Patient-Specific Opioid Volume Calculations Reduce Discharge Prescriptions After Hepato-Pancreato-Biliary Surgery

Presenter: Ryan Day MD | The University of Texas MD Anderson Cancer Center Background: Patients who undergo hepato-pancreato-biliary (HPB) surgery and receive discharge opioid prescriptions beyond their actual needs are at risk for persistent use and diversion to family and community. In the context of the United States opioid epidemic, this study’s aim was to quantify and compare HPB surgery discharge prescribing patterns balanced against 30-day refill rates to compare a novel standardized prescription calculation vs. usual care. Methods: This is a non-randomized retrospective cohort study of consecutive HPB operations at a Comprehensive Cancer Center (September 2018 – June 2019). These dates were based on departmental opioid reduction education initiatives in August 2018, which included the introduction of a novel patient-specific “5x multiplier” concept, wherein patients received 5-times their actual opioid use in the last 24 hours of their inpatient stay. This concept was adopted by a voluntary proportion of our individual HPB providers. Thus, there were two prescribing patterns: a usual care group which typically received a round number of opioid pills based on provider preference vs. the “5x multiplier” calculation. Refills were abstracted from the electronic medical record. Actual opioid doses (oral morphine equivalents, OME) in the last 24 hours of inpatient stay and discharge prescriptions were recorded and converted with institution-approved conversion tables (e.g. one 50-mg tramadol equals 5mg OME, and one 5-mg oxycodone equals 7.5mg OME). Descriptive statistics were used to summarize and compare the usual care and 5x-multiplier cohorts. Results: There were 278 consecutive HPB operations for 276 unique patients. These included 152 (55%) liver and 126 (45%) pancreas resections. Of 278 operations, 125 (45%) were in the 5x-multiplier cohort. Both groups had similar demographic and peri-operative characteristics including length of stay and use of non-opioid medication bundles. The median OME consumed during the last 24 hours before discharge was 10mg (range 0-360mg; IQR 4-20mg) in the total 278 index hospitalizations. The median last 24-hr OME was 10mg (range 0-200mg, interquartile range [IQR] 5-20mg) for usual care vs. 10mg (range 0-360mg, IQR 0-20mg) for the 5x-multiplier. The median discharge prescription OME was 3-times higher in the usual care group (150mg, range 0-3,150mg, IQR 100-150mg) vs. 50mg (0-1,800mg, IQR 0-100mg, p<0.001). Despite this reduction, rates of opioid refills within 30 days were similar for the usual care and 5x-multiplier groups (20.9% vs. 16.8%, p=0.385). Pancreas surgery patients were more likely to receive a postoperative 30-day refill compared to liver surgery patients (32/126, 25%, vs. 21/152, 14%, p<0.001). Among all patients, 69 (25%) patients used zero OME in the last 24 hours. However, 31/69 (45%), were still discharged with opioids. Using the 5x multiplier for the 153 usual care discharge prescriptions would have hypothetically saved 15,761mg OME, or 3,152 pills of 50-mg tramadol over this 10-month study period. Conclusion: In patients undergoing HPB operations, a simple, reproducible strategy using a patient-specific 5x-multiplier of actual last 24-hour inpatient opioid consumption reduced median discharge opioid prescriptions by 67% over usual care, with no measurable increase in refills. The next step will be a department-wide prospective quality improvement project implementing this novel 5x multiplier.