Introduction Adequate pain control is a mainstay in enhanced recovery after surgery (ERAS) protocols. ERAS protocols are widely accepted in colorectal and gynecologic surgeries and are increasingly implemented in the obstetric setting. Multimodal analgesia incorporating non-steroidal anti-inflammatory drugs (NSAIDs) and acetaminophen is a mainstay of ERAS protocols for cesarean delivery, but little research has focused on the choice of NSAIDs or timing of initiation in women undergoing cesarean delivery. At our institution, patients undergoing cesarean delivery receive a standardized multimodal analgesic regimen consisting of neuraxial morphine with NSAIDs and acetaminophen. Our initial protocol involved starting the oral analgesics in the recovery room. There was variability in whether these medications were given in a timely manner or withheld in the setting of postoperative nausea and vomiting. We modified this protocol and performed a retrospective analysis to assess the impact of this change on postoperative opioid rescue requirements in women undergoing cesarean delivery under neuraxial anesthesia. Methods This retrospective analysis included patients who underwent cesarean deliveries from 1 July 2014 to 22 August 2017. With the initial analgesic protocol, patients received neuraxial morphine, followed by naproxen 500 mg PO Q12 hours and acetaminophen 650–975 mg PO Q6 hours initiated in the recovery room. After protocol revision in January 2016, the same neuraxial morphine dose was used in addition to acetaminophen 975 mg PR at the start of the case and ketorolac 15–30 mg IV at the end of the case. Postoperatively, patients received acetaminophen PO 975 mg Q6 hours, ketorolac IV 15 mg Q6 hours for 3 doses, transitioning to ibuprofen 600 mg Q6 hours. Fentanyl, oxycodone, and intravenous hydromorphone were given for breakthrough pain with both protocols. The primary outcome of the study is the need for rescue opioid analgesia. Secondary outcomes are total opioid usage, time to first rescue opioid, maximum reported pain scores, and need for rescue antiemetics. Univariate and multivariate analyses were performed controlling for variables significantly different between the two cohorts. Results 3250 patients were included in our analysis (1574 in the old protocol and 1676 in the new protocol). There was no significant difference in patient demographics or intraoperative characteristics between the two cohorts except for more primiparous women (25% vs. 17%), more Pfannenstiel incision (98% vs. 96%), and less repeat cesarean deliveries (40% vs. 44%) in the new protocol cohort. Need for rescue opioids was reduced with the new protocol at 2, 24, and 48 h [(36.46% vs. 75.73%, p < .0001), (74.28% vs 91.99%, p < .0001), (87.53% vs 95.49% p < .0001), respectively]. Among those who received opioids, opioid consumption over 48 h was reduced (median [IQR]: 55 [30, 95] vs. 40 [20, 70] mg oxycodone equivalents) after protocol revision (GMR 0.75, 95% CI 0.7, 0.80, p < .0001). The time to first rescue opioid medication was significantly longer in the new protocol compared to the old protocol (175 [79, 1057] min vs 51 [28, 104] min, p < .001). Conclusion There was a significant decrease in the need for and the dose of rescue opioid medications with the new protocol. This highlights the importance of optimizing the choice of agents, as well as route and timing of administration of the components of the postoperative multimodal analgesic regimen.