Standardization and Reduction of Narcotics After Pediatric Tonsillectomy

Abstract

Objectives.

(1) To measure caregiver satisfaction with a non-standardized postoperative pain regimen after pediatric tonsillectomy. (2) To implement a quality improvement project (QIP) to reduce the number and volume of narcotics prescribed and to describe the effect on caregiver satisfaction.

Methods.

A prospective cohort study at a tertiary children’s hospital examined postoperative narcotics prescribed to children following adenotonsillectomy. A QIP was implemented 3 months into the observation, with the goal to standardize nonnarcotic analgesics and reduce the volume of narcotics prescribed. Caregivers were called 2 to 3 weeks postoperatively to assess pain control and caregiver satisfaction.

Results.

Over an 8-month period, 118 patients were recruited (66 before the QIP, 52 after induction). Prior to the QIP, 47% of patients were prescribed postoperative narcotics, as opposed to 27% after the QIP (P < .05). There was a significant reduction in the volume of narcotics prescribed before (mean ± SD, 300 ± 150 mL) versus after (180 ± 111 mL) the initiative (P < .05). The per-kilogram dose did not change over the study time frame. On a 5-point Likert scale, there was no difference in the caregivers’ satisfaction regarding pain control before (4.37 ± 0.85) versus after (4.35 ± 1.0) the project started.

Discussion.

A system shift was identified with the establishment of a posttonsillectomy pain control protocol associated with a reduction in prescribed narcotics without a significant change in caregiver satisfaction.

Implications for Practice.

Implementing a standardized plan for the use of nonnarcotic medications was associated with reduced frequency and volume of narcotics prescribed. Future work will further standardize our postoperative pain regimen.

Tonsillectomy, with or without adenoidectomy, is the second-most common pediatric surgery in the United States, with between 250,000 and 500,000 cases performed annually.^1–3 In 2019, the American Academy of Otolaryngology–Head and Neck Surgery released updated clinical practice guidelines regarding tonsillectomy that included a review of the indications, perioperative recommendations, and treatment of postoperative complications.⁴ The new guidelines recommended the use of acetaminophen and ibuprofen for postoperative pain but offered little guidance regarding postoperative narcotics. Aside from recommending against codeine use, the lack of comments about postoperative pain and narcotic control in the pediatric population likely speaks to the controversial nature of this subject.

Pain control following pediatric tonsillectomy and adenoidectomy (pTA) can be very challenging for providers and caregivers. In a survey by Persino, Saleh, and Walner published in 2017, the caregivers for patients who underwent tonsillectomy with or without adenoidectomy did not feel that over the counter medications alone were adequate in controlling postoperative pain.⁵ Due to the opioid epidemic, recent media attention, and increasing public knowledge, caregivers are increasingly more cautious and skeptical regarding narcotics use, including for postoperative pain control.^6,7 In West Virginia, a state that has been significantly affected by the opioid epidemic, caregivers have sought to avoid these medications if and when possible for children.⁶ Alternative measures to improve pain control and limit potential risk have been studied, including acupuncture, honey, and perioperative pain counseling.^8–10 Outside of advocating for increased prescription of over-the-counter medication, few methods have been widely adopted.

In the backdrop of the opioid epidemic during the summer of 2017, we developed a plan to study the current prescribing practices at our institution regarding narcotics and pediatric tonsillectomy and to assess patient satisfaction with the previous regimen. Using this information, we hoped to reduce our narcotic prescriptions, in terms of prescription number and volume, by prescribing a standardized nonnarcotic pain regimen with simultaneous counseling and expectance guidance for the patients’ caregivers. Our hypothesis was that the number of narcotic prescriptions could be reduced posttonsillectomy but that patient and caregiver satisfaction may decrease following this intervention.

