The role of recovery peer navigators in retention in outpatient buprenorphine treatment: a retrospective cohort study

Abstract

Background

Racial and ethnic disparities are evident in the accessibility of treatment for opioid use disorder (OUD). Even when medications for OUD (MOUD) are accessible, racially and ethnically minoritized groups have higher attrition rates from treatment. Existing literature has primarily identified the specific racial and ethnic groups affected by these disparities, but has not thoroughly examined interventions to address this gap. Recovery peer navigators (RPNs) have been shown to improve access and overall retention on MOUD.

Patients and Methods

In this retrospective cohort study, we evaluate the role of RPNs on patient retention in clinical care at an outpatient program in a racially and ethnically diverse urban community. Charts were reviewed of new patients seen from January 1, 2019 through December 31, 2019. Sociodemographic and clinical visit data, including which providers and services were utilized, were collected, and the primary outcome of interest was continuous retention in care. Bivariate analysis was done to test for statistically significant associations between variables by racial/ethnic group and continuous retention in care using Student’s t-test or Pearson’s chi-square test. Variables with p value ≤0.10 were included in a multivariable regression model.

Results

A total of 131 new patients were included in the study. RPNs improved continuous retention in all-group analysis (27.6% pre-RPN compared to 80.2% post-RPN). Improvements in continuous retention were observed in all racial/ethnic subgroups but were statistically significant in the non-Hispanic Black (NHB) group (p < 0.001). Among NHB, increases in continuous retention were observed post-RPN in patients with male sex (p < 0.001), public health insurance (p < 0.001), additional substance use (p < 0.001), medical comorbidities (p < 0.001), psychiatric comorbidities (p = 0.001), and unstable housing (p = 0.005). Multivariate logistic regression demonstrated that patients who lacked insurance had lower odds of continuous retention compared to patients with public insurance (aOR = 0.17, 95% CI 0.039-0.70, p = 0.015)

Conclusions

RPNs can improve clinical retention for patients with OUD, particularly for individuals experiencing several sociodemographic and clinical factors that are typically correlated with discontinuation of care.

KEY MESSAGES

Recovery peer navigators improve continuous clinical retention following initiation of outpatient treatment for opioid use disorder.

Recovery peer navigators may be especially beneficial for patients with factors and identifiers commonly associated with discontinuation of care.

Introduction

Though the use of buprenorphine has clearly reduced opioid use disorder-related morbidity and mortality, [1, 2] the availability of buprenorphine in low-income, racially/ethnically minoritized communities has failed to grow at the same rates seen in high-income, predominantly white neighborhoods [3]. Rather, expansion of buprenorphine has selectively occurred in private physicians’ offices serving privately insured patients. As a result, patients receiving buprenorphine are largely white, college-educated individuals seeking care for prescription opioid dependence [4]. Effective methods for treatment with buprenorphine thus may systematically favor the populations where access to this medication has expanded the most, limiting the translatability of these methods when working with more disenfranchized communities of color. Therefore, interventions with demonstrated effectiveness in racially/ethnically minoritized patients receiving buprenorphine are urgently needed.

Racial and ethnic minority populations not only have limited access to treatment with buprenorphine, but are also vulnerable to prematurely discontinuing treatment [1, 2, 5]. Though there is considerable variability in retention on buprenorphine and overall rates are discouragingly low, racial/ethnic minorities are 31% more likely to discontinue buprenorphine compared to their white counterparts [6, 7]. However, racial/ethnic minority groups are not within themselves homogenous. Other sociodemographic factors intersect with structural barriers to increase the risk of poor treatment outcomes [4, 8]. Consequently, studying differences within these minority populations is important for understanding who experiences disparities to retention in opioid use disorder (OUD) treatment, and which factors contribute most to these differences in outcomes.

