Investigate whether patients with shorter hospital stays are more likely to leave the hospital with undiagnosed diabetes mellitus.

The Kentucky Appalachian Stroke Registry (KApSR) is used to study cerebrovascular disease in Kentucky—comorbidities, diagnostic tests, demographics, and outcome data. KApSR currently provides access to data from over 554,300 individuals seen between 2010 to the present. Diabetes, a common comorbidity in ICVD, is important to diagnose and effectively manage to optimize health outcomes.   

Individuals with ischemic cerebrovascular disease (ICVD) were identified through KApSR using previously published methods.Between 2010 and 2018, KApSR recorded 10,209 hospitalizations for ICVD. The dataset was used to identify common characteristics among patients with undiagnosed diabetes (Hemoglobin A1C level ≥ 6.5). Mean age was 64 (SD: 13.8), BMI 29.3 (7.7), and 48% were female (n=4935). Over 40 demographic factors and comorbidities were identified and studied in the patient population. Chi squared analysis and multinomial regression tests were utilized to study correlations between factors and undiagnosed diabetes. 

6902 (67.6%) subjects had a HbA1C test performed. Of those, 2639 (38.2%) had a HbA1C value ≥ 6.5, and 205 (7.8%) of whom were not given a diagnosis of diabetes despite meeting diagnostic criteria, confirming a gap of undiagnosed diabetes in cerebrovascular disease. Chi square analysis showed a correlation of undiagnosed diabetes with a hospital length of stay < 2 days (P<0.001). Correlation was further confirmed by multinomial logistic regression adjusted for age and sex. 

ICVD short hospitalization is correlated with a higher likelihood of undiagnosed diabetes, suggesting an opportunity to improve care quality. Diabetes was an independent predictor of long-term hospitalization, considering other predictors like NIHSS score, HbA1c, and c-LDL. Shorter hospital stays were found to be associated with undiagnosed diabetes in ischemic cerebrovascular disease patients. This suggests the need for more diligent clinical follow up on HbA1C studies.