Methods

The methods are reported in adherence to the SQUIRE 2.0 guidelines (Standards for Quality Improvement Reporting Excellence).¹¹

Context

Ruby Memorial Hospital is a tertiary care academic children’s hospital that performs between 400 and 600 pediatric tonsillectomies each year. It is 1 of 2 pediatric hospitals in the state of West Virginia, serving patients from within the state as well as southwestern Pennsylvania, western Maryland, and eastern Ohio. Until January 2020, the majority of West Virginia, a rural Appalachian state, did not have the infrastructure for digital or e-prescription of narcotics or other controlled substances either within the state or across state lines. Secondary to the state’s rural nature and limited access to care, many patients travel >2 hours away for routine pTA. Secondary to the inability to electronically prescribe narcotics, many patients were offered narcotics at the time of discharge in addition to acetaminophen and ibuprofen. West Virginia was especially hard hit by the opioid epidemic, with the highest rate of opioid-related fatalities in the country.¹² As such, beginning in 2017, there was a significant push within the institution to limit opioid prescriptions. Furthermore, there was no standardized nonnarcotic regimen and minimal, if any, postoperative teaching regarding nonnarcotic pain management following pTA at our institution. Prior to and throughout the experimental portion of the procedure, hydrocodone/acetaminophen elixir was the primary narcotic used at our institution secondary to its ease of dosing to patient weight and its availability in a liquid form.

Interventions

Interdisciplinary Team.

The ambulatory pharmacy services team approached the Department of Otolaryngology regarding the development of an initiative to reduce the number of opioid prescriptions as well as the total volume of liquid hydrocodone/acetaminophen prescribed following pediatric tonsillectomy. An interdisciplinary committee was assembled, which included 1 pediatric otolaryngologist, 1 general otolaryngologist, the director of ambulatory pharmacy services, and 2 otolaryngology residents.

Intervention Design.

The interdisciplinary committee developed a combined observational cohort–quality improvement project with the intention of reducing narcotic prescriptions without causing undo postoperative suffering among our patients. The project was begun in November 2017 and completed in June 2018. Consecutive pediatric patients (≤ 18 years old) undergoing tonsillectomy with or without adenoidectomy were eligible to enroll. Exclusion criteria included significant craniofacial abnormalities, severe developmental delays, underlying clotting, and/or hematologic disorders that precluded the use of narcotics or nonsteroidal anti-inflammatory drugs (NSAIDs). The first portion of the project was an observational cohort, which occurred from November 2017 until January 2018. All providers at our institution who perform pTA were observed (2 pediatric trained otolaryngologists, 1 sleep boarded otolaryngologist, 5 general otolaryngologists), and their practice habits regarding postoperative pain control were documented. Two to 3 weeks following the procedure, the caregiver for the patient was called on the telephone. After verbal consent, caregivers were enrolled in a phone survey of their satisfaction with their children’s pain control regimen. This survey included the amount of medication consumed, how much was left over, and if a refill was required. The survey can be found in Supplemental Appendix A (available online). Data were recorded in a REDCap data set.¹³ The medical chart was queried for volume and dose of narcotic, if prescribed, as well as complications such as admissions for dehydration or reports of bleeding.

Change of Practice 1: Standardization of Nonnarcotic Medications and Reduction of Opioid Prescriptions.

The second portion of the study was the intervention group, which spanned March to June 2018. Following the observational lead-in, a quality improvement project was instituted where all patients undergoing pTA were given a standardized nonnarcotic pain regimen. This included acetaminophen, 15 mg/kg up to 500 mg every 6 hours, alternating with ibuprofen, 10 mg/kg up to 400 mg every 6 hours. These medications were staggered every 3 hours, and caregivers were encouraged to awaken patients to provide medication. The operating surgeon could continue to prescribe narcotics if deemed necessary, but all providers were recommended to reduce narcotic use at faculty and departmental meetings.

Change of Practice 2: Postoperative Pain Control Education.

An instructional handout was created and provided to caregivers prior to discharge. The handout contained a section to track a child’s dosing, and it reinforced staying ahead of the patient’s pain, including awakening to provide these medications (Supplemental Appendix B).

All but 1 (7 of 8) of the surgeons who perform tonsillectomies participated in the intervention portion of the study. Like the first arm, our care team telephoned caregivers 2 to 3 weeks after surgery for administration of our phone survey following verbal consent (Supplemental Appendix A). The medical chart was queried for volume and dose of narcotic, if prescribed, as well as complications such as admissions for dehydration or reports of bleeding.