Negative perceptions of opioid use, stigma, and paucity of culturally responsive and respectful care present additional challenges to recovery treatment in these populations [5]. The provision of recovery peer support may help to mitigate these barriers to retention in care. Recovery peer support is expert emotional or social support that is provided by persons with lived experience of a substance use disorder (SUD) to others sharing similar conditions to bring about their self-motivated change [9]. There are local, state, and national certifications for this role that require training in core competency domains. Peer recovery specialists are trained to provide services across a range of domains that are recovery-oriented, patient-centered, voluntary, relationship-focused, and trauma-informed [10, 11]. Recent literature reviews on recovery peer support demonstrate significant improvements in measures of health status, such as decreases in substance use, hospitalization rates, psychotic and depressive symptoms; social determinants of health, such as involvement in the legal system, homelessness, internalized stigma; and adherence with mental health and medical outpatient care [12]. For those with OUD, peer navigation has shown to improve outpatient maintenance on medications for opioid use disorder (MOUD) and reduced substance use. [13–15]. The effectiveness of peer navigators is thought to be due to the help they provide with navigating structural barriers and social determinants of health that contribute to treatment discontinuation [16]. Additionally, their lived experiences may aid in creating a welcoming treatment environment and reduce stigma about MOUD [17]. In the context of treatment for SUDs, peer recovery support can be an important addition to multidisciplinary clinical care models. Given the numerous challenges that individuals with OUD face while in treatment, the benefits that peer recovery support offers are promising and can prove especially beneficial in retaining individuals with compounding risk factors for poorer outcomes.

The aim of this paper is to analyze the impact of incorporating recovery peer support into clinical practice on retention in care in a low-barrier outpatient clinic serving a majority minority population. Additionally, the study aims to examine differences within racial/ethnic groups to better identify intersectional factors that may require the development of tailored interventions to support retention in care for each group. We hypothesize that there will be an overall increase in retention and that these benefits will extend to individuals who demonstrate increased risk of treatment discontinuation based on socioeconomic vulnerability.

Patients and methods

Study design

This study was designed as a retrospective chart review of new patients with at least 1 visit seen between January 1, 2019 and December 31, 2019. Follow-up data were collected through March 15, 2020. On July 1, 2019, recovery peer navigators were introduced into the clinical workflow. Those patients seen prior to implementation of the recovery peer navigators (RPN) were considered in the pre-RPN period while those seen after were considered in the post-RPN period.

Setting and participants

Essex County, in New Jersey, has the highest number of heroin and fentanyl-related deaths in New Jersey and Newark is the municipality with the most residents admitted for substance abuse in Essex County [18, 19]. Newark is an urban city. Approximately 89% of Newark residents are racial/ethnic minorities and 29% of residents live in poverty [20]. The Comprehensive Addiction Resources and Education (CARE) Center is an outpatient substance use disorder treatment program at Rutgers New Jersey Medical School, an academic center in Newark that primarily focuses on providing care to individuals with an opioid use disorder. The clinical team is comprised of medical providers, social workers, RPNs and a clinical pharmacist, who work together to assess and support the patients’ treatment plan. The clinic follows a low-barrier harm reduction model that prioritizes same day prescribing of buprenorphine, walk-in availability, no treatment termination for urine drug screens that are positive for other substances, and MOUD prescribing was not contingent on receiving psychotherapy. The frequency of patient visits was individualized but typically patients were seen at least monthly with patients that required stabilization due to ongoing hazardous opioid use receive received more frequent visits. Buprenorphine refills occurred at >90% of patient visits.

Data collected

Medical records of patients seen at the CARE Center were manually abstracted by 4 trainees (medical students and graduate students) and data were stored in REDCap, a secure, web-based electronic data capture application [21, 22]. Sociodemographic data were abstracted which included age, ethnicity, race, and gender, insurance, other substance use, and presence of psychiatric and medical comorbidities. Dates of initial and subsequent visits, visit type, whether the visit was scheduled or walk-in, and receipt of medication refills was also collected. Visit type indicated which providers were seen along with which services were utilized.

Statistical analysis

Demographic and clinical data were analyzed using descriptive statistics. Continuous data were described using mean and standard deviation and categorical data were described using frequency of counts. The primary outcome of interest was continuous retention in care, which was characterized based on consistency of clinic visits. Patients with any gap in care of greater than 30 consecutive days were considered not continuously retained. Bivariate analyses using Student’s t-test (Mann Whitney U for non-parametric data) or Pearson’s chi-square test were done to test for statistically significantly differences between groups. Variables identified in these analyses with p value ≤ 0.10 were included in a regression model to assess for statistically significant associations with continuous retention in care. Results were reported as odds ratios with 95% confidence intervals (95% CI). Statistical analyses were performed using Stata (version 17, StataCorp LLC, College Station, Texas, USA).

