Whether the risk of hospital readmission within 30 days of discharge (early readmission, ER) varies by primary or secondary discharge diagnosis of diabetes (DcDxDM) is unknown.

In a retrospective cohort study of adults with diabetes discharged from an academic medical center, we analyzed 4,027 discharges with a primary DcDxDM and 4,027 discharges with a secondary DcDxDM. A total of 49 characteristics were evaluated for association with ER by multivariable logistic regression models.

The ER rate of discharges with a primary DcDxDM was significantly higher than the ER rate of discharges with a secondary DcDxDM (22.2% vs. 16.2%, p<0.01). Discharges with a primary DcDxDM were significantly different from discharges with a secondary DcDxDM in 44 characteristics (all p<0.04). Among both primary and secondary DcDxDM discharges, lack of an outpatient visit within 30 days of discharge, length of stay, being unemployed, being discharged within 90 days before admission, and anemia were associated with higher ER risk, and being uninsured was associated with lower ER risk (all p<0.05). Among only primary DcDxDM discharges, lower education, gastroparesis, and higher serum creatinine were associated with higher ER risk, and inpatient diabetes consultation, thiazolidinedione use, and higher hematocrit were associated with lower ER risk (all p<0.02). Among only secondary DcDxDM discharges, urgent/emergent admission, abnormal serum sodium or albumin, discharge against medical advice, and pancreatitis were associated with higher ER risk (all p<0.05). C-statistics for the multivariable models of ER were similar between primary and secondary DcDxDM discharges (0.834 vs. 0.822, p=0.15).

Risk factors for ER after a primary DcDxDM discharge differ from risk factors after a secondary DcDxDM discharge. Patients with a primary DcDxDM are at a higher risk of ER and may benefit from dedicated effort to reduce their ER risk.


D.J. Rubin: Research Support; Self; AstraZeneca, Boehringer Ingelheim Pharmaceuticals, Inc. H. Zhao: None. E. Miller: None.


National Institute of Diabetes and Digestive and Kidney Diseases (K23DK102963 to D.J.R.)

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at http://www.diabetesjournals.org/content/license.