Clinician's Capsule
What is known about the topic?
The emergency department is a key contact and intervention site for persons at risk of opioid death.
What did this study ask?
This quality improvement study evaluated the impacts of a multi-site program to initiate opioid use disorder treatment in emergency departments.
What did this study find?
Forty-seven discharged patients were given buprenorphine/naloxone and 35 continued to fill prescriptions for opioid agonist treatment 30 and 60 days after.
Why does this study matter to clinicians?
A multi-site approach to opioid use disorder in emergency departments can be effective and support standardized evidence-based care across sites.
INTRODUCTION
In 2018, when our intervention began, 746 Albertans died from an apparent accidental opioid overdose.1 The rate of emergency department (ED) visits related to opioid use and substance misuse increased by 65% from January 1, 2015 to September 30, 2018.1 Nine percent of individuals who died from an apparent accidental opioid poisoning in Alberta had a substance use-related ED visit in the 30 days before death.1
In the summer of 2017, leaders within Alberta began working with clinician experts on how best to serve ED patients who live with opioid use disorder. Project coordinators established an expert working group, including administrators, pharmacists, addiction medicine specialists, and emergency medicine physicians. Based on a rapid review of research evidence, the group set goals of increasing prescription of buprenorphine/naloxone (sublingual) in ED sites and establishing rapid referral of opioid use disorder patients to continuing treatment.
The Canadian Research Initiative in Substance Misuse strongly recommends initiating opioid agonist treatment with buprenorphine/naloxone “whenever feasible” as the preferred first-line treatment for opioid use disorder.Reference Bruneau, Ahamad and Goyer2,Reference Wood, Ahamad and Fairbairn3 Opioid agonist treatment “is significantly more effective than nonpharmacological treatments in retaining individuals in treatment and suppressing illicit opioid use”.Reference Wood, Ahamad and Fairbairn3 Opioid agonist treatment is also effective in reducing morbidity and mortality,Reference Wood, Ahamad and Fairbairn3 and can be initiated in the ED.Reference D'Onofrio, O'Connor and Pantalon4–Reference Hu, Snider-Adler, Nijmeh and Pyle6 A randomized control trial conducted in the United States showed that individuals who are initiated on buprenorphine/naloxone in the ED and referred to treatment had better outcomes than those who receive brief intervention and referral or facilitated referral.Reference D'Onofrio, O'Connor and Pantalon4,Reference D'Onofrio, Chawarski and O'Connor5 Specifically, patients who initiated buprenorphine/naloxone in the ED had increased engagement in addiction recovery programs and reduced self-reported illicit opioid use at 30 days since ED initiation.Reference D'Onofrio, O'Connor and Pantalon4 A Canadian observational study published after our project began shows 88% of eligible patients were prepared to start buprenorphine/naloxone treatment in the ED, and 54% attended clinic treatment.Reference Hu, Snider-Adler, Nijmeh and Pyle6
This article reports initial results of a multi-site effort to improve buprenorphine/naloxone initiation rates and develop processes for referring patients to community care.
METHODS
The project is quality improvement and formal exemption from ethics review was received from the University of Alberta Health Research Ethics Board. The project team worked with the Alberta Health Services Privacy Office to protect patient and clinician privacy from risks arising from data collection. Administrative data were used for measurement, as these data are available across sites in the province.
The program theoryReference Funnell and Rogers7 informing our intervention was that ED sites would benefit from common processes and tools that could be implemented locally, and that patients must be referred to and attend follow-up care in order for ED intervention to make a difference over time for patients. Our evaluation objectives were consequently to determine uptake of the initiation and referral pathway at the site level and to examine process, outcome, and balancing measures related to buprenorphine/naloxone initiation at ED visit and patient levels. Our initial evaluation, as reported here, was intended to establish whether the program could be spread to ED sites across the province.
Study design and time period
Evaluation used retrospective observational methods during May to September 2018.
Study setting
Our province has a single health authority with regional geographic subdivisions for operational decision making. Some electronic health records are captured within a provincial information system. ED sites rely on a variety of record keeping systems. Independent community addiction clinics and health authority operated clinics serve patients who live with opioid use disorder. The Emergency Strategic Clinical NetworkTM brings together stakeholders across the province to create and promote standardized best practices in emergency care.Reference McLane, Holroyd and Lang8
The intervention was developed with and piloted in three urban ED sites in the province. Each ED referred patients to a single opioid use disorder treatment clinic. Clinic and ED selection was based on including both health authority and independent sites, availability of local clinicians who would work on and advocate for the intervention (expert working group members), and on reports of prior referrals from selected ED sites to paired clinics.
