Implementing a Protocol to Reduce Opioid Prescriptions in Military Otolaryngology: A Quality Improvement Initiative

Introduction: To compare pain medication refill rates for adult septoplasty and rhinoplasty patients before and after initiating a multimodal analgesic protocol for reducing opioid prescriptions (PROP). Materials and Methods: Data from 58 adult patients were retrieved by retrospective chart review (19 septoplasties and 10 rhinoplasties before initiating PROP in September 2018 and 21 septoplasties and 8 rhinoplasties after PROP). We selected consecutive sep- toplasties and rhinoplasties, at which time a new discharge order set was implemented. The new order set consisted of 10 oxycodone tabs (5 mg), 100 acetaminophen tabs (325 mg), and 28 celecoxib tabs (200 mg). The primary outcome variable was the number of initial opioid prescriptions and refills filled by any provider. Results: Among the septoplasties, there was a 46% decrease in total morphine milligram equivalent (MME) prescribed, from a mean of 202.0 mg in the non-PROP group (95% CI, 235.4, 174.6) to 108.6 mg in the PROP group (95% CI, 135.8, 81.4), with no difference in refill rates. Among the rhinoplasties, there was a 51% decrease in total MME prescribed, from a mean of 258.8 mg in the non-PROP group (95% CI, 333.4, 184.1) to 126.6 mg in the PROP group (95% CI, 168.1, 85.0) with no difference in refill rates. Conclusions: The outcomes after PROP implementation for septoplasty and rhinoplasty at our institution suggest that opioid prescription rates can be significantly decreased to manage postoperative pain, with no difference in opioid refill rates. The results also warrant further investigation into patient pain, satisfaction, provider efficiency, and healthcare costs.


INTRODUCTION
The opioid epidemic is well known among the medical and surgical communities nationwide. According to the U.S. Department of Health and Human Services Centers for Disease Control and Prevention, a record number of 63,632 drugrelated deaths occurred in 2016 in the United States. 1 The CDC reports that 66.4% of these drug-related deaths involved prescription and/or illicit opioid drugs. 1 Recent studies show that 1 in 16 of postoperative patients becomes a long-term user of opioids with many patients receiving their first exposure to opioids after surgery. Meanwhile, approximately 50 million U.S. citizens are estimated to undergo ambulatory surgery annually in the United States. 2 Opioid-prescribing trends appear to be declining in the past decade. The CDC reports that from 2006 to 2017, the annual prescribing rate per 100 persons decreased from 72.4 to 58.5 for all opioids, which is an overall relative reduction of 19.2%. 1 Several observational and randomized controlled studies across the country are increasingly focused on decreasing opioid rates in the postoperative period. Many of these studies are utilizing the practice of multimodal analgesia in order to encourage both patients and providers to manage pain with nonopioid medications as first-line alternatives. [3][4][5] Studies across general surgery and orthopedics have influenced other surgical specialties to investigate the use of nonopioid pain alternatives, such as acetaminophen (Tylenol) and nonsteroidal anti-inflammatory medications (NSAIDs), as first-line alternative medications for managing outpatient surgical postoperative pain and the surgeon's role in decreasing opioid prescription rates for a surgical outpatient. [5][6][7] In the last few years, otolaryngology has joined the movement of investigating how to primarily manage patients on nonopioid medications. A handful of studies have analyzed the opioid-prescribing patterns for outpatient otolaryngology surgeries. Patel and colleagues demonstrated that some rhinoplasty patients might not require more than 15 tablets of hydrocodone-acetaminophen tablets, whereas others may only require as few as eight hydrocodone-acetaminophen tablets postoperatively. 8 A recent single-blinded randomized control trial by Nguyen and colleagues demonstrated that ibuprofen might be used as an alternative postoperative pain medication and be just as effective as opioid medication in outpatient otolaryngology procedures. 9 An evidence-based review by Nguyen and colleagues in 2018 compared 37 studies utilizing various pain medication regimens for septoplasty and rhinoplasty, and the majority of these studies focused on the utilization of local anesthesia intraoperatively and its impact on perioperative and postoperative analgesic requirements. Meanwhile, five of the studies in the review compared the use of NSAIDs with other methods of pain control, and none of these studies reported excessive bleeding or a postoperative hematoma, which is often a concern that surgeons consider when using NSAIDs postoperatively. 10 Overall, there is increasing support in the literature to adopt the practice of multimodal analgesia and reduction of opioid prescriptions across the surgical subspecialty of otolaryngology.
Septoplasty and rhinoplasty are two of the most common surgeries performed in our tertiary academic military hospital within otolaryngology. In 2019, a study by Aulet and colleagues demonstrated a statistically significant reduction in morphine milligram equivalent (MME) prescribed to patients following septoplasty and rhinoplasty without an increase in postoperative patient complaints, pain, or a need for a second prescription. 11 We hypothesized that rhinoplasty and septoplasty were reasonable target outpatient surgeries to develop a standardized protocol for postoperative pain control in order to reduce opioid prescription rates without increasing refill rates. As a quality improvement (QI) initiative, our long-term goals are to reduce the opioid prescription rates in our department and determine the effect of this reduction on refill rates without increasing pain or patient dissatisfaction outcomes.

