Poor opioid stewardship contributes to opioid misuse and adverse health outcomes. We sought to decrease opioid prescriptions in children 0 to 18 years treated for pain after fractures and cutaneous abscess drainage from 13.5% to 8%. Our secondary aims were to reduce opioid prescriptions written for >3 days from 41% to 10%, eliminate codeine prescriptions, increase safe opioid storage and disposal discharge instructions from 0% to 70%, and enroll all emergency department (ED) physicians in the state prescription drug monitoring program.
We implemented an intervention bundle on the basis of 4 key drivers at a pediatric ED: ED-wide education, changes in the electronic medical record, discharge resources, and process standardization. Two plan-do-study-act cycles were performed. Interventions included provider feedback on prescribing, safe opioid storage and disposal instructions, and streamlined electronic medical record functions. Run charts were used to analyze the effect of interventions on outcomes. Our balance measure was return ED or clinic visits for inadequate analgesia within 3 days.
During the intervention period, 249 of 3402 (7.3%) patients with fractures and cutaneous abscesses were prescribed opioids. The percentage of opioid prescriptions >3 days decreased from 41% to 13.2% (P < .0001), codeine prescription dropped from 1.1% to 0% (P = .09), opioid discharge instructions increased 0% to 100% (P < .0001), and all physicians enrolled in the prescription drug monitoring program. There was no change in return visits for uncontrolled analgesia compared with the baseline (P = .79).
A comprehensive opioid stewardship program can improve opioid prescribing practices of ED physicians and deliver information on safe storage and disposal of prescription opioids with a negligible effect on return visits for uncontrolled pain.
Opioid misuse is a major cause of morbidity and mortality in the United States. One-third of US opioid overdose fatalities are due to prescription opioid overdoses.1 In 2018, there were 826 prescription opioid overdose deaths in persons <25 years of age.1 One-fifth of adolescents report using a prescription opioid in the past year and 3.8% of adolescents engage in opioid misuse or have an opioid use disorder.2 In younger children, exposure to opioids prescribed for other patients can lead to adverse health effects and even death.3
A key contributor to the nonmedical use of prescription opioids is access. Among adolescents and young adults who misuse opioids, 56% obtained them from friends or relatives and 25% obtained them from the health care system.2 Parents may inadvertently allow access to prescription opioids through improper storage and disposal of unused opioids.4–6 Caregivers are often unaware of safe disposal options or the rationale behind disposal of leftover opioids.5
Physician prescribing habits also contribute to the availability of prescription opioids. Clinicians may prescribe greater amounts of opioids than is necessary for the treatment of acute pain in children7 because of a lack of evidence-based guidelines regarding the appropriate type and quantity of opioids to be prescribed for medical conditions.7,8 This can lead to marked variability in opioid prescribing.9–11 In a pediatric acute-care setting, the odds of receiving an opioid prescription for >5 days were higher among infants and among patients whose prescriptions were written by a resident physician.12 In addition, prescriber use of the prescription drug monitoring program (PMP) is low13 and its use in pediatric emergency medicine is unknown. Mandatory querying of the PMP for a patient’s previous use of opioids is associated with significant lowering of prescription drug misuse in young adults.14 Opioid use in adolescence is associated with nonmedical use of prescription opioids in young adulthood.15 Therefore, it is important to limit exposure of adolescents to prescription opioids.
Injuries and infections are the primary reasons why children seek emergency care and pain is a common symptom with these conditions.16 The overarching goal of acute pain management is to administer the safest analgesic at an appropriate dose and duration within the context of the natural course of the illness. There must be a balance between achieving satisfactory analgesia and prescribing an appropriate quantity of opioids to reduce the risk of diversion (transfer of one’s prescribed controlled substance to another for illicit use).
In adult emergency departments (EDs), the comparison of individual and group prescribing rates, as well as education on pain management and opioid prescribing guidelines, significantly reduced the number of opioid prescriptions and prescribed doses.17–19 However, opioid stewardship has not been studied in the pediatric ED. With this study, our aim was to render appropriate analgesia to children post-ED discharge while reducing the risk of opioid diversion by standardizing opioid prescribing and educating caregivers about safe storage and disposal of prescription opioids.