Each paired set (ED and clinic) started the intervention at a different time. The North East Community Health Centre began the program on May 15, 2018, the Grey Nuns Community Hospital began on June 11, 2018, and the Rockyview General Hospital began on July 5, 2018. The sites are located in different cities with distinct clinical and operational emergency medicine leadership teams, and have unique staff mixes, medical record systems, and pharmacy systems. Rockyview General Hospital had the third highest number of opioid visits within the province from January 1, 2015, to September 30, 2018; Grey Nuns Community Hospital had the ninth highest; and North East Community Health Centre was not among the top ten sites for opioid-related ED visits provincially.1
Population
Patient eligibility criteria were developed by the expert working group. To be eligible, patients had to meet the following three criteria: “Age 18 years old and older, Suspicion of opioid use disorder,” and “Patient willing to engage in buprenorphine/naloxone.” Exclusion criteria were any of: “Allergy to buprenorphine or naloxone, Being admitted for medical/psychiatric concern, Severe liver dysfunction, Currently prescribed or using methadone or buprenorphine/naloxone,” or “Clinical signs of sedative/depressant impairment or intoxication.” Pregnant patients were included, and consultation with an addiction medicine specialist was recommended.9 Patients not prescribed buprenorphine/naloxone could still be referred to addiction clinics.
For evaluation purposes, discharged visits where patients received buprenorphine/naloxone are compared with visits for opioid-related reasons where patients did not receive buprenorphine/naloxone. See online Appendix 1 for diagnosis codes counted as opioid related. Only discharged patients are included because the ED referral protocol is directed toward these patients. Patients with other dispositions, such as admissions, transfers, and left without treatment, are excluded because there is no development of an outpatient follow-up plan for them within the ED. See online Appendix 2 for disposition codes counted as discharged and admitted.
Intervention
We considered the Consolidated Framework for Implementation Research as a conceptual model for change in developing our intervention.Reference Damschroder, Aron, Keith, Kirsh, Alexander and Lowery10 This model considers a range of factors that impact the success of an intervention, including the credibility of its origin, evidence for the intervention, perceived need for change, relative advantage of the intervention over alternatives, perceived difficulty for practitioners, and individuals’ capacity and willingness to take up the intervention. To address these and other aspects of the model, local implementation teams were formed in each ED site. These teams included a physician champion, a nursing lead, an administrative lead (e.g., ED manager), and representatives from pharmacy services and social work. Teams worked with experts to create the initial treatment protocol, order set, discharge instructions, education materials, fax referral forms and referral pathways based on available evidence. A full-time project manager and the interdisciplinary Emergency Strategic Clinical NetworkTM (ESCN) leadership team provided expertise and infrastructure to develop materials, processes, and operational relationships. All processes were intended to be evidence based, promote standard practices, maximize detection of patients with opioid use disorder, minimize use of ED resources, and reduce barriers to clinic attendance for patients.
Local teams led education for physicians and nurses on the benefits of opioid use disorder treatment, buprenorphine/naloxone medication initiation, referral protocols, and supporting documents. As each site is different, local teams also worked to tailor the program to local resources (e.g., social work hours), and information systems, and to coordinate with local services outside the ED (e.g., accounting for clinic and pharmacy opening hours, and ensuring referral options or bridging practices during off hours). Figure 1 shows a high-level overview of our intervention protocol.
Figure 1. High level overview of protocol.
Outcome measures
The monthly rate of buprenorphine/naloxone initiation per ED was assessed from January to April 2018 (baseline) and May to September 2018 (intervention period). Rates were compared with eight nonintervention sites in Calgary and Edmonton. Monthly initiation rate was our main outcome at the site level.
One process and two outcome measures were assessed at the patient level. The process measure was the number of referrals to community clinics from intervention ED sites. The outcome measures were attendance at first appointment at community clinic and continuity in care.
We used ED electronic medical records to track patients given buprenorphine/naloxone in the ED. Participating community clinics manually tracked the number of referrals received by means of fax from participating ED sites and the number of referred patients who attended their first follow-up appointment, as no existing data system captures such information across sites.
