Table 4 Previous investigation on the external validation of prediction models for undiagnosed/ incident type 2 diabetes using Tehran Lipid and glucose study data
Prediction model | Population/year | Validation Data | Risk predictors in the final model | Statistical models | C-index/AUC | Calibration |
|---|---|---|---|---|---|---|
San Antonio heart study diabetes prediction model [22] | Mexican Americans and Hispanic white/2010 | TLGS cohort / the 6.3-year incidence of T2DM | age, sex, ethnicity, SBP, HDL-C, BMI, FH-DM, FPG | Logistic regression | 0.83 | Acceptable after recalibrated |
The Saint Antonio Diabetes Prediction Model (SA) [23] | Mexican Americans and Hispanic white/2021 | TLGS cohort / The 5-year incidence of T2DM | age, sex, ethnicity, SBP, HDL-C, BMI, FH-DM, FPG | Logistic regression | 0.81 | Acceptable after recalibrated |
Atherosclerosis Risk in Communities Study (ARIC) [24] | American population/2013 | TLGS cohort / The 6-year incidence of T2DM | Age, FH-DM, hypertension, WC, height TG, HDL-C, FPG, race | Cox regression | Men 0.790 Women 0.829 | Acceptable after recalibrated |
Atherosclerosis Risk in Communities Study (ARIC) [23] | American population/2021 | TLGS cohort / The 5-year incidence of T2DM | Age, FH-DM, hypertension, WC, height TG, HDL-C, FPG, race | Logistic regression | 0.83 | Acceptable after recalibrated |
Finnish Diabetes Risk Score (FINDRISC) (4) | Finish population/2019 | TLGS cohort / For undiagnosed T2DM | age, BMI, WC, physical activity, daily consumption of fruits, berries, or vegetables, and the history of antihypertensive drug treatment and history of high blood glucose to predict drug-treated diabetes | Logistic regression | 0.75 | Acceptable after recalibrated |
Australian Type 2 Diabetes Risk Assessment Tool (AUSDRISK) (4) | Australian population/2019 | TLGS cohort / For undiagnosed T2DM | Non-invasive model: age, sex, ethnicity, FH-DM, history of high blood glucose level, use of antihypertensive medications, smoking, physical inactivity, and WC | Logistic regression | 0.77 | Acceptable calibration |
Australian Type 2 Diabetes Risk Assessment Tool (AUSDRISK) for undiagnosed diabetes [23] | Australian population/2021 | TLGS cohort / The 5-year incidence of T2DM | Invasive model: age, race, FH-DM, FPG, SBP, WC, height, HDL-C, and TG | Logistic regression | 0.77 | Acceptable after recalibrated |
American Diabetes Association Risk Score (ADA) [4] | American population/2019 | TLGS cohort/ For undiagnosed T2DM | Age, sex, FH-DM, history of hypertension, obesity, and physical activity | Logistic regression | 0.73 | Acceptable after recalibrated |
risk assessment tool for cardiovascular disease, type 2 diabetes, and chronic kidney disease [25] | Dutch population/2020 | TLGS cohort / For undiagnosed T2DM | Sex stratified analysis: age, BMI, WC, use of antihypertensive medications, current smoking, parent and/or sibling with MI or stroke (age < 65 years), FH-DM | Logistic regression | Men 0.65 Women 0.69 | Not acceptable |
American Diabetes Association screening tool [21] | American population/2020 | national survey of risk factors for non-communicable diseases / For undiagnosed T2DM | Age, sex, FH-DM, history of hypertension, obesity, and physical activity | Logistic regression | 0.737 | Not reported |
The Framingham Offspring Study (FOS) risk score [23] | American population/2021 | TLGS cohort / The 5-year incidence of T2DM | age, gender, FPG, BMI, WC, HDL-C, SBP, FH-DM | Logistic regression | 0.82 | Acceptable after recalibrated |
REasons for Geographic And Racial Differences in Stroke (REGARDS) / Current study | American population | TLGS cohort / The 10-year incidence of T2DM | Age, sex, BMI, SBP, DBP, HDL-C, TG, FPG, and race | Baysian logistic regression | 0.79 | Acceptable calibration |