Methods
We performed this quality improvement project in the pediatric EDs of 1 urban and 2 suburban, tertiary-care children’s hospitals within the same health system (cumulative annual ED census: 140 000). The EDs serve the greater Houston area, with a population of 3 million. In our EDs, opioids are only prescribed by attending physicians. The study was approved by the Baylor College of Medicine Institutional Review Board.
We formed a multidisciplinary quality improvement team of ED physicians, nurses, pharmacists, and information technology professionals. In our discussions, we identified several barriers to opioid stewardship in our EDs, including (1) a lack of evidence-based guidelines on the appropriate type and quantity of prescribed opioids for the treatment of acute pain in children,7,8 (2) the harmful safety profile of codeine in pediatrics,20 (3) nonenrollment of ED physicians in the state PMP, and (4) the absence of after-visit summary (AVS) instructions on safe storage and disposal of prescription opioids.
Current literature describes the equianalgesic efficacy and better side-effect profile of ibuprofen compared with opioids in children with extremity fractures or after minor surgery.21,22 In addition, opioids are prescribed for a median of 2 to 3 days in pediatric acute-care settings,12,23 and parents often discontinue opioids and use over-the-counter analgesics by the third postoperative day in children who undergo minor surgery.24 Therefore, we addressed the above problems on the basis of the literature25–27 and key drivers (Fig 1) by educating physicians as follows: (1) use safer nonopioid analgesic alternatives, (2) avoid prescribing codeine, (3) identify youth at risk for nonmedical use of prescription opioids by enrolling in and querying the PMP before prescribing an opioid, (4) limit prescribing opioids, (5) avoid prescribing opioids for >3 days when opioids are absolutely necessary, and (6) provide safe storage and disposal instructions after prescribing opioids.
Opioids are most commonly prescribed for pediatric patients with extremity fractures and cutaneous abscesses.12,23,28,29 Thus, we standardized opioid prescribing for the treatment of these conditions at our EDs. In May 2018, we initiated the opioid stewardship program for the treatment of children aged 0 to 18 years who sustained extremity fractures or who underwent drainage of a cutaneous abscess. By reducing the duration of prescribed opioids and by educating families on safe storage and disposal of opioids, our goal was to administer effective analgesia while reducing the risk of opioid diversion. However, because a reduction in prescribed opioid doses could lead to incomplete analgesia, we evaluated all return ED or clinic (primary care or orthopedic) visits within 3 days for uncontrolled pain as a balance measure.25
Baseline Measurement
Baseline data for 2017 extracted from the electronic medical records (EMRs) consisted of age, sex, diagnosis, procedure performed, opioid discharge prescription (written or e-prescribed), prescription duration (days), and total volume (milliliters) and/or number of doses prescribed. Opioids included morphine, hydrocodone, oxycodone, and codeine as a single ingredient or in combination with acetaminophen. We excluded complex, open, and femur fractures because they require hospital admission and/or operative care and excluded conditions for which opioids are rarely prescribed, such as fractures involving the clavicle, upper humerus, and foot Supplemental Table 3). Baseline data revealed that of the 1965 patients with fractures (1645) or abscesses (320), opioids were prescribed in 266 (13.5%) encounters and for >3 days in 110 (41.4%) of those encounters, codeine constituted 1.1% of prescribed opioids (3 of 266), no ED physician was enrolled in the PMP, and no discharge instructions on safe storage and disposal of prescription opioids were available. Our goals, which were prespecified before project initiation, were to achieve the following quality measures within 24 months of project commencement: primary process measure and balance measure.
Primary Process Measure
For children treated for pain after sustaining extremity fractures and after drainage of cutaneous abscesses, we would reduce the percentage of opioid prescriptions from 13.5% to 8% of encounters (40% reduction). Secondary process measures were to (1) limit prescribed opioid doses of >3 days’ supply to up to 10% of encounters, (2) eliminate all codeine prescriptions, (3) enroll 100% of ED physicians in the PMP, and (4) include discharge instructions on the safe storage and disposal of prescribed opioids in 70% of the AVS of patients who were prescribed opioids.
Balance Measure
To study the effect of uncontrolled analgesia, we measured the number of return ED or clinic (primary care or orthopedic) visits for inadequate analgesia within 3 days of ED discharge.