The Pharmaceutical Information Network11 dataset was used to determine the number of patients filling any opioid agonist treatment prescription 30, 60, and 90 days after being discharged from ED. This dataset includes information on prescriptions filled at pharmacies in Alberta. Prescription filling was understood as a proxy measure for continuity in care. See online Appendix 3 for the medications and doses counted as opioid agonist treatment prescriptions.
Patient and emergency visit characteristics are also reported. Diagnosis, patient age, patient sex, Canadian Triage Acuity Score (CTAS),Reference Bullard, Musgrave and Warren12 disposition (e.g. admitted or discharged), and length of stay were extracted from the National Ambulatory Care Reporting System.13 Diagnoses were classified using International Classification of Disease-10 codes.14
Two balancing measures were developed to assess ED resource use and adverse patient events related to buprenorphine/naloxone. Patient length of stay was recorded as a proxy measure for ED resource use caused by buprenorphine/naloxone initiation in the ED. Numbers and descriptions of adverse events involving buprenorphine/naloxone were derived from the provincial Reporting and Learning System.15
Data analysis
Descriptive statistics are presented for all measures. A run chart, i.e., a line graph showing data over time, was used to compare the number of ED visits where buprenorphine/naloxone was given in intervention sites before and after the intervention.Reference Rocco, Lloyd and Sandy16 Comparison to eight other sites was conducted to establish whether changes observed were due to the intervention or to wider system changes at a time when there is an increasing focus on opioids. One Edmonton hospital is excluded from the analysis as an outlier, as it has a very successful and high-profile pre-existing hospital-wide addiction medicine program.
When reporting prescription filling for patients who were given buprenorphine/naloxone in the ED, the first ED visit where buprenorphine/naloxone is given was used. For patient-level analysis of patients not given buprenorphine/naloxone in the ED, their first ED visit for an opioid-related reason was used. A nonparametric Wilcoxon rank sum test comparing means was used to compare lengths of stay for visits where patients are given buprenorphine/naloxone in the ED and not given buprenorphine/naloxone in the ED. Sample size for ED visits is 427.
RESULTS
Table 1 gives characteristics of facilities participating in the intervention. Figure 2 shows the number of buprenorphine/naloxone starts in intervention and control sites by month from January 2018 to September 2018.
Figure 2. Buprenorphine/naloxone average monthly initiations (2018).
Table 1. Facility characteristics
Before the intervention, from January to April 2018, ED sites participating in the intervention gave buprenorphine/naloxone tablets during four unique ED visits (an average of 0.25 visits per month per site). From program launch dates until the end of September 2018, these sites gave buprenorphine/naloxone tablets during 66 visits (an average of 5.9 visits per month per site, after intervention launch dates). From May to September 2018, the eight other sites in Edmonton and Calgary gave buprenorphine/naloxone during 84 unique visits (an average of 2.1 visits per month per site).
The demographics of patients receiving buprenorphine/naloxone in the ED and their prescription filling rates at 30, 60, and 90 days, are given in Table 2. Thirty-five (74.4%) patients discharged from the ED with buprenorphine/naloxone continued to fill opioid agonist treatment prescriptions 30 days and 60 days after their index ED visit. Thirty-four (72.3%) filled prescriptions at 90 days.
Table 2. Patient characteristics
The participating ED sites referred 37 patients to the three participating clinics, and 16 of those individuals (43%) attended their first community follow-up appointments.
Characteristics of ED visits during which buprenorphine/naloxone was given are presented in Table 3. Median length of stay for ED visits resulting in discharge with buprenorphine/naloxone was not significantly longer than for opioid-related visits where buprenorphine/naloxone was not given (p = 0.35).
Table 3. Visit characteristics (all visits)
One patient safety event involving buprenorphine/naloxone was reported, which was a dosing error with no apparent harm to the patient. This event occurred because clinicians provided a dose based on the naloxone component of buprenorphine/naloxone tablets. Wording in the protocol was subsequently clarified, to specify that buprenorphine/naloxone is a combination drug and quantity given should be based on the buprenorphine component.