Measures

The process measures following the intervention were the frequency of narcotic prescriptions, the volume of narcotic prescribed (mL), and the dose of narcotic prescribed (hydrocodone; mg/kg). Balancing measures included caregiver satisfaction and complications following tonsillectomy, specifically hospitalizations for dehydration or reports of bleeding episodes. The outcome measure was caregiver-reported narcotic remaining.

Analysis

An a priori power analysis was not performed prior to the study. A post hoc power analysis showed that with an α value of 0.05, the β values for narcotic volume, prescription, and dose were 95%, 60%, and 65%, respectively. All statistical analysis was preformed with R studio (v 3.3.1; R Foundation). The results were analyzed with a pooled Student’s t test and Mann-Whitney U test to compare the means of the 2 populations where appropriate. P values were not calculated for postoperative tonsillectomy hemorrhage or hospitalization for dehydration. A P value <.05 was deemed to be statistically significant.

Ethical Considerations

This project, including the observational lead-in and the quality improvement project, was approved following expedited review by the West Virginia University Institutional Review Board (protocol 1702446187).

Results

During the 3-month observational lead into the study, 110 patients met inclusion criteria. Of these, 66 were willing to participate in the phone survey (60%). Patient demographics are shown in Table 1. Of this observational cohort, 47% of the patients received a narcotic prescription (hydrocodone/acetaminophen elixir, 7.5 mg / 325 mg per 15 mL), with a mean dose of 300 mL and approximately 3 mg/kg hydrocodone. (Table 2). During this initial observational cohort, 72% of the patients were prescribed acetaminophen, and 66% were prescribed ibuprofen. At the time of the phone survey, none of the patients who were prescribed narcotics endorsed having unused medication. During this observational cohort, 1 patient presented to the emergency department for dehydration, and 3 reported episodes of bleeding. Caregiver satisfaction regarding postoperative pain control was 4.4 out of a 5-point Likert scale.

Table 1.

Demographic Data.

	No. or mean ± SD
	Lead-in	Intervention	P value
Patients	66	52	—
Sex			.17
Male	31	31
Female	35	21
Age, y	7.2 ± 4.7	5.8 ± 3.3	.079
Body mass index	19.7 ± 6.0	19.2 ± 6.6	.69

Table 2.

Narcotic Results.

	% (No.) or mean ± SD
	Lead-in	Intervention	P value
Receiving postoperative narcotics	47 (32)	27(14)	.019
Caregiver satisfaction, 5-point Likert scale	4.4 ± 1.0	4.4 ± 0.85	.95
Dose, mg/kg	3.0 ± 1.6	2.28 ± 1.2	.190
Hydrocodone/acetaminophen (7.5/325), mL	300 ± 150	180 ± 111	.015
>50% of narcotic remaining	0	43	—
Dehydration events	1.5(1)	1.9(1)	—
Bleeding events	4.5 (3)	5.7 (3)	—

The quality improvement project was implemented in February 2018, and a 1-month run-in was observed to allow for implementation of the interventions. Over the ensuing 3 months, 95 patients qualified for the study with 52 (54.7%) agreeing to participate. Following the implementation of the quality improvement project, 27% of patients were prescribed narcotics, 87% acetaminophen, and 85% ibuprofen. At the time of the phone survey, 43% of the patients who were prescribed narcotics endorsed having >50% of the medication remaining. The mean volume of narcotic prescribed following the intervention was 180 mL, with a mean dose of 2.28-mg/kg hydrocodone. Similar complications were observed to the lead-in, with 1 episode of dehydration and 3 episodes of reported bleeding. Caregiver satisfaction regarding postoperative pain control was 4.4 out of a 5-point Likert scale.

