A Nonparametric Bootstrap Confidence Interval for the Prevalence of Undiagnosed Type 2 Diabetes Mellitus in Hamadan, West of Iran

Background: The prevalence of type 2 diabetes mellitus (T2DM) is increasing in Iran. This study determined the prevalence of undiagnosed type 2 diabetes mellitus in apparently normal individuals in Hamadan, west of Iran. Study design: A cross-sectional study. Methods: A sample of 106 apparently normal volunteers aged 18 yr and more were enrolled, and T2DM was diagnosed using hemoglobin A1c (HbA1c) levels from 2015-2016. The nonparametric bootstrap method was used to eliminate the undesirable effect of small sample size on the estimation of standard error of multiple logistic regression coefficients and confidence interval for the prevalence of undiagnosed T2DM. Results: Overall, 23 (21.69%) were male. The mean (±sd) age of the participants was 43.76 ±14.01 year. In 78.3% of individuals, HbA1c level was within normal range (<5.7), 13.21% was in the range of 5.7-6.4 (undiagnosed pre-T2DM), and 8.49% was ≥6.5 (undiagnosed T2DM). Multiple logistic regression gave the characteristic distribution of volunteers such as sedentary hour (P=0.001), family history of diabetes (P=0.001), smoking (P=0.002), and age (P=0.012) had the odds on the significant effect on undiagnosed T2DM. Conclusions: The prevalence of undiagnosed T2DM among apparently normal individuals in Hamadan was relatively high. Addition to age, factors such as sedentary, exposure to smoking and having a history of diabetes in family can be a prognosis for undiagnosed T2DM in apparently normal individuals