Ethics approval

This study was approved by the Rutgers Biomedical and Health Sciences Institutional Review Board. (IRB ID: Pro2020001928)

Results

131 new patients were included in the study. The average ages were 54.5 for non-Hispanic Black (NHB), 37.4 for non-Hispanic White (NHW), 49.7 for Hispanic and 45.4 for other race/ethnicity patients. Public insurance was utilized by 45 (50%) NHB, 8 (53.3%) NHW, 13 (72.2%) Hispanic and 5 (62.5%) other race/ethnicity patients. Most patients identified as male: 65%, 60%, 83%, and 100%, for NHB, NHW, Hispanic, and other race/ethnicity, respectively. Nearly all patients reported additional substance use and intravenous route of opioid use was reported in 15 (16.7%) NHB, 9 (60%) NHW, 9 (50%) Hispanic, and 1 (12.5%) other race/ethnicity patients. Psychiatric comorbidities were reported in 47 (52.2%) NHB patients, 13 (86.7%) NHW, 13 (72.2%) Hispanic patients, and 5 (62.5%) other race/ethnicity patients. Medical comorbidities were present in 63 (70%) NHB, 7 (46.7%) NHW, 9 (50%) Hispanic, and 3 (37.5%) other race/ethnicity patients as well. Housing instability was reported in 18 (20%) NHB, 3 (20%) NHW and 3 (16.7%) Hispanic patients. See Table 1 for additional sociodemographic information.

Table 1.

Demographic characteristics of patients by race/ethnicity and retention status.

	Non-Hispanic Black		Non-Hispanic White		Hispanic		Other/unknown
	Retained (n = 62)	Not retained (n = 28)	Retained (n = 9)	Not retained (n = 6)	Retained (n = 14)	Not retained (n = 4)	Retained (n = 5)	Not retained (n = 3)
Age, mean (SD)	54.2 (9.4)	55.2 (7.9)	38.4 (9.5)	38.8 (6.3)	49 (12.1)	52 (5)	51.8 (13.3)	34.7 (4)
Sex, n (%)
Male	39 (66.1)	20 (33.9)	5 (55.6)	4 (44.4)	11 (73.3)	4 (26.7)	5 (62.5)	3 (37.5)
Female	23 (74.2)	8 (25.8)	4 (66.7)	2 (33.3)	3 (100)	0	0	0
Health insurance, n(%)
Medicaid	42 (67.7)	28 (32.3)	7 (63.6)	4 (36.4)	13 (76.5)	4 (25.5)	3 (75)	1 (25)
Medicare	2 (50)	2 (50)	1 (100)	0	0	0	0	0
Dual	1 (100)	0	0	0	0	0	1 (100)	0
Uninsured	12 (85.7)	2 (14.3)	0	0	0	0	1 (33.3)	2 (66.7)
Commercial	5 (62.5)	3 (37.5)	1 (50)	1 (50)	1 (100)	0	0	0
Missing	0	1 (100)	1 (100)	0	0	0	0	0
Opioid IDU, n (%)	9 (60)	6 (40)	6 (66.7)	3 (3.3)	6 (66.7)	3 (33.3)	1 (100)	0
Unstable housing	12 (66.7)	6 (33.3)	2 (66.7)	1 (33.3)	3 (100)	0	0	0
Additional substance use, n (%)a	61 (69.3)	27 (30.7)	8 (57.1)	6 (42.9)	14 (77.8)	4 (22.2)	5 (62.5)	3 (37.5)
Psychiatric comorbidityb	35 (74.5)	12 (25.5)	7 (58.3)	5 (41.7)	10 (76.9)	3 (23.1)	4 (80)	1 (20)
Medical comorbidityc	44 (69.8)	19 (30.2)	4 (57.1)	3 (42.9)	7 (77.78)	2 (22.2)	2 (66.7)	1 (33.3)

^a alcohol, cannabis, tobacco, cocaine, methamphetamines, phencyclidine, illicit benzodiazepines.

^b schizophrenia, bipolar disorder, depression, anxiety, post-traumatic stress disorder, personality disorders.

^c hypertension, heart disease, chronic obstructive pulmonary disease, asthma, diabetes mellitus, HIV, chronic hepatitis C, other medical conditions.

