Implementation of an opioid detoxification management pathway reduces emergency department length of stay

Abstract

Objectives:

With the rise of opioid use in the United States, the increasing demand for treatment for opioid use disorders presents both a challenge and an opportunity to develop new care pathways for ED patients seeking opioid detoxification. We set out to improve the care of patients presenting to our ED seeking opioid detoxification by implementing a standardized management pathway and to measure the effects of this intervention.

Methods:

We conducted a before-after study of the effects of an opioid detoxification management pathway on ED length of stay, use of resources (social worker consultation, laboratory tests obtained), and return visits to the same ED within 30 days of discharge. All data were collected retrospectively by review of the electronic health record.

Results:

Ultimately, 107 patients presented to the ED that met criteria, 52 in the intervention period and 55 in the pre-intervention period. Median ED length of stay in the intervention period was 152 (IQR 93–237) minutes compared to 312 (IQR 187–468) minutes in the pre-intervention period (p<0.001). Patients in the intervention period less frequently had a social work consultation (32.7% vs. 83.6%, p<0.001) or had laboratory tests obtained (32.7% vs 74.5%, p<0.001) and more frequently were prescribed a medication for withdrawal symptoms (57.7% vs. 29.1%, p=0.003).

Conclusions:

Implementation of an opioid detoxification management pathway reduced ED length of stay, reduced utilization of resources, and increased the proportion of patients prescribed medications for symptom relief.

Introduction

With the growth of opioid use disorders in the United States, an increasing number of patients are presenting to emergency departments (EDs) requesting opioid detoxification.¹ Traditional recommendations for ED management of opioid withdrawal have focused on supportive care and referral to outpatient treatment. Unfortunately, the demand for detoxification services has outpaced supply, leaving ED clinicians with limited options for immediate referral and a need to manage patients independently.^2,3 This is complicated by a lack of specific training in opioid detoxification counseling and practice specific guidelines.

The goals of this project were to improve, standardize, and expedite the care of patients presenting to our ED with a chief concern of seeking opioid detoxification by implementing a management pathway and to measure the impact of this intervention.

Methods

Study Design:

We conducted a single-center before-after study to evaluate the effect of an opioid detoxification management pathway on ED length of stay, prescribing, use of resources (social worker consultation, laboratory tests obtained), and return visits to the ED within 30 days of discharge. The primary outcome was ED length of stay (EDLOS). The project protocol was approved by the institutional review board.

Study Setting and Population:

The study was conducted at a tertiary-care academic hospital with approximately 75,000 annual adult ED visits. The study hospital does not offer inpatient treatment for uncomplicated opioid withdrawal. Prior to the intervention, patients requesting opioid detoxification were first evaluated and medically cleared by an emergency clinician. Then they were typically assessed by a mental health care specialist (MHCS), a specially trained social worker, who would provide information about outpatient treatment. There is only one facility within an hour drive from the hospital with grant-funded inpatient detoxification services available for the uninsured. These services are in high demand and are often unavailable.

The study included patients who presented to the adult ED at the study hospital between August 1, 2016 to January 31, 2017 with a chief concern of requesting detoxification from opioids. Patients were excluded if they were: < 18 years old; pregnant; had an acute psychiatric emergency (endorsed suicidality, homicidal ideations, or psychosis); reported concomitant benzodiazepine dependence, alcohol, or barbiturates; or had a coincident medical condition requiring additional care.

Intervention:

Prior to July 2016, MCHS were available in the ED to assist with the evaluation and counseling of patients with substance use disorders. They were situated in an area connected to the main ED, referred to as the Psychiatric Transition Unit (PTU). In July of 2016, the PTU closed and these staff and services were moved to a new acute care area located within the psychiatry hospital on campus, about 0.3 miles away. While affording more beds for patients awaiting definitive placement at a psychiatric facility, this geographical separation of mental health care specialty services from the ED created prolonged ED lengths of stay for patients presenting to the ED for opioid withdrawal, as social workers now had to travel to the ED from the psychiatric hospital. Consequently, patients were often waiting hours to be seen by a social worker, who, given the lack of available beds in the community for opioid use disorder treatment, could often only offer the patient a list of resources for them to pursue independently after discharge. These prolonged lengths of stay, compounded by limited available options for immediate detoxification treatment, became a source of frustration for both patients and providers.

