Background and aims: Diabetes is a growing global health problem, and its early detection is very important. Due to low sensitivity, blood glucose and HaA1c screening is not recommended. Therefore, a number of country-specific risk scores have been developed as useful primary screening tools for detecting diabetes. The aim of this study was to develop and measure the accuracy of a diabetes risk score model for type 2 diabetes mellitus in a Kuwaiti population.

Methods: A total of 3915 randomly selected Kuwaiti men and women aged 18-69 years were surveyed in 2014 using the WHO Instrument for Chronic Disease Risk Factor Surveillance comprising demographics, medical and family history, physical measurements and blood biochemistry. Diabetes was defined as fasting plasma glucose ≥7mmol/L, HbA1c ≥6.5% or being treated with glucose-lowering drugs.  A logistic regression model of known diabetes risk factors was created and the specificity, sensitivity and receiver-operating characteristics (ROC) curves were assessed.

Results: The final Kuwait risk score model included age >45, BMI>25 kg/m2, waist circumference >88cm for women and >102cm for men and history of hyperglycemia and cholesterol. Optimum sensitivity and specificity of the Kuwait risk score (cutoff point 8 out of 20) was 67.1% and 57.2%. The area under the curve was 0.68. This Kuwaiti risk score performed better than the Finnish, Thai, Colombian and Chinese risk scores but not as well as the American or Omani risk scores. 

Conclusion: A simple, inexpensive screening tool designed to detect individuals at high risk for type 2 diabetes in Kuwait was created.

A Kuwaiti risk score for detecting diabetes in adults