29 new patients were seen during the pre-RPN period and 102 during the post-RPN period. Eight of the 29 patients (27.6%) in the pre-RPN period were continuously retained in care, while in the post-RPN period 82 of the 102 patients (80.2%) were continuously retained in care (p < 0.001). Similarly, when the analysis was restricted to patients with 2 or more visits, 7 of the 26 patients (26.9%) in the pre-RPN period versus 38 of 59 patients (64.4%) in the post-RPN period were continuously retained in care (p = 0.001). Overall retention by racial/ethnic group for the entire duration of the study period was found to be: 62 of 90 (68.9%) NHB patients, 9 of 15 (60.0%) NHW patients, 14 of 18 (77.8%) Hispanic patients, and 5 of the 8 (62.5%) patients who identified as other race/ethnicity were retained in care. The proportion continuously retained was increased for all racial/ethnic groups during the post-RPN: by 73.1% for NHB patients, 32.1% for NHW patients, 53.3% for Hispanic patients, and 6.7% other race/ethnicity patients and was a statistically significant change for NHB patients (p < 0.001) (Table 2). Statistically significant increases in continuously retained patients were seen among male sex, female sex, individuals with public and commercial insurance, reported intravenous drug use, housing instability, additional substance use, psychiatric comorbidity, and medical comorbidity. Uninsured patients were retained more frequently post-RPN, but this improvement was not statistically significant.

Table 2.

Retention in care by race/ethnicity and presence of recovery peer navigation support services.

	Pre-Recovery Peer Navigation			Post-Recovery Peer Navigation
	Retained	Not Retained	Percent retained	Retained	Not retained	Percent retained	p-value
Race/ethnicity
Non-Hispanic Black	1	13	7.1	61	15	80.3	<0.001
Non-Hispanic White	3	4	42.9	6	2	75	0.32
Hispanic	1	2	33.3	13	2	86.7	0.11
Other	3	2	60	2	1	66.7	1
Sex
Male	6	16	27.3	54	15	78.3	<0.001
Female	2	5	28.6	28	5	84.9	0.002
Insurance
Public	7	15	31.8	63	16	79.8	<0.001
Commercial	0	3	0	7	1	87.5	0.024
Uninsured/missing	1	3	25	12	3	80	0.071
IDU
No	4	14	22.2	64	15	81.01	<0.001
Yes	4	7	36.4	18	5	78.3	0.026
Unstable housing
No	8	17	32	65	17	79.3	<0.001
Yes	0	4	0	17	3	85	0.003
Additional substance use
No	1	1	50	1	0	100	1
Yes	7	20	25.9	81	20	80.2	<0.001
Psychiatric comorbidity
No	2	11	15.4	32	9	78.1	<0.001
Yes	6	10	37.5	50	11	82	<0.001
Medical comorbidity
No	7	10	41.2	26	6	81.3	0.004
Yes	1	11	8.3	56	14	80	<0.001

42 of the 102 patients (41.2%) in the post-RPN period had at least 1 visit with an RPN. There was no statistically significant differences between contact with an RPN by race/ethnicity, sex at birth, insurance status, housing status, intravenous drug use, psychiatric comorbidity, or medical comorbidity.

Given the large sample size of the NHB subgroup, further analyses were performed to evaluate relationships between the sociodemographic characteristics of patients with continuous retention between the two time periods. In bivariate analysis, there were differences in proportion of patients with continuous retention in the post-RPN when compared by sex, health insurance, history of injection drug use, additional substance use, current unstable housing, and presence of comorbid psychiatric and medical conditions. Sex, receipt of public health insurance, additional substance use, medical comorbidities and current unstable housing were statistically significant (Figure 1). A sensitivity analysis found that using a more liberal definition for continuous retention, no more than 45 days between appointments, did not affect these findings and there was no change (data not shown). All sociodemographic factors, including patient gender, insurance status, housing insecurity, injection drug use, concurrent substance use, and medical and psychiatric comorbidity, were included in a multivariate logistic regression. NHB patients without insurance had a statistically significant lower odds of continuous retention as compared to patients with public insurance (aOR = 0.17, 95% CI 0.039-0.70, p = 0.015, Online Appendix 1).

Figure 1.

NHB patients continuously retained before and after RPNs by sociodemographic variables.