In August of 2016 a multidisciplinary team representing emergency medicine, psychiatry, nursing, social work, and pharmacy convened with the aim to expedite and standardize care for patients presenting requesting opioid detoxification. The team decided to create a standardized management pathway that would empower emergency clinicians to directly assess, manage, and refer patients requesting treatment for opioid dependence, without reliance on social-work.

The intervention included a new medication order set and discharge instructions as well as provider education about using these tools. The medication order set, which was integrated into the EHR, included a clonidine taper, dicyclomine, promethazine, and ibuprofen (Figure 1). These were supplied as outpatient prescriptions and typically not administered in the ED. The new discharge instructions included information about: coping with withdrawal and how to obtain naloxone. The instructions also included information about outpatient resources, which was obtained via semi-structured interviews of regional facilities about: each facility’s offered treatments, patient eligibility, insurance requirements, and out-of-pocket costs.

Figure 1:

Opioid detoxification management pathway including recommended outpatient prescription medications

Provider education was disseminated in lecture and group discussion format during departmental meetings and email-based tutorials of the pathway tools to faculty, residents, and advanced practice providers. In our department, most of these patients are treated by a triage provider and often are not placed in an ED bed. Our triage area is staffed only by faculty physicians. During our education of providers, we explained laboratory tests were unnecessary for medical clearance for outpatient care for this population. The only caveat was a pregnancy test for women of child-bearing age. This recommendation was supported by our interviews of local facilities, as none of them required or felt that lab work done in the ED would be helpful for patients to secure outpatient care. An illustration of the pathway was provided to clinicians in print copy and by email, and was available on the departmental website. The pathway was implemented on October 25, 2016.

Data Collection and Analysis:

To study the effects of the intervention, charts with any ED diagnosis ICD-10 code starting with F11, “opioid related disorders,” were abstracted from the electronic health record (EHR) and reviewed for exclusion criteria, first by ICD-10 code, then manually by one study author. Patients were excluded for benzodiazepine dependence if there was an ICD-10 code diagnosis starting with F13, “Sedative, hypnotic, or anxiolytic related disorders,” or if provider documentation described concern for benzodiazepine dependence. Only patients who were requesting opioid detoxification were included, which was determined by manual chart review as well. Charts were further reviewed for: EDLOS, age, sex, payer status, exclusion criteria, final disposition, whether a social work consultation was obtained, laboratory tests ordered, prescriptions written, and return visits to the same hospital within 30 days.

Patient characteristics and outcomes were compared between a 3-month intervention period after implementation and a 3-month pre-intervention period immediately preceding implementation of the pathway. Continuous data were summarized as medians and interquartile ranges (IQRs) and discrete data with counts and percentages. EDLOS was compared using the Wilcoxon Rank-Sum test. Only patients who were discharged from the ED in both periods were included in the calculation of EDLOS. Categorical data were compared using the Chi Square or Fischer Exact test. Statistical analyses were performed using Stata Statistical Software (StataCorp, College Station, TX). All tests of significance were two-sided. Alpha was set at 0.05.

Results

From August 1, 2016 to January 31, 2017, 548 patients presented to the ED with at least one opioid related diagnosis by IDC-10 Code. Of these, 107 patients met final eligibility criteria, including 55 in the pre-intervention period and 52 in the intervention period. The most common reason for exclusion was concomitant benzodiazepine use (n=80). Patients in the preintervention and intervention periods were similar with respect to age, gender, and payer status (Table 1).

