the occurrence of type 2 diabetes complications and comorbidities can improve the health status of patients and reduce the burden of family and country.This study aimed to explore the prevalence and risk factors of comorbidities and complications of diabetes among patients with type2 diabetes in Tianjin,Evaluating the applicability of SF-8 in patients with diabetes,after that,discussing the quality of life of respondents.Method: 1.The data used in this study was abstract from the 2015 survey database of patients with type 2 diabetes in Tianjin.Descriptive analyzing the prevalence of target diseases,based on the incidence of complications and comorbidities,patients were classified into several subcategories using Latent Class Analysis.After the univariate analysis of patient characteristics,meaningful variables were selected into multiple nominal Logistic regression to find the risk factors of each latent class.2.The factors influencing the group were treated as covariates,carrying out Latent Class analysis with covariates and comparing it with method 1.3.Reliability and validity was examined with Cronbach’s α coefficient and factor analysis separately,what’s more,descripting the quality of life among patients with type2 diabetes in Tianjin.4.Latent Class analysis and Latent Class analysis with Covariates were analyzed by Latent gold 5.0 software,what’s more,descriptive analysis,univariate analysis,multinomial logistic regression analysis,the descriptive of reliability and validity were performed with SPSS 20.0.Results: 1.The prevalence of complications related to type 2 diabetes in Tianjin were listed as below: 21.67%(cardiovascular disease),21.58%(diabetic peripheral neuropathy),18.23%(diabetic lower extremity vascular disease),17.51%(diabetic retinopathy),11.06%(diabetic nephropathy),7.28%(cerebrovascular disease)and 1.82%(diabetic foot).The prevalence of comorbidities of diabetes were 37.88%(hypertension),18.15%(dyslipidemia)and 9.32%(metabolic syndrome).2.Patients with type 2 diabetes mellitus were classified into five classes using both latent class analysis and latent class analysis with covariates,however,the second way had lower BIC(37.76.83 vs 36469.06)and lower classification error(0.17 vs 0.14)than the first way.According to the results of latent class analysis with covariates,five latent classes were termed “severity group”(7.84%),“peripheral neuropathy & lower extremity vascular disease group”(6.56%),“comorbidities & cardiovascular and cerebrovascular disease group”(12.43%),“microvascular disease group”(21.00%)and “diabetes without complications & comorbidities group”(52.17%).3.Overall consideration of the results derived from logistic regression analysis and gamma parameter,older age,longer duration of diabetes and higher 2h postprandial plasma glucose tested in the diagnosis were risk factors of latent class 1-4;female was the risk factor of latent class3;family history was the risk factor of latent class1,latent class2 and latent class4;higher BMI was the risk factor of latent class 1-3 and central obesity was the risk factor of latent class2;living in around district of Tianjin was the risk factor of latent class1 and latent class3-5.4.Cronbach’s α coefficient was 0.89 and the Cronbach’s α coefficients between 8 items and total score were higher than 0.60,which means the internal consistency of SF-8 was well.Extracting one factor which eigenvalue larger than 1 using exploratory factor analysis,the cumulative variance proportion was 73.23%.Multiple correlation coefficients between 8 items and one factor were relatively large,which means the construct validity was well.Quality life of type 2 diabetes patients in Tianjin was low,and the scores of 8 items were all less than 50.The score of general health was the highest(48.80),and the score of role-emotional was the lowest(21.13).Conclusion: 1.The prevalence of comorbidities and complications of patients with type 2 diabetes in Tianjin were lower than other cities in China,and the disease was controlled well.2.Based on the prevalence of complications and comorbidities,this study used the latent class analysis with or without covariates clustering the patients with type 2 diabetes,and model fitting was good.Latent class analysis without covariates underestimated the prevalence of comorbidities and complications,therefore,latent class analysis with covariates was recommended because it had superior statistic index and was more reasonable in clinical.3.Risk factor related to complications and comorbidities consisting of older age,female,family history,higher BMI,central obesity,living in around city,higher 2hPG in diagnosis and smoking.4.Items of SF-8 was designed reasonable,having good reliability and validity.Quality life of type 2 diabetes patients in Tianjin was bad,especially the emotion item.Hence,family members and physicians should improve the attention to patient’s emotions.