Discussion

This retrospective cohort analysis evaluated the impact of RPNs on continuous retention in care among racial/ethnic groups in a low-barrier clinic. RPNs improved continuous retention in care overall and across all racial/ethnic groups. When examining within-group differences among NHB individuals, the availability of RPNs significantly increased retention in care for both sexes, individuals who rely on public health insurance, patients who reported additional substance use, those with medical comorbidities, and those experiencing unstable housing. In multivariable analysis, NHB patients without health insurance remained most at-risk for not being retained in care.

While our findings demonstrate that RPNs increase retention in care, the setting in which patients are first engaged may impact the effect of the RPNs. RPNs have shown benefit in linkage and retention of individuals with MOUD from the emergency room and carceral setting following MOUD treatment initiation [13, 15]. Similarly, referrals made to RPNs through the primary care setting improved engagement with outpatient buprenorphine and decreased acute care utilization [23]. However, another study did not demonstrate a significant change in outpatient clinic retention based on initial RPN engagement during the inpatient setting [24]. Further investigation is needed to determine the most effective forms of support provided by RPNs.

Given that racial/ethnic minoritized groups continue to disproportionately experience overdoses, strategies are urgently needed that increase treatment retention, while being culturally responsive, provide adequate psychosocial support, and effectively address treatment accessibility 5, 25–27]. Failure to attend appointments results in gaps in buprenorphine treatment and these gaps were associated with a 1.56 to 4.3 times increased risk of overdose [28]. Non-White patients experience significantly shorter durations of continuous buprenorphine receipt than White patients and had a higher risk of complete discontinuation of buprenorphine [29, 30]. Our findings indicate that RPNs increased retention in care for all racial/ethnic groups and adds to the dearth of studies that examine the impact of RPNs on retention in care by race and/or ethnicity. One study examined racial differences in community-based peer recovery support as an adjunct to SUD treatment for women and found that though NHB women were less likely to receive any form of SUD treatment, they had higher retention in the peer intervention program. The qualitative feedback from the NHB participants centered on the benefits of the additional emotional and instrumental support (i.e. assistance with accessing transportation, housing, and employment resources) as well as referrals to treatment programs for their recovery process [31]. Our findings suggest that incorporation of RPN into clinical teams may be a strategy to improve buprenorphine engagement across various racial/ethnic minoritized groups. Further longitudinal evaluation of the effect of RPNs on care retention and resultant overdoses is needed.

The positive effects of RPNs are likely related to psychosocial support that RPNs provide due to shared lived-experience. Studies of race-stratified perceptions of treatment suggest that racial/ethnic minoritized peoples tend to have negative views of opioid agonist therapies and demonstrate a preference for residential rehab, support groups and “cold turkey” abstinence; but all racial groups describe both internalized and external stigma as major barriers to engagement with opioid agonist therapies [32]. This reluctance to participate in treatment with MOUD speaks to impression management, or regulation of information individuals share within a specific context, and employing this to manage their experience with social stigma [4, 33]. Because negative impressions is a barrier to engagement with MOUD, meaningful interventions to improve retention must mitigate the way patients navigate these stigma. A qualitative analysis of RPNs indicated their particular advantages are their shared experience, their role in making care accessible, linking patients to social services, and in motivating behavioral change [24]. RPNs benefit across racial groups may in part be explained by their ability to mitigate stigma associated with treatment for OUD, serving to minimize the harmful experience of stigma that would otherwise lead to discontinuation of treatment.

Additionally, RPNs provide critical instrumental supports which make them especially effective in engaging individuals with sociodemographic factors that have been empirically associated with loss to care such as minority race, male sex, and comorbid substance use disorders [6, 29]. Using reliance on public insurance, housing instability, and living with psychiatric comorbidity as proxy indicators for socioeconomic vulnerability, our study suggests that recovery peer support improves treatment retention among the most vulnerable patients, even within racial/ethnic minority groups. We consider these measures of socioeconomic vulnerability to represent a subset of the population that has been systematically excluded in traditional healthcare models, whether because of poor means to access care, or treatment models that do not consider the intersecting challenges patients face in obtaining care. Based on our findings, having RPNs in clinics may help overcome socioeconomic factors that typically are associated with poor care retention. This may be because RPNs provide assistance that is tailored to patient needs which helps to overcome some of the structural barriers to health care [34]. Our results put forward the idea that peer navigation may have a role in retaining socioeconomically vulnerable populations in care.

Within our cohort, for NHB, lack of insurance was associated with increased odds of