Table 1:

Characteristics and outcomes before and after implementation of management guideline for patients requesting opioid detoxification

Characteristic/Outcome	Pre-intervention (n=55)	Intervention (n=52)	p-value
Age (years)	33 (26–41)	30.5 (26–37.5)	0.19
Female	27 (49.1)	19 (36.5)	0.19
Payer Status
None	19 (34.6)	19 (34.6)	0.12
Medicaid	25 (45.5)	15 (28.9)
Medicare	2 (3.6)	4 (7.7)
Commercial	9 (16.4)	15 (28.9)
Military	0	2 (3.9)
EDLOS* (minutes, n=45, n=50)	312 (187–468)	152 (93–237)	<0.001
Social work consulted	46 (83.6)	17 (32.7)	<0.001
Labs obtained
≥ 1 test ordered	41 (74.5)	17 (32.7)	<0.001
Urine	31 (56.4)	11 (21.2)
Blood	10 (18.2)	6 (11.5)
Disposition
Discharged	45 (81.8)	50 (96.2)	0.019
Admitted	3 (5.5)	0 (0)
Left prior to formal discharge or AMA	7 (12.7)	2 (3.8)
Prescribed ≥ 1 medication for withdrawal symptoms	16 (29.1)	30 (57.7)	0.003
Medications prescribed
Clonidine	12 (21.8)	25 (48.1)	0.005
Anti-emetic	16 (29.1)	28 (53.9)	0.009
Anti-diarrheal	2 (3.6)	18 (34.6)	<0.001
Return ED visit to same hospital within 30 days	1 (1.8)	3 (5.8)	0.35

Continuous data presented as: median (interquartile range). Descriptive data reported as: number (percentile).

^*ED length of stay: only patients who were discharged were included in the analysis of length of stay (pre-intervention n=45, intervention n=50).

In the pre-intervention period 7 patients either left prior to formal discharge or left against medical advice (AMA) and 3 patients were admitted to the psychiatric hospital. These 3 admitted patients initially denied suicidal ideation to the ED physician and met the criteria for the pathway but later endorsed suicidal ideation or a history of possible recent intentional overdose to the social worker. The median EDLOS of these 3 patients was 670 (IQR 614–891) minutes. In the intervention period 2 patients either left prior to formal discharge or left AMA and none were admitted. A total of 45 patients in the preintervention period and 50 in the intervention period were discharged from the ED and therefore included in the final comparison of ED length of stay before and after the intervention.

Median EDLOS in the pre-intervention period was 312 (IQR 187 – 468) minutes as compared to 151 (IQR 93–237) minutes after the intervention (p<0.001) (Figure 2). Patients in the intervention period less frequently had a social work consultation (32.7% vs. 83.6, p<0.001), had any laboratory tests obtained (32.7% vs. 74.5, p<0.001), and more frequently received medication prescriptions for withdrawal symptoms (57.7% vs. 29.1%, p<0.05) (Table 1). More patients were discharged (did not leave prior to formal discharge, leave AMA, or be admitted) in the intervention period (96.2% vs. 81.8%, p=0.019). There was no difference in return ED visits within 30 days between periods (5.8% vs. 1.8%, p=0.354).

Figure 2:

ED length of stay (LOS) before and after implementation of a guideline for the management of patients requesting opioid detoxification. Blue dots represent individual consecutive visits. The red line represents the predicted ED LOS via linear regression by week, in the pre-intervention and intervention period. The 95% confidence interval is shown in dashed lines.

Discussion

Our study of the implementation of a management pathway for patients presenting to the ED requesting detoxification from opioids had three major findings. First, our pathway resulted in a significant reduction in EDLOS. Second, there was a significant increase in the proportion of patients who were discharged from the ED with a prescription to assist them in managing their symptoms (29% vs 58%). Third, we significantly reduced ED resource utilization, including a reduction in social work consults (84% vs 33%) and laboratory testing (75% vs 33%).

