OBJECTIVES: To apply a risk assessment algorithm to a national US population sample (i) to determine the prevalence of at-risk pre-diabetic and undiagnosed diabetic participants and (ii) to validate the role of the risk factors included in the scoring algorithm in the prediction of diagnosed type 2 diabetes. METHODS: Data from the 2009 US National Health and Wellness Survey (NHWS) were used to apply a modified version of a recently developed scoring algorithm/prediction model to identify patients at-risk for screening for pre-diabetes and undiagnosed diabetes. The algorithm combines six risk factors to generate total risk scores: age, gender, family history of hypertension, high blood pressure, obesity, and physical activity. Estimates of the national prevalence of those at risk together with key health status attributes were generated. RESULTS: Among those 18 years of age and over in the NHWS sample (N=74,474) without a confirmed diagnosis of diabetes, 27.3% of participants were estimated to be at elevated risk for diabetes and a further 16.5% for pre-diabetes. Among those with a confirmed diagnosis of diabetes (excluding persons with type 1 diabetes), 73.6% were confirmed by the scoring algorithm with a further 13.8% meeting pre-diabetic criteria. Overall, 10.0% had a confirmed type 2 diabetes diagnosis. The fitted prediction model indicated the following variables were strongest risk factors in the NHWS sample: i) age 60 years and over (odds ratio 6.94); ii) morbid obesity (odds ratio 6.11); iii) a family history of diabetes (odds ratio 4.79); and (iv) being obese (odds ratio 3.04). All variables significant at the 1% level. The sensitivity was 23.2%, specificity 97.8%, positive predictive value 57.3% and negative predictive value 91.0%, with area under the ROC curve of 0.847. CONCLUSIONS: This study provides strong support for the new scoring algorithm to identify both pre-diabetic and diabetic at-risk populations.