Impact of an Alcohol Withdrawal Treatment Pathway on Hospital Length of Stay: A Retrospective Observational Study Comparing Pre and Post Pathway Implementation.

OBJECTIVE: To determine if the implementation of a hospital-specific alcohol withdrawal treatment pathway used in a medical-surgical patient population decreased hospital length of stay (LOS) compared with the standard of care.

METHODS: This retrospective observational study, conducted in a large academic tertiary care hospital, involved 582 subjects who met criteria for study inclusion, with 275 subjects in the 2010 cohort and 307 in the 2012 cohort. The Alcohol Withdrawal Project Team was formed with the goal of creating a standardized approach to the recognition and treatment of alcohol withdrawal at Duke University Hospital. The group created a computerized physician order entry alcohol withdrawal treatment pathway with 4 possible treatment paths chosen on the basis of current withdrawal symptoms, vital signs, and alcohol withdrawal history. The 4 treatment paths are 1 prophylaxis; 2 mild-to-moderate withdrawal; 3 moderate-to-severe withdrawal, and 4 severe withdrawal/alcohol withdrawal delirium. Each treatment path corresponds to a different lorazepam dose and dose schedule and symptom assessment. This pathway was implemented in the hospital at the end of 2011.

RESULTS: Using a Cox proportional hazards model and adjusting for covariates, there was a 1 day [95% confidence interval (CI), 1-2 d] reduction in median hospital LOS between the 2010 and 2012 cohorts, 5 versus 4 days, respectively. The average ratio in hospital LOS between the 2 cohorts was 1.25 (95% CI, 1.25-1.67). The CI was estimated by bootstrapping and indicated a significantly longer LOS in the 2010 cohort compared with the 2012 cohort. Nonsignificant changes were found in the proportion of subjects admitted to the intensive care unit (24% in 2010 vs. 29.3% in 2012), LOS in the intensive care unit (7.1±8 d in 2010 vs. 5.6±6.9 d in 2012), and proportion of patients discharged with a diagnosis of delirium tremens (17.8% in 2010 vs. 15.3% in 2012).

CONCLUSIONS: This study demonstrates the successful implementation of an alcohol withdrawal treatment pathway in a medical-surgical population hospitalized in a large tertiary care facility with significant impact on hospital LOS.