On the basis of earlier data, for a 40% reduction in the percentage of opioid prescriptions from a baseline of 13.5% to 8%, a sample of 740 patients would be needed to achieve 80% power in a 2-sided t test with an α of .01. We constructed run charts using Minitab 19 (State College, PA) to identify a nonrandom pattern for primary and secondary process measures. We used the following rules to identify a nonrandom pattern for the run charts30 : (1) shift: ≥6 consecutive points either all above or all below the median; (2) trend: ≥5 consecutive points all going up or all going down; and (3) runs: above or below the standardized limits for the number of allowable runs for a set number of observations.
Preintervention
Before project implementation, we developed a document in English and Spanish on safe storage and disposal of prescription opioids with a link to Drug Enforcement Administration–approved disposal locations for inclusion in the AVS.31,32 This document was uploaded as a discharge resource in the EMR for distribution to parents whose children were prescribed opioids. We also standardized the process for prescribing opioids (document on choice, indications, and duration of opioids) and issuing safe opioid storage and disposal discharge instructions. Finally, we obtained approval to launch a voluntary American Board of Pediatrics Maintenance of Certification (MOC) quality project on opioid stewardship for interested ED physicians.
After 3 months of project design and feedback, we commenced the opioid stewardship project for ED physicians in May 2018. The project goals, interventions, and metrics were shared with ED physicians, nurses, and nursing leadership at each site during section meetings and retreats and through in-person meetings and flyers. Invitations were extended to all 86 ED physicians to participate in the MOC project, and 12 enrolled in August 2018. We maintained a rolling enrollment to enable others to enroll at any time. Participating physicians attended a 1-hour lecture on opioid stewardship delivered by one of the authors (R.P.S.) and another session on project aims and metrics. The lecture discussed the epidemiology of nonmedical use of prescription opioids, neurobiology of opioids, evidence-based opioid prescribing practices,12,23,26,28 and safe storage and disposal of prescription opioids.31,32
Interventions
The interventions consisted of physician education and feedback, enrollment in the PMP, rollout of revised discharge instructions, and changes in EMR functionality. The first plan-do-study-act cycle for the opioid stewardship project began in May 2018. Physicians were educated on safe prescribing, storage, and disposal of opioids at commencement and every quarter through e-mail. We analyzed the EMR of all eligible patients for primary, secondary, and balance measures every 2 months. Aggregate metrics were shared every quarter with all ED physicians via e-mail irrespective of their enrollment in the MOC project. Despite sharing monthly metrics, we observed that we were not meeting project goals, particularly with regard to opioid treatment duration and discharge instructions. Therefore, in October 2019, we began the second plan-do-study-act cycle. In this cycle, we began providing bimonthly confidential and nonpunitive feedback via e-mail to physicians who failed to adhere to project goals. The feedback included the specific patient encounters in which they prescribed codeine, prescribed opioids for >3 days, and/or did not provide opioid discharge instructions.
In August 2018, we began reaching out to all ED physicians to ensure that they enrolled in the PMP. This continued until all physicians were enrolled. In the second year of the project (May 2019), we began providing information on the opioid stewardship project to newly hired ED physicians and assisting them with PMP enrollment during the onboarding process. In March 2020, Texas made it mandatory for all physicians who prescribed controlled substances to enroll in the PMP and query their patient’s past use of controlled substances.
In December 2019, we revised the AVS instructions on the basis of input and feedback from project stakeholders. The new document had the same information as the previous iteration but was more concise and included illustrations. The finalized document was translated into Spanish, Vietnamese, and Arabic (most common non-English languages spoken in our hospital) and uploaded in the EMR.
The final intervention consisted of improving EMR functionality to assist physicians in reviewing the PMP before prescribing opioids and in providing opioid discharge instructions. In January 2019, the hospital began an opioid stewardship program to examine prescribing patterns by service line, methods of disposal of unwanted opioids,33 and the possibility of including a PMP query tool within the EMR. On February 25, 2020, the hospital information services (IS) implemented a best-practice alert (BPA) in the EMR for use by all hospital physicians who prescribed opioids. This prompt directed physicians to the PMP site to verify their patient’s previous controlled substance prescription history and attest to doing so. This BPA became functional before the state-mandated deadline of March 1, 2020.
