EXTERNAL VALIDATION OF THE ADA DIABETES RISK SCORE FOR POPULATION-LEVEL DIABETES SCREENING IN THE UNITED ARAB EMIRATES
Author(s)
Buthaina AlMuhairi, MD1, Marwa Mahmoud Khalil, MD2.
1Non-Communicable Diseases and Mental Health Department, Ministry of Health and Prevention (MoHAP), Dubai, United Arab Emirates, 2Non-Communicable Diseases and Mental Health Department, 1- Ministry of Health and Prevention (MoHAP), 2-Public Health Department, Faculty of Medcine, 1- Dubai 2- Zagazig, Egypt, United Arab Emirates.
1Non-Communicable Diseases and Mental Health Department, Ministry of Health and Prevention (MoHAP), Dubai, United Arab Emirates, 2Non-Communicable Diseases and Mental Health Department, 1- Ministry of Health and Prevention (MoHAP), 2-Public Health Department, Faculty of Medcine, 1- Dubai 2- Zagazig, Egypt, United Arab Emirates.
OBJECTIVES: Prediabetes and diabetes are priority public health concern in the UAE. This study evaluated the diagnostic accuracy of the ADA Diabetes Risk Score to identify undiagnosed dysglycemia.
METHODS: A cross-sectional community campaign (September 2023-February 2024) assessed asymptomatic adults using the ADA Diabetes Risk Score. Venous HbA1c test was used to classify people as normal (<5.7%), prediabetic (5.7-6.4%), or diabetic (≥6.5%). The risk score was evaluated at cutoffs ≥4, ≥5, and ≥6 to calculate diagnostic performance parameters (sensitivity, specificity, positive predictive value [PPV], negative predictive value [NPV]). Independent predictors were evaluated using adjusted multivariable logistic regression. The score’s discriminatory power was assessed using the Receiver Operating Characteristic curves.
RESULTS: Among 4,203 participants (mean age 42.8±10.4 years; 67.8% male), prediabetes prevalence was 30.3% vs. 11.8% for diabetes. A risk score ≥5 was observed in 87.4% of the cohort. A score ≥4 demonstrated 97.3% sensitivity, 7.4% specificity, 43.3% PPV, and 79.2% NPV; score ≥5 showed 90.9% sensitivity, 15% specificity, 43.8% PPV, and 69.4% NPV. A score ≥6 provided the best balance between sensitivity (44%) and specificity (65.9%), with 48.4% PPV and 61.8% NPV. Discriminatory power was poor (AUC=0.566). Older age (OR=1.03, 95%CI: 1.02-1.04), male sex (OR=1.97; 95% CI: 1.66-2.34), family history (OR=1.23; 95%CI: 1.07-1.42), gestational diabetes history (OR=1.41; 95%CI: 1.11-1.81), obesity (OR=1.25; 95%CI: 1.10-1.43), and elevated blood pressure were significantly associated with increased odds of dysglycemia.
CONCLUSIONS: The ADA Diabetes Risk Score poorly discriminated between dysglycemic and nondiabetic persons in the UAE. This emphasizes the need for a country-specific risk score developed from local real-world data and informed by AI predictive models. This will allow for timely implementation of lifestyle modifications and targeted pharmacotherapy, to prevent progression to diabetes. Until then, routine HbA1c screening is recommended for early identification of at-risk individuals in the UAE.
METHODS: A cross-sectional community campaign (September 2023-February 2024) assessed asymptomatic adults using the ADA Diabetes Risk Score. Venous HbA1c test was used to classify people as normal (<5.7%), prediabetic (5.7-6.4%), or diabetic (≥6.5%). The risk score was evaluated at cutoffs ≥4, ≥5, and ≥6 to calculate diagnostic performance parameters (sensitivity, specificity, positive predictive value [PPV], negative predictive value [NPV]). Independent predictors were evaluated using adjusted multivariable logistic regression. The score’s discriminatory power was assessed using the Receiver Operating Characteristic curves.
RESULTS: Among 4,203 participants (mean age 42.8±10.4 years; 67.8% male), prediabetes prevalence was 30.3% vs. 11.8% for diabetes. A risk score ≥5 was observed in 87.4% of the cohort. A score ≥4 demonstrated 97.3% sensitivity, 7.4% specificity, 43.3% PPV, and 79.2% NPV; score ≥5 showed 90.9% sensitivity, 15% specificity, 43.8% PPV, and 69.4% NPV. A score ≥6 provided the best balance between sensitivity (44%) and specificity (65.9%), with 48.4% PPV and 61.8% NPV. Discriminatory power was poor (AUC=0.566). Older age (OR=1.03, 95%CI: 1.02-1.04), male sex (OR=1.97; 95% CI: 1.66-2.34), family history (OR=1.23; 95%CI: 1.07-1.42), gestational diabetes history (OR=1.41; 95%CI: 1.11-1.81), obesity (OR=1.25; 95%CI: 1.10-1.43), and elevated blood pressure were significantly associated with increased odds of dysglycemia.
CONCLUSIONS: The ADA Diabetes Risk Score poorly discriminated between dysglycemic and nondiabetic persons in the UAE. This emphasizes the need for a country-specific risk score developed from local real-world data and informed by AI predictive models. This will allow for timely implementation of lifestyle modifications and targeted pharmacotherapy, to prevent progression to diabetes. Until then, routine HbA1c screening is recommended for early identification of at-risk individuals in the UAE.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
EPH59
Topic
Epidemiology & Public Health, Patient-Centered Research, Real World Data & Information Systems
Topic Subcategory
Public Health
Disease
Diabetes/Endocrine/Metabolic Disorders (including obesity)