Background: Length of stay (LOS) outliers are a focus of cost and resource utilization for hospitals in the United States. Limited research has been done to characterize outliers, risk factors, and barriers to discharge.
Methods: We conducted a case-control study of inpatient admissions to the general medicine service line between September 2015 and August 2016. Outliers were defined as observed LOS three standard deviations or higher than predicted and were then matched one-to-one to inlier controls on primary diagnosis and severity of illness. Paired t-tests were used to compare the difference between cases and controls. Conditional logistic regression models adjusted for age, gender and race were used to examine significant predictors of LOS outliers.
Results: 85 outlier patients were matched to inlier controls. Compared to inliers, outlier cases stayed 31.52 (95% CI; 25.01, 33.91) days longer, had 2.55 (95% CI; 2.01, 3.09) higher number of consultants, and were 7.02 (95% CI; 3.83, 10.19) years younger. There was no significant difference between number of comorbidities. Higher odds of being an outlier were observed for patients with mental illness Odds Ratio (OR) 3.24 (95% CI; 1.10, 9.56) and acquired in-hospital complications OR 12.58 (95% CI; 2.46, 64.30). Compared to home discharge, there were higher odds of being an outlier for discharges to a rehab facility OR 15.04 (95% CI; 3.96, 57.14) and hospice OR 30.08 (95% CI; 2.99, 302.24). The odds of being an outlier were higher for Medicaid 29.92 (95% CI; 3.66, 244.32) and Medicare 9.80 (95% CI; 2.27, 42.33) compared to patients with commercial insurance.
Conclusions: Age, race, comorbidities, and lack of insurance are often cited anecdotally as reasons for increased length of stay; however, our study found that when adjusted for age, gender, and race, there was no change to any significant findings and almost all outliers had insurance. Mental illness does appear, however, to be a risk factor for being an outlier. Our study did confirm that in-hospital complications was associated with being a LOS outlier. The higher odds of being an outlier if a patient is discharged to rehab or hospice as opposed to home may point to a delay in placement to outside facilities due to lack of social work resources and a delay in identifying patients who should be transitioned to hospice. Of the outliers with Medicare, only 16% were under 65, which does not entirely account for higher odds of being an outlier with Medicare despite the outlier cohort being younger. Based on our data, it is likely that reducing in-hospital infections and complications, early identification of those needing placement to outside facilities, and increased resources directed to those with mental illness as a comorbidity would reduce hospital costs by decreasing risk factors for LOS outliers.