We received feedback from MOC participants that the need to manually click on the opioid discharge document in the EMR for inclusion in the AVS made it difficult to consistently include these instructions after prescribing an opioid. Thus, on March 22, 2020, the IS established a new EMR functionality that permitted automatic printing of the document in the AVS when an opioid was prescribed.
Results
There were 1965 encounters between January 1 and December 31, 2017 (baseline), and 3402 encounters between May 1, 2018 and April 30, 2020 (intervention), by children aged 0 to 18 years for treatment of extremity fractures and cutaneous abscesses. Demographics of the 2 cohorts were similar (Table 1).
Demographics of Patients at Baseline and During Intervention Period
| Variable | Time Period | |||
|---|---|---|---|---|
| Baseline Eligible Conditions Jan 2017–Dec 2017 (n = 1965) | Intervention Eligible Conditions May 2018–Apr 2020 (n = 3402) | Baseline Eligible Conditions Prescribed Opioids Jan 2017–Dec 2017 (n = 264) | Intervention Eligible Conditions Prescribed Opioids May 2018–Apr 2020 (n = 249) | |
| Male sex, n (%) | 1247 (63.4) | 2187 (64.2) | 196 (74) | 165 (66) |
| Median age in y (IQR) | 6.7 (4.6–11.7) | 10.3 (4.8–11.7) | 9.3 (6.4–12.3) | 7.5 (6.7–13.2) |
| Diagnosis, n (%) | ||||
| Extremity fracture | 1645 (83.7) | 3198 (94.0) | 252 (95.5) | 246 (98.7) |
| I&D skin abscess | 320 (16.3) | 204 (6.0) | 12 (4.5) | 3 (0.1) |
| Variable | Time Period | |||
|---|---|---|---|---|
| Baseline Eligible Conditions Jan 2017–Dec 2017 (n = 1965) | Intervention Eligible Conditions May 2018–Apr 2020 (n = 3402) | Baseline Eligible Conditions Prescribed Opioids Jan 2017–Dec 2017 (n = 264) | Intervention Eligible Conditions Prescribed Opioids May 2018–Apr 2020 (n = 249) | |
| Male sex, n (%) | 1247 (63.4) | 2187 (64.2) | 196 (74) | 165 (66) |
| Median age in y (IQR) | 6.7 (4.6–11.7) | 10.3 (4.8–11.7) | 9.3 (6.4–12.3) | 7.5 (6.7–13.2) |
| Diagnosis, n (%) | ||||
| Extremity fracture | 1645 (83.7) | 3198 (94.0) | 252 (95.5) | 246 (98.7) |
| I&D skin abscess | 320 (16.3) | 204 (6.0) | 12 (4.5) | 3 (0.1) |
I&D, incision and drainage.
Pre- and postintervention outcomes are summarized in Table 2. Figures 2–5 refer to the intervention period only. For the primary outcome, the percentage of opioid prescriptions decreased significantly from 13.5% to 7.3% (P < .0001) with a median of 6.7% (interquartile range [IQR]: 5.2–8.7). The run chart (Fig 2) reveals a decrease in the rate of opioid prescriptions with a nonrandom shift early in the project (education of ED staff on metrics) and no nonrandom pattern of trends or number of runs.