While overall the implementation of the pathway was felt to be successful, 33% of patients who qualified for inclusion still had social work consults. This indicates that there was nonadherence to the pathway, consistent with contemporary studies of pathway implementation.^4,5 Using multiple linear regression, a social work consult was associated with a 221.1 (95% CI; 116.4 – 325.8) minute increased EDLOS, indicating that nonadherence attenuated the maximum possible impact of the intervention on EDLOS.

Nonadherence may have been secondary to lack of awareness of the pathway, extenuating patient or departmental circumstances, or clinician preference. Clinicians may feel reluctant to disappoint patients who present with the expectation of admission.⁶ The ability to reference a departmental policy or pathway can be helpful in these situations. Adherence was facilitated by EHR integration such that clinicians could use the discharge instructions as an outline for counseling the patient on symptomatic treatment, return precautions, and outpatient treatment.

Medication-assisted treatment (MAT) with opioid agonists, such as buprenorphine, is widely considered to be the most effective treatment for opioid dependence.⁷ While there is a recent movement towards considering initiation of these medications from the ED,^7,8 EDs face significant barriers in establishing pathways accomplish this. While not as effective as MAT, alpha-2-adrenergic agonists, such as clonidine, have been shown to reduce the severity of withdrawal and increase the likelihood that a patient will complete treatment.⁹ While we continue to work on increasing access to MAT, including through the ED, pathways including clonidine, such as ours, represent an initial step towards improving the care of patients requesting opioid detoxification.

We identified fewer patients for our analysis than we expected, likely because we underestimated how many patients would meet exclusion criteria, especially concomitant benzodiazepine use. The ubiquity of benzodiazepine use, as well as challenges in defining benzodiazepine dependence, was a barrier to successful pathway implementation. Benzodiazepine use is pervasive amongst opioid users, with as many as 40–61% reporting concomitant use.^10,11 Commonly cited reasons for use include enhanced euphoria, anxiolysis, and to treat withdrawal symptoms.^11,12 Benzodiazepine withdrawal, like alcohol withdrawal, is potentially life-threatening, so we originally excluded patients who reported benzodiazepine use from the pathway. However, given the high prevalence of opioid and benzodiazepine co-use, this exclusion may have ultimately limited the impact of our intervention. In realizing this limitation, we took away two important lessons for the future success of caring for patients with this presentation.

First, some patients may want to discontinue opioids but not benzodiazepines; these patients can still be managed with the opioid detoxification pathway as they continue their benzodiazepine use. Therefore, clinicians should determine which substances the patient is requesting detoxification from instead of assuming the patient is seeking total abstinence.

Second, not all benzodiazepine use is equivalent in producing pharmacologic dependence. Patients with a longer duration of regular use and those taking higher doses of benzodiazepines are more likely to experience withdrawal.¹³ In a recent study of nonmedical opioid users, the majority reported benzodiazepine use in the past 30 days but the minority (13%) were every-day users.¹⁰ Developing a clear pathway for assessing a patients’ risk of withdrawal based on use characteristics, would be helpful in determining which patients can be discharged safely. Without this, ED providers will have difficulty triaging patients to appropriate levels of care, which will become more and more problematic as the nation confronts growing numbers of patients requesting opioid detoxification, many of whom co-abuse benzodiazepines.

Limitations

This study has several important limitations. First, we were not able to assess if our intervention had any impact on patient likelihood to complete detoxification or to abstain from opioids. Nor did we assess patient satisfaction with the process. We only accounted for visits to our ED and are unsure if patients presented to other local hospitals after the index visit. Data was gathered via retrospective review of the electronic health record and is therefore subject to errors of documentation or coding. Our pathway was tailored for our local context, which includes a paucity of opioid detoxification resources, particularly for underinsured patients; an alternative intervention may be more appropriate in a different context.

Conclusions

Implementation of a standardized management pathway for patients presenting to the ED requesting detoxification from opioids reduced ED length of stay, reduced utilization of resources, and increased the proportion of patients prescribed medications for symptom relief.