Factors associated with glycemic control after an inpatient program.

In this study, we investigated the factors predicting poor glycemic control after an inpatient program. Using the hospital database from April 1999 to May 2003, we retrospectively identified patients with type 2 diabetes mellitus and hemoglobin A(1C) (HbA(1C)) of at least 8.0% at the time of admission for an inpatient program. In the primary analysis, factors potentially related to poor glycemic control (HbA(1C) > or =7.0%) at 6 months after admission were investigated. Stepwise multivariate regression analysis identified the duration of diabetes (odds ratio, 2.43; 95% confidence interval [CI], 1.54-3.82; P < .001), period from the first attendance at our hospital until admission (odds ratio, 1.60; 95% CI, 1.01-2.54; P = .047), and number of admissions (odds ratio, 2.28; 95% CI, 1.36-3.82; P = .002) as predictors of poor glycemic control. In the secondary analysis, factors related to poor glycemic response (an absolute decrease of HbA(1C) by <1.5% from the baseline) at 6 months after admission were investigated. Stepwise multivariate regression analysis identified the duration of diabetes (odds ratio, 2.17; 95% CI, 1.19-3.93; P = .011), period from the first attendance at our hospital until admission (odds ratio, 2.17; 95% CI, 1.43-3.29; P < .001), treatment of diabetes at discharge (oral hypoglycemic agents: odds ratio, 2.52; 95% CI, 1.15-5.51; P = .021; insulin: odds ratio, 4.44; 95% CI, 1.96-10.07; P < .001), baseline HbA(1C) (odds ratio, 0.44; 95% CI, 0.37-0.53; P < .001), and addition of new medications (odds ratio, 0.41; 95% CI, 0.27-0.62; P < .001) as predictors of poor glycemic control.