Baseline and Intervention Outcome Measures
| Outcome | Baseline January | Intervention May 2018 to April 2020 | |||
|---|---|---|---|---|---|
| to December 2017 n of N (%) | Goal % | Median (IQR), % | Postintervention, n of N (%) | P | |
| Primary process measure | |||||
| Patients prescribed opioids at discharge | 266 of 1965 (13.5) | 8 | 6.7 (5.2–8.7) | 249 of 3402 (7.3) | <.0001 |
| Secondary process measures | |||||
| Opioid prescriptions exceeding a 3-d supply | 110 of 266 (41) | 10 | 9.7 (0–23.3) | 33 of 249 (13.2) | <.0001 |
| Codeine prescriptions | 3 of 266 (1.1) | 0 | 0 | 0 of 249 (0) | .09 |
| Enrollment of ED physicians in the PMP | 0 | 100 | — | 100 | <.0001 |
| Discharge instructions on safe storage and disposal of opioids in AVS | 0 | 70 | 44.4 (22.9–53.4) | 100 | <.0001 |
| Balance measure | |||||
| Return ED or clinic visit for inadequate pain control within 3 d | 5 of 266 (1.9) | — | 4 of 249 (1.6) | .79 | |
| Outcome | Baseline January | Intervention May 2018 to April 2020 | |||
|---|---|---|---|---|---|
| to December 2017 n of N (%) | Goal % | Median (IQR), % | Postintervention, n of N (%) | P | |
| Primary process measure | |||||
| Patients prescribed opioids at discharge | 266 of 1965 (13.5) | 8 | 6.7 (5.2–8.7) | 249 of 3402 (7.3) | <.0001 |
| Secondary process measures | |||||
| Opioid prescriptions exceeding a 3-d supply | 110 of 266 (41) | 10 | 9.7 (0–23.3) | 33 of 249 (13.2) | <.0001 |
| Codeine prescriptions | 3 of 266 (1.1) | 0 | 0 | 0 of 249 (0) | .09 |
| Enrollment of ED physicians in the PMP | 0 | 100 | — | 100 | <.0001 |
| Discharge instructions on safe storage and disposal of opioids in AVS | 0 | 70 | 44.4 (22.9–53.4) | 100 | <.0001 |
| Balance measure | |||||
| Return ED or clinic visit for inadequate pain control within 3 d | 5 of 266 (1.9) | — | 4 of 249 (1.6) | .79 | |
—, not applicable.
Opioids prescribed for fractures and abscesses during the intervention period. PDSA, plan-do-study-act.
Opioids prescribed for fractures and abscesses during the intervention period. PDSA, plan-do-study-act.
Opioids prescribed for >3 days’ supply for fractures and abscesses during the intervention period. PDSA, plan-do-study-act.
Opioids prescribed for >3 days’ supply for fractures and abscesses during the intervention period. PDSA, plan-do-study-act.
After-visit instructions on safe storage and disposal of opioids during the intervention period. PDSA, plan-do-study-act.
After-visit instructions on safe storage and disposal of opioids during the intervention period. PDSA, plan-do-study-act.
For the secondary outcomes, the percentage of opioid prescriptions exceeding a 3-day supply (Fig 3) decreased from 41% to 13.2% (P < .0001), with a median of 9.7% (IQR: 0–23.3). There was a nonrandom shift and 2 downward trends (education and confidential feedback of physicians) but no nonrandom number of runs. Codeine prescriptions as a percentage of all opioid prescriptions dropped from 1.1% to 0% (P = .09). Figure 4 reveals the increase in enrollment of ED physicians in the PMP from 0% to 100%. The percentage of AVS with discharge instructions (Fig 5) increased from 0% to 100% (P < .0001), with a median of 44.4% (IQR: 22.9–53.4). There was 1 nonrandom shift (confidential feedback to physicians and changes in EMR functionality).
There was no significant difference in the number of return ED or clinic visits for uncontrolled analgesia at baseline and during the intervention (1.9% vs 1.6%; Table 2). In Supplemental Table 4, we describe the patients who sought care for uncontrolled analgesia at baseline and during the project period. Return visits involved tight splints or opioid adherence in patients with lower extremity fractures. No cases of compartment syndrome occurred.
Discussion
Successful opioid stewardship requires a comprehensive approach and process standardization pertaining to opioid prescribing, storage, and disposal.9,27 It is an iterative process necessitating the cooperation of multiple stakeholders through all stages of the project. Through multidisciplinary engagement, we were able to identify potential barriers and develop methods to address them. This is the first example of opioid stewardship in pediatric emergency medicine.
At the conclusion of the project, there were significant reductions in overall opioid prescribing and in opioid prescriptions exceeding a 3-day supply. Furthermore, codeine prescribing was eliminated. These changes in opioid prescribing patterns were not associated with a concomitant increase in return visits for uncontrolled analgesia. Our results in reducing the number of opioid prescriptions (45%) and prescribed doses (67%) are comparable with opioid stewardship programs in adult EDs (28%–56% and 15%–20%, respectively)17–19 and could potentially reduce opioid prescribing at pediatric centers with similar baseline opioid prescribing rates (20%–42%)12,23 A longer duration of prescribed doses may be necessary in some instances, such as in patients with complicated fractures. Prescribing a higher number of opioid doses should be individualized and physicians must query the PMP for possible prescription duplication.
