Background: Prevalence of alcohol use disorder (AUD) on inpatient medicine wards exceeds 40%. Only 50% of these patients develop alcohol withdrawal, of whom only 5-20% develop complicated alcohol withdrawal syndromes (AWS) that warrant treatment and monitoring. Overutilization of the Clinical Institute Withdrawal Assessment of Alcohol Scale, Revised (CIWA-Ar) scale is well-reported in the literature. CIWA-Ar is labor-intensive, and inappropriate use may increase benzodiazepine exposure and associated complications. The Prediction of Alcohol Withdrawal Severity Scale (PAWSS) is a validated screening tool used to identify patients at high-risk of complicated AWS.
Purpose: Our goal was to implement PAWSS screening in patients with AUD to select for patients at high-risk for complicated AWS. We aimed to limit CIWA-Ar monitoring to only those patients with positive PAWSS screening, thereby reducing unnecessary CIWA-Ar usage, and inappropriate benzodiazepine exposure and associated complications.
Description: Our study is a pre- and post-intervention quality improvement study done in patients with AUD admitted to medicine units from March 2018 to September 2019 at a single, large, urban academic medical center. An order was created in the electronic medical record for PAWSS screening with reflex CIWA-Ar if PAWSS scores was > 4. We implemented a broad educational initiative regarding use of PAWSS screening. Our pre-intervention data from 03/2018-06/2019 (n=1190) showed nearly all patients with documented AUD had CIWA-Ar protocol ordered (95.8%), however only 48.1% of these patients ever had a CIWA-Ar score high enough to necessitate treatment (CIWA-Ar >8). Despite this, 64% of these patients were treated with benzodiazepines during admission. Post-intervention, PAWSS screening was ordered for 30.3% of patients with AUD from 06/2019-09/2019 (n=145). In patients with PAWSS ordered, post-intervention data showed a decrease in the duration of CIWA-Ar orders (59 hours vs 69 hours) as well as an increase in peak CIWA-Ar scores (10.32 vs 9.55). Compared to our pre-intervention group, patients with PAWSS screening had decreased benzodiazepine exposure (57% vs 63%), decreased mortality (2.38% vs 3.87%), decreased ICU admissions (9.52% vs 18.15%), decreased delirium (2.38% vs 3.95%), decreased falls (0 vs 2.77%), and decreased length of stay (4.21 days vs 5.61 days).
Conclusions: Our data suggest CIWA-Ar protocol is overutilized in patients at low-risk of complicated AWS. Consequently, the majority of patients are exposed to benzodiazepines during admission despite low CIWA-Ar scores, suggesting unnecessary benzodiazepine exposure. Implementation of a protocol to screen for high-risk patients using PAWSS and limiting CIWA-Ar usage to patients with positive PAWSS scores led to an increase in peak CIWA-Ar scores, confirming PAWSS successfully selected for patients at high-risk for complicated AWS. Further, use of CIWA-Ar protocol appropriately decreased in low-risk patients. Results from our initiative suggest that the PAWSS screening tool can safely be introduced on inpatient medicine services to reduce unnecessary CIWA-Ar monitoring, as well as reduce benzodiazepine exposure and associated complications.