DISCUSSION
Interpretation of findings
As the intervention site increased monthly buprenorphine/naloxone prescribing rates more than comparison sites, the increase in buprenorphine/naloxone dispensing at participating sites appears to be the result of the project. Nearly three-quarters of the patients started on the medication in the ED continued to fill prescriptions for opioid agonist treatment at 30, 60, and 90 days. We regard the fact that 43% of referrals attended their first follow-up appointment as evidence of success of the referral pathway. Some patients who received the medication were not referred to participating clinics, and these patients may have preferred referral to primary care or another clinic.
Clinical implications
Results suggest that a multi-site effort can initiate patients on opioid agonist treatment, and effectively connect patients to treatment. In this project, ED sites worked together while avoiding duplication of efforts, and patients received standardized care across sites.
The median difference in length of stay for visits leading to discharge with buprenorphine/naloxone and opioid-related visits leading to discharge without buprenorphine/naloxone was not statistically significant. This suggests that dispensing buprenorphine/naloxone does not increase ED length of stay for discharged patients. The expert working group regarded the fact that only one minor patient safety event was reported positively.
Comparison to previous studies
We are not aware of other studies reporting buprenorphine/naloxone monthly prescribing rates at the ED level across sites. Our findings on continuity of care may be compared with findings reported by D'Onofrio et al. and Hu et al.Reference D'Onofrio, Chawarski and O'Connor5,Reference Hu, Snider-Adler, Nijmeh and Pyle6 D'Onofrio et al. report 74% of patients given buprenorphine/naloxone and referred to treatment were engaged in treatment 60 days after randomization.Reference D'Onofrio, Chawarski and O'Connor5 However, D'Onofrio et al. excluded patients with active psychiatric conditions. Hu et al. reports no such exclusion and found 35% of patients (15/43) initiated on buprenorphine/naloxone and referred to treatment were continuing on buprenorphine/naloxone 6 months later.Reference Hu, Snider-Adler, Nijmeh and Pyle6
Our 43% referral attendance rate was similar to the 54% rate reported by Hu et al.Reference Hu, Snider-Adler, Nijmeh and Pyle6 We believe reducing the time and effort required to make referrals through standardized fax referral forms encouraged uptake of the program in ED sites. We believe that next day walk-in options made transitions in care straightforward for patients.
Strengths and limitations
A particular strength of our work was that it examined a multi-site approach to initiating opioid use disorder treatment across ED sites. A limitation of our evaluation was an inability to determine the number of eligible patients who could have received buprenorphine/naloxone in the ED during our intervention period. A further limit of our analysis was that Reporting and Learning System use is voluntary. Therefore, some patient safety events may not have been captured in the data relied on.
Research implications
While this was a quality improvement project, results suggest a need for research to determine the numbers and characteristics of patients who would benefit from buprenorphine/naloxone and referral to opioid use disorder treatment in Alberta ED sites. Further quality improvement efforts could then focus on ensuring that these patients receive buprenorphine/naloxone.
CONCLUSION
Emergency clinicians can effectively initiate patients on buprenorphine/naloxone when supports for this standardized evidence-based care and pathways for follow-up in community are available.
Supplemental material
The supplemental material for this article can be found at https://doi.org/10.1017/cem.2020.438.
Acknowledgments
The authors thank Denise Boudreau for information design for project materials, Dr. Katherine Rittenbach with the Additions and Mental Health Strategic Clinical NetworkTM, for providing scientific feedback on our evaluation plan, and the members of the Emergency Department Opioid Response Expert Working Group for providing input on all aspects of program design, implementation, and evaluation.
Financial support
This research was funded by Alberta Health through the Minister's Opioid Emergency Response Commission (MOERC). The funder did not direct the reporting of the intervention or the contents of this manuscript.
Competing interests
Dr. Deol reports grant funding from AHS and Health Canada to provide treatment guidelines and education on opioid use disorder to emergency physicians, during the conduct of the study; he is part owner and part of the leadership team of Metro City Medical Clinic Calgary. Dr. Fanaeian reports grant funding from AHS and Health Canada to provide treatment guidelines and education on opioid use disorder to emergency physicians, during the conduct of the study; and clinical practice in treating patients with opioid use disorder. Dr. Ghosh reports he is currently an addiction physician at a publicly funded clinic supporting those with substance use disorders. Dr. Holroyd reports that he is the Senior Medical Director of the AHS Emergency Strategic Clinical Network.