MATERIALS AND METHODS
A multimodal analgesic protocol was implemented as a QI initiative at our tertiary academic military hospital in September 2018. The initial targeted population comprised of adults 18 years and older undergoing an outpatient septoplasty (with or without turbinoplasty) or rhinoplasty (with or without rib graft harvest and/or osteotomies). This study was reviewed by the Tripler Army Medical Center institutional review board and approved for an exemption; patient consent was not obtained as data were acquired in a deidentified manner. SQUIRE 2.0 guidelines were adapted for implementing our intervention and reporting our findings. 12

Intervention
We selected consecutive septoplasties and rhinoplasties before and after which time a new postoperative medication discharge order set was made available for providers to utilize in the hospital-wide computer system, CHCS. The pain medications included in the new order set consisted of 10 oxycodone immediate-release tabs (5 mg), 100 acetaminophen (Tylenol) tabs (325 mg), and 28 celecoxib (Celebrex) tabs (200 mg). Patients were instructed in the discharge paperwork to use acetaminophen and celecoxib tablets as first-line scheduled pain medications unless they had drug allergies and/or other medical contraindications; acetaminophen was instructed to be used in the dosage of 650 mg every 6 hours and celecoxib in the dosage of 200 mg every 12 hours. Patients were instructed to use oxycodone immediate-release tablets for severe breakthrough pain (5 mg every 4-6 hours), only if pain was not controlled with the first-line pain medications. As part of our routine postoperative discharge instructions, patients were instructed to contact our department through our day or night patient call lines for any postoperative issues, medication refill requests, and/or to return to our emergency department for any acute or urgent medical needs. At the time of this study, surgeons from our department consisted of 7 otolaryngology attending physicians (including 1 fellowshiptrained facial plastics surgeon) and 10 otolaryngology resident physicians. Resident physicians were either the primary surgeon or the assistant surgeon in the cases in which they were involved. Nursing staff provided postoperative education regarding medication administration without deviating from prior standard of care; preoperative education varied from provider to provider and was not implemented as a part of the protocol.

Outcomes and Variables
The primary outcome measured was the number of opioids prescribed and the proportion of any and all types of pain medication refills by any provider within our institution for 21 days postoperatively. Other data collected included patient gender, patient age, surgery performed, medications filled at the time of discharge with details such as medication name, dose, and quantity. The authors also investigated if the patient returned to the emergency room, another provider within the hospital (i.e., primary care provider), a different otolaryngologist within the same department, or the original surgeon for a refill of opioid or nonopioid refill medication within the first 21 days following surgery. The authors also reviewed each chart for any postoperative complications and/or mortality within the first 30 days postoperatively. The authors also noted if the patient was already prescribed opioid medication at the time of surgery; however, these patients were not excluded from the study or were their results stratified as the number of these patients was negligible.

