The Effect of an Automated Order on Postpartum Opioid Use After Uncomplicated Vaginal Deliveries.

Abstract

Background: To address the opioid epidemic, physicians are encouraged to identify means of reducing patient opioid exposure. Electronic medical records (EMRs) often include default order sets with automated orders for opioid medications, which may influence how much opioids physicians prescribe. We sought to evaluate the impact of de-selecting an automated order for oxycodone-acetaminophen from an EMR order set for postpartum vaginal deliveries on inpatient opioid exposure by comparing the proportion of patients who received an opioid after an uncomplicated vaginal delivery before and after the EMR change. As secondary outcomes, the impact on average total morphine milligram equivalents (MMEs) and discharge opioid prescriptions was investigated. Materials and Methods: A quality improvement study was conducted through retrospective chart review of uncomplicated vaginal deliveries for the four quarters before and after the EMR order set change occurred. The total proportion of patients who received an opioid in the postpartum period was then determined for the preexposure and postexposure groups. The total average MME consumed for patients who received an opioid in each group was determined and the total proportion of patients who received an opioid prescription at discharge was compared. Results: A total of 5826 records of uncomplicated vaginal deliveries met the criteria for analysis. In the preintervention group, 32.9% of patients received an opioid postpartum, compared to 12.5% of patients in the postintervention group, representing a decrease of 62.0% (p < 0.001). Of those who received opioids, the preintervention mean total opioid consumption was 28.4 MME (±27.6) compared to 33.6 MME (±46.4) postintervention, and there was no significant difference in median total opioid consumption: 22.5 MME (interquartile range [IQR]: 7.5-47.5) preintervention compared with 20.8 MME (IQR: 7.5-45.0) postintervention (p = 0.902). No significant difference was found with discharge opioid prescriptions between the two groups. Conclusion: Order sets within EMR systems appear to have a significant influence on physician prescribing behaviors and removing these automated orders for opioids should be considered.