Category: Professional Posters
Purpose: To evaluate the impact of intravenous acetaminophen on patient outcomes.
Methods: In this retrospective observational analysis, 54,742 patients were identified from 19 hospitals over a 12 month period. The treatment population for this project was defined as patients who received IV acetaminophen, and the control group were those patients who did not. Propensity score matching was used to identify sets of matched pairs for each outcome. Matching was intended to address the underlying biases in our populations due to patient characteristics. These characteristics included basic patient demographics; structural factors such as facility and provider; and comorbidities, expected length of stay, and average income by zip code. There were 5 outcomes analyzed: total length of stay, ICU length of stay, total narcotic use during admission (measured in morphine milligram equivalents [MME]), whether a patient was discharged with a narcotic prescription, and 30-day readmission rate. Tests were run against balanced populations for 5 procedures: Total Knee Replacement, Total Hip Replacement, Caesarean Section, Coronary Artery Bypass Graft (CABG) surgery, and Gallbladder Resection. A combined population with all procedures balanced between groups was also tested for each outcome. With 5 outcomes and 6 populations, there was a total of 30 possible tests. Due to the limited number of observations with ICU length of stay, only the combined population was tested for that outcome. Therefore, matched populations for 25 tests were created.
Results: After matching, there was no significant difference found between the treatment and control groups for the aforementioned patient characteristics. Six of the 25 tests on the clinical effect of IV acetaminophen showed a significant negative treatment effect. Length of stay was shorter for the control group in the combined population (-0.18 day [4 hours], p<0.001). Total narcotic use was lower for the control group in the Caesarean Section (-10 MME, p<0.001), CABG (-26 MME, p <0.001), and combined (-13 MME, p <0.001) populations. The control group was less likely to be discharged with a narcotic prescription for the Caesarean Section (-5%, p=0.01) and combined (-4%, p<0.001) populations. Three additional tests showed no significant effect on total length of stay, but had power greater than 80%. These tests were on the Total Knee Replacement, Total Hip Replacement, and Caesarean Section populations.
Conclusion: Patients in the control group had a shorter length of stay, lower total narcotic use, and fewer narcotic prescriptions at discharge. No tests showed a significant difference favoring the IV acetaminophen group.