Another aim was to enroll all ED physicians in the PMP. When we commenced this project, there was no regulatory requirement for physicians to enroll in the PMP. Although we were able to reach our goal of enrolling all ED physicians in the PMP, we were unable to verify that physicians were actually querying the system to determine their patient’s past opioid prescription history. Fortuitously, in March 2020, Texas passed a law mandating that all physicians enroll in and query the state PMP and attest that they reviewed their patient’s previous controlled substance prescription history before prescribing an opioid. This, along with the introduction of a BPA and changes in the EMR, enabled us to ensure that all physicians queried the PMP every time.
We exceeded our project goal of including instructions on opioid safe storage and disposal in at least 70% of encounters in which opioids were prescribed. Initially, physicians were required to manually select inclusion of this information in the AVS. Given the busy ED, the low compliance was unsurprising. We therefore reached out to another hospital that uses the same EMR as our institution and has implemented an automated process to print the opioid discharge information in their AVS when an opioid is prescribed. Our hospital IS was able to implement similar changes in our EMR with excellent results.
Some quality interventions, such as the BPA, state-mandated requirement to query the PMP, and EMR changes to include opioid discharge instructions, were successful and required minimal to no effort. Interventions that were related to process standardization, namely, prescribing opioids or limiting opioid doses, were more challenging. Success was possible through physician engagement and translating evidence into bedside practice.34
In terms of sustainability and dissemination of the quality project, we are educating new hires about the project, encouraging participation in the MOC course, and disseminating our results internally and with colleagues statewide. State regulations and EMR changes will ensure long-lasting effects without active oversight. Future considerations are to use the car, relax, alone, forget, family and friends, and trouble questionnaire35 to risk stratify youth for substance use whenever an opioid is prescribed.
The study has several limitations. First, we did not contact parents after discharge to verify whether they had filled their opioid prescription, whether their child had adequate analgesia, or whether the family had used the safe storage and disposal information in the AVS. The fill rate for outpatient opioids in children who have had surgery is 60%.36 Second, hospital policy and access to specialty care allow us to reduce most displaced fractures in the ED and arrange follow-up appointments with orthopedics within 1 to 3 days of the ED visit. This enables patients to receive definitive care early but may not be possible elsewhere. Third, we did not verify whether patients were prescribed opioids at their follow-up visit or whether they sought care at other facilities for uncontrolled pain. Querying the PMP for repeat opioid prescriptions can ascertain this fact. Fourth, we studied the top 2 conditions for which opioids are prescribed in the ED, but we did not study sickle cell disease or cancer. Opioids are prescribed to patients with the latter conditions by the hematology-oncology service for longer durations. Fifth, data reviewers for the balancing measure were not blinded to the study objectives. Finally, the project was performed at 3 children’s hospitals and will need to be replicated at other sites to assess generalizability.
Conclusions
A comprehensive opioid stewardship program in the pediatric ED can improve emergency medicine physician opioid prescribing practices and deliver information on safe storage and disposal of prescription opioids to caregivers with a negligible effect on return visits for uncontrolled pain. This has the potential to reduce adolescent prescription opioid misuse.
Acknowledgments
We appreciate Mr Kuo-Rei C (Cory) Mao for his help in obtaining the data and all the physicians and nursing staff who participated in the quality project.
Ms Lavingia participated in the design of the 2 plan-do-study-act cycles, created after-visit summary discharge instructions and educational materials for emergency department staff, conducted the analysis, and drafted the initial manuscript; Dr Mondragon participated in the design of the 2 plan-do-study-act cycles and created after-visit summary discharge instructions and educational materials for emergency department staff; Dr McFarlane-Johansson conceptualized and designed the study, participated in the design of the 2 plan-do-study-act cycles, and collected the data; Dr Shenoi conceptualized and designed the study, led the design of and oversaw the 2 plan-do-study-act cycles, conducted the analysis, and drafted the initial manuscript; and all authors reviewed and revised the manuscript, approved the final manuscript as submitted, and agree to be accountable for all aspects of the work.
FUNDING: No external funding.
Comments