Inclusion and Exclusion Criteria
Data from 58 adult patients were retrieved retrospectively and deidentified (all patients were assigned a randomized number). There were 19 septoplasties and 10 rhinoplasties before initiating protocol for reducing opioid prescriptions (PROP) and 21 consecutive septoplasties and 8 consecutive rhinoplasties after the PROP period. Regardless of whether or not providers opted to use the new order set, any adult patient undergoing an outpatient septoplasty or rhinoplasty during the study period was included. Patients undergoing additional surgeries (i.e., rhinoplasty combined with uvulopalatoplasty and/or tonsillectomy) were excluded from the study. Surgeries performed by all providers were included. Patients meeting the criteria of age of 18 years and older, any gender, and with any preexisting medical condition(s) were included in the study, provided they were discharged from the hospital the same day of the surgery.

Data Analysis
The data abstracted from clinical records was deidentified and entered into a data collection system using Microsoft Excel. The total amount of opioid tablets for each patient and tablets were converted into the mean prescribed total MME doses using the CDC MME conversion table. MME was recorded for initial prescription, refills, and total amount for each patient. The MME was compared before and after the PROP period using one-tailed independent t-tests. The means, 95% CIs, and comparisons were made in the PROP group and non-PROP group. The proportion of patients that received a pain medication refill was compared before and after the PROP implementation using Fischer's exact test for small sample sizes. A post hoc power analysis was undergone for both the prescribed MME difference and refill rates before and after the PROP implementation period.  Table I.

RESULTS
Post hoc power analysis of the difference between prescribed opioids from the non-PROP group to the PROP group revealed a power of 0.99 and 0.917 for the septoplasty group and rhinoplasty group, respectively. Post hoc power analysis of the difference between pain medication refill rates from the non-PROP group to the PROP group revealed a power of 0.03 and 0.06 for the septoplasty group and rhinoplasty group, respectively. There were no postoperative complications or mortalities observed in either group.

