Objectives: The Canadian prevalence and incidence of diabetes has increased by the greatest extent in young adulthood. The original Canadian Diabetes Risk Questionnaire (CANRISK) was created to assess dysglycemia risk among adults ≥40 years of age, but it has not been validated among younger adults. Furthermore, it is unclear whether a young adult-specific risk score would better identify dysglycemia in this age group.
Methods: Analyses were done on participants who completed the self-administered CANRISK and underwent anthropometric and blood glucose measurements, were 18 to 39 years of age, were not pregnant and had no previous diabetes diagnosis. A risk model was generated from a lenient stepwise function fit with predictors identified through univariate analyses. Risk scores were produced from adjusted odds ratios. Model performance was internally validated using bootstrap methods and compared with the original CANRISK prognostic tool.
Results: Of the 3,334 participants included in the study, 194 (5.8%) and 51 (1.5%) were living with prediabetes or undiagnosed diabetes, respectively. The model displayed an area under the curve of 73.0%, adjusted to 72.9% after bootstrapping; however, using the original CANRISK model resulted in similar results (area under the curve, 71.4%). Sensitivity and specificity of the new and original models were also comparable (78.8% vs 77.1% and 54.0% vs 55.0%, respectively).
Conclusions: The original CANRISK score performed well among young adults, even when compared with a young adult-specific model. We suggest that the cut point be lowered for young adults and the tool be permitted for use in this age group.
Keywords: CANRISK; diabète de type 2; dépistage; jeune adulte; prediabetes; prédiabète; risk assessment; screening; type 2 diabetes; young adult; évaluation des risques.
Crown Copyright © 2019. Published by Elsevier Inc. All rights reserved.