Process, Balancing, and Outcome Measures

The process measures for this quality improvement project were percentage of patients prescribed narcotics, the volume of narcotics prescribed, and the dose of narcotic prescribed. This project arose following a discussion with the ambulatory pharmacy services regarding the volume of narcotic prescribed following pTA. As such, the primary measure of the study is the volume of narcotic prescribed, with the frequency and dose being secondary measures. Prior to beginning this project, we used a hydrocodone/acetaminophen elixir as the patient’s postoperative narcotic of choice, in part because of the ability to easily titrate this medication to patient weight. As such, we did not expect much, if any, reductions in the weight-based dosing of these medications.

While we desired to reduce the frequency, volume, and dose of narcotic, we first wanted to make sure that patients were not undergoing any undo discomfort. As such, our first balancing measure was caregiver satisfaction with the patient’s pain control. In the observational portion of the study, we sought to assess our current practice regarding nonnarcotic pain management and associated caregiver satisfaction. Prior to the intervention, 72% of the patients were prescribed acetaminophen and 66% ibuprofen as compared with 87% and 85% following the intervention, respectively. What is lost in the raw numbers is that following the intervention, acetaminophen and ibuprofen were prescribed on a scheduled basis as compared with an as-needed basis prior to the intervention. The previously mentioned educational handout was created to show how to schedule the nonnarcotic medication, and it was provided to the patient upon discharge (Supplemental Appendix B).

The second balancing measure was complications. Since we were attempting to decrease the frequency, amount, and volume of narcotics prescribed and increase the dosing and prescription of ibuprofen and acetaminophen, we were interested in assessing dehydration secondary to poor pain control and the possible increased risk of bleeding from the ibuprofen. In each arm of the study, there was 1 emergency department visit for dehydration and 3 episodes of reported posttonsillectomy hemorrhage. All 6 episodes of bleeding were self-resolved, and there were no repeat trips to the operating theater for any of the patients included in the study. The study is not powered to assess for rare events, such as posttonsillectomy hemorrhage or dehydration, and a larger study may find small differences that are missed in this current assessment.

The outcome measure of the study was the volume of remaining narcotic at the conclusion of the patient’s postoperative course as reported by the caregiver. If the nonnarcotic regimen and caregiver education interventions were successful, then even the patients who do receive the lower volume of narcotics should have some remaining at the conclusion of the their convalescence. What we found was that of those in the observational cohort who were prescribed narcotics, none of the caregivers reported using <50% of the narcotic. In the interventional cohort, 43% (6 of 14 patients) endorsed using less than half the prescribed narcotic.

Summary of Findings

The rate of narcotic prescription and volume of narcotic prescription were reduced between the observational and interventional cohorts (rate, P = .019; volume, P = .015). The dose of narcotic as measured in mg/kg of hydrocodone did not change between the observation and intervention. Despite the decrease in narcotic frequency and volume, caregiver satisfaction did not change between the cohorts (P = .95), and there were no obvious changes in complications as measured by dehydration and bleeding events. Even though the volume of narcotic was reduced in the interventional cohort, 43% of the interventional cohort used <50% of the prescribed narcotic, while 0% of the patients in the observation group used <50% of the narcotic prescribed.

Discussion

Interpretation of Results

From the results of our study, standardization of nonnarcotic prescriptions and providing instructions postoperatively were associated with a reduction in the number and volume of prescription narcotics without a significant change in caregiver satisfaction. Despite the intervention, patients in the treatment arm had more remaining narcotic at the conclusion of the study as compared with the lead-in group. This finding may partially explain why there was no reduction in satisfaction scores following the intervention. This finding is not unique, however, as previous retrospective surveys found that patients use significantly less narcotic medication than that prescribed following tonsillectomy.¹⁴ While previous studies showed that patient caregivers prefer to have a narcotic script for backup pain control, recent literature has shown that opinions on this topic may be changing.^5–7 To this end, a recent database study discovered that patients who received narcotics had a slightly increased risk of complication but a lower overall readmission rate.¹⁵