DISCUSSION
This retrospective study was designed to evaluate current opioid-prescribing practices for outpatient adult septoplasty and rhinoplasty surgeries and to propose a multimodal protocol in order to reduce the amount of total opioid prescribed postoperatively. Our QI initiative led to a 46% decrease in MME prescribed in the septoplasty group and a 51% decrease in the MME prescribed in the rhinoplasty group with no significant increase in refill rates among the 58 patients included. The mean MME prescribed post-PROP was 108.6 ± 63.6 mg for the septoplasty group, and the mean MME prescribed post-PROP was 126.6 ± 59.9 mg for the rhinoplasty group.
An expert panel consensus released by Johns Hopkins University in 2018 reviewed 20 surgeries, recommending ibuprofen use in all surgeries with a maximum of 20 doses of opioid prescriptions as well as 0 dose for 3 outpatient surgeries. 13 Surgical specialties such as orthopedics, obstetrics and gynecology, and general surgery have demonstrated success in decreasing opioid prescription rates across various studies through written guidelines for providers, standardization of prescription rates, and educational handouts for patients. 4,5,7,14 Sekhri and colleagues demonstrated that the probability of refilling prescription opioids after surgery was not correlated with initial prescription strength, implying that surgeons may be overprescribing opioids and could likely prescribe less narcotics postoperatively without experiencing requests for refills. 7 In 2016, the American Pain Society, in conjunction with the American Society of Regional Anesthesia and Pain Medicine, and American Society of Anesthesiologists, released guidelines on the management of postoperative pain. In these guidelines, acetaminophen and NSAIDs are recommended as part of a multimodal analgesia for postoperative pain. The authors depict a pattern from multiple studies that pain seems to be more effectively managed when multimodal analgesia is utilized postoperatively. 15 The underlying concept of multimodal analgesia is that pain may be managed with less opioid overall through targeting different pain receptor pathways. 15 For example, acetaminophen and NSAIDs have different mechanisms of action and therefore are able to blockade different pain pathways other than the mu receptor targeted by opioids. Although acetaminophen's exact mechanism of action is yet to be elucidated, it theoretically lacks peripheral anti-inflammatory properties and thus works primarily in the central nervous system, inhibiting prostaglandin production, possibly stimulating serotonergic and cannabinoid pathways, which leads to its antipyretic and analgesic effects. 16 NSAIDs have become a popular first-line postoperative pain alternative to opioids and are categorized into nonselective (ibuprofen and naproxen) and selective (celecoxib) classes. 17 Nonselective NSAIDs work by reversibly inhibiting the cyclooxygenase enzymes and decreasing the formation of prostaglandin precursors. Celecoxib is a U.S. FDA-approved selective COX-2 inhibitor and has become popular because of its comparable analgesia to nonselective NSAIDs with the reduction in risk of gastrointestinal (GI) bleeding. Pharmacologic studies and a small randomized controlled trial have demonstrated that because of celecoxib's absence of COX-1 activity, it may not interfere with normal platelet aggregation and hemostasis, and this has been demonstrated in a systematic review and meta-analysis of 35 studies, which demonstrated no increased risk of postoperative bleeding. 18,19 Cochrane Reviews of randomized controlled trials demonstrated that the use of celecoxib (200 or 400 mg orally) decreases the need for rescue opioid pain medication without serious side effects. 17 This makes celecoxib a much more appealing postoperative NSAID choice as compared to COX-1 inhibitors. We have chosen celecoxib rather than ibuprofen as part of the multimodal protocol because of theoretic decreased risk of potential surgical bleeding and GI bleed rates, given that we hope to expand our protocol to other otolaryngology outpatient procedures to include other same day procedures, which pose a risk of increased postoperative bleeding. 20 In our study, there were no incidents of postoperative epistaxis, septal hematoma, or GI bleeds.
Recent trends in otolaryngology demonstrate a growing interest in decreasing postoperative opioid prescriptions by using alternative nonopioid medications. 3,8,9 Recent studies have demonstrated that fewer than 10 tablets of hydrocodoneacetaminophen (5 mg/325 mg) can be prescribed postoperatively for select outpatient otolaryngology surgeries without impacting patient satisfaction or reported pain scores. 8,9 Scalfani and colleagues report that surgeons can reliably prescribe as few as 11 opioid tablets to septoplasty and rhinoplasty patients without increasing pain scores, which is supported through the use of the study obtaining postoperative 10-point visual analog scale pain scores. 21 Most published protocols in otolaryngology at this current time primarily focus on head and neck-related surgery with the adoption of multimodal analgesia protocols translating well to thyroidectomy and parathyroidectomy surgeries, with studies in the last 2 to 3 years focusing more on elective outpatient surgeries such as septoplasty and rhinoplasty. 22 Although there has been a shift of attention toward postoperative opioid prescription patterns, especially in elective outpatient procedures, some providers have reported hesitation in limiting patients' initial opioid prescription out of concern that it may cause additional postoperative visits and unnecessary inconvenience to patients and/or providers. 7 Patel and colleagues published that providers may be overprescribing opioid as patients were prescribed between 20 and 30 hydrocodone-acetaminophen tablets after rhinoplasty; however, they only used an average of 8.7%. 8 Our primary outcome measure was to analyze refill rates in order to encourage other providers in our department that limiting the initial opioid prescription may not necessarily lead to more work in the future or more burden on the patient.