Pediatric pain management following tonsillectomy is variable from patient to patient and highly dependent on provider preferences. Historically, codeine represented one of the most common medications for posttonsillectomy pain, but this treatment was abandoned following a black box warning by the Food and Drug Administration in 2013.¹⁶ Since then, consensus on the ideal pain regimen for pediatric tonsillectomy has yet to emerge, leaving considerable variability in postoperative routines. Based on a 2018 survey of otolaryngologists by Schwartz et al, the most commonly prescribed pain medications in children were NSAIDs (67%) and acetaminophen (65%).¹⁷ In that same population, 37% of respondents prescribed hydrocodone-acetaminophen, and 9% still prescribed acetaminophen with codeine.¹⁷ When pediatric providers were asked how frequently they prescribed narcotics, there was considerable variability among the many pediatric surgical subspecialities.^17,18

To our knowledge, this is the first study to examine caregiver satisfaction following a program to reduce narcotic utilization among patients undergoing tonsillectomy. Luk et al looked at a similar initiative based on a standardized order set for acetaminophen and ibuprofen with or without narcotic prescription.¹⁹ In their investigation, Luk and colleagues reduced narcotic prescriptions from 82% to 15% of patients following implementation. The authors used urgent care and emergency care visits as a proxy for patient satisfaction, without a significant increase in utilization following the initiation of the study.¹⁹ Sowder et al evaluated the combined use of ibuprofen and narcotics with caregiver satisfaction.²⁰ In their investigation, they looked at patients who were prescribed ibuprofen alone versus ibuprofen with hydrocodone/acetaminophen, with the study markers being postoperative pain and recovery. They found that the patients in the ibuprofen-only group recovered more quickly, but it is worth noting that these patients were not randomized and were significantly younger than those receiving ibuprofen with narcotic.²⁰

Nonnarcotic medications are not without risk. The most common alternative nonnarcotic medications are ibuprofen and acetaminophen, with ibuprofen being shown to be as effective as narcotics for postoperative pain control in certain settings.²¹ In addition to its anti-inflammatory properties, COX inhibitors such as ibuprofen decrease platelet function and potentially worsen the frequency and severity of posttonsillectomy hemorrhage.²² The literature has shown mixed results for postoperative bleeding rates in patients taking ibuprofen. While early studies were encouraging, recent studies and systematic reviews found an increased risk for severe bleeding among patients who use ibuprofen after tonsillectomy.^23–26 In light of this new research, the risks and benefits of NSAIDs or narcotics use for postoperative pain control should be reviewed with the patient’s caregiver prior to surgery.

Limitations

Limitations of this study include a relatively small sample size (118 patients) as well as the subjective nature of the outcome measurement. The primary outcome measured in our study was pain control as assessed by caregivers using a 5-point Likert scale. Pain is difficult to measure among patients, given the subjective nature of pain and wide variations in pain tolerance. This effect is augmented in pediatric patients, who often cannot be asked directly about their pain. The interpretation of pediatric pain can also be skewed by caretaker biases and perceptions. Second, while this study prospectively collected data, we still asked the patients’ caregivers to assess their experience with the patients’ pain control over the preceding weeks. Due to this design, there is some risk of recall bias among the caregivers. Future areas of research include better understanding and quantifying the appropriate dose of narcotics for each patient and what, if any, pre- or perioperative factors can be used to determine who will require narcotic pain medications postoperatively.

Implications for Practice

Pediatric tonsillectomy is the second-most common pediatric surgery in the United States. Postoperative pain control can be challenging after a tonsillectomy, and the use of narcotics in pediatrics is controversial. In this investigation, we first sought to better understand our practices at a tertiary care children’s hospital. Once we understood our practice, we sought to reduce the volume and frequency of narcotic prescriptions in our practice. We were able to do this through increased caregiver education and a standardized nonnarcotic regiment of ibuprofen and acetaminophen. This reduction of narcotics was balanced against caregiver satisfaction regarding pain control and patient complications. Through this initiative, we were able to reduce the frequency and volume of narcotics prescribed without negatively affecting caregiver satisfaction or grossly worsening the complication rate. The current study is not powered to assess complication rate, and larger studies are needed to elucidate the risks and benefits of narcotics and ibuprofen for pain control following tonsillectomy. Going forward, we hope to periodically reassess our prescribing patterns to better understand our use of narcotics and how to best care for our pediatric patients undergoing tonsillectomy.