Septoplasty and rhinoplasty within our hospital are primarily performed for functional reasons to address nasal obstructive breathing, given that a large portion of our patient population are active duty military who have endured nasal trauma secondary to combat, training injuries, and/or prior trauma during sports. This population is typically in their twenties or thirties and therefore subject to a higher rate of prescription opioid abuse and/or accidental drug overdose as well as self-prescribing in the future, lending to greater risk of chronic opioid dependency. 23 The patient, friends, and/or family member may use leftover opioids recreationally in the future. 24 The significance of reducing leftover opioid impacts short-and long-term risks for patients and dependents.
Another significant component of reducing postoperative pain is managing patient expectations by discussing pain management and how to effectively manage pain in the preoperative appointment. The preoperative appointment is a critical time to perform safety checks by reviewing medical history and medical reconciliation but to also have an open discussion with the patient about the postoperative plan including follow-up and prescriptions. We believe that our protocol is best implemented when the patient is educated before surgery rather than the day of or after surgery; however, further research is necessary to best determine the impact of patient education. Educational intervention is supported in multiple other studies in addition to the American Pain Society Guidelines on the Management of Postoperative Pain. 6,8,13 These panelists cite decreased postoperative opioid consumption, less preoperative anxiety, fewer requests for sedatives, and reduced postoperative hospital inpatient admission when clinicians provide patient-and family-centered education to a patient, which includes information on treatment options for the management of postoperative pain (strong recommendation, low evidence). 13 Murphey and colleagues performed a single institution pilot study that investigated prescribing patterns within an institution among both residents and attending physicians. Of 42 survey respondents, they found the most commonly prescribed narcotic was hydrocodone-acetaminophen. On average, residents prescribed 14.1 tablets for septoplasty, whereas attending physicians prescribed 20 tablets. 25 Data on rationale for their prescription practices were not included; however, the authors noted the importance of educating providers on the practices of multimodal analgesia before the implementation, which could infer another reason as to why those who have been practicing with postoperative opioid for years may hesitate to adopt a different and new practice without buy-in and/or provider education. There may be benefit for institutions to establish opioid educational programs for those who wish to implement standardized perioperative and/or postoperative protocols.
As a retrospective study, there were several limitations to include having a limited number of patients that were subjected to the protocol, reflected by the power, specifically when analyzing the refill rates. Our power would be larger if our sample sizes were approximately four or five times larger. However, the overall trend is helpful in order to build a study for our department. Meanwhile, this project lacks postoperative patient survey data such as validated pain scores as well as how much opioid was used. Given the retrospective nature of obtaining data, we were not able to reliably query patients as to how much opioid was leftover several months following surgery. We are currently implementing surveys in our practice in order to obtain these specific data points. Although it is plausible that patients could have obtained opioids from an outside facility, typically those enrolled in Tricare fill their prescriptions at a military facility; however, no external databases were used during this study, and therefore, it is possible that patients who may have obtained opioids from private institutions and/or physicians were not captured. The data extraction process cross-referenced both inpatient and outpatient records in order to capture any postoperative refills that were filled by any provider within the Tricare system. To help eliminate the risk of bias, we used consecutive patients and selected a specific point in time before and after the protocol was initiated. Finally, our population largely consisted of active duty military who may have developed chronic pain conditions or on the other spectrum, who have developed a greater tolerance or threshold to pain, after time in combat. Our study did not stratify patients in accordance with baseline pain levels-further research is warranted in this population to determine the impact of chronic pain, opioid naivety, and susceptibility to opioid tolerance and dependency.

CONCLUSION
The results from implementing a multimodal analgesic protocol suggest that opioid prescription rates may be significantly decreased to manage postoperative pain for septoplasties and rhinoplasties with no significant difference in opioid refill rates. In addition, educating surgeons in regard to a multimodal analgesic approach in conjunction with adopting standardized order sets may reduce opioid prescription rates. All of these efforts combined may potentially decrease the risk of recreational opioid use in the future and related overdose events by helping to eliminate leftover opioid. This QI project warrants further investigation for a future study within our department to determine if further reduction in postoperative opioid prescription dosing can be made while gaining additional insight from patients particularly in regard to their postoperative pain scores using a validated assessment scale and satisfaction outcomes. Expanding this multimodal protocol to other surgeries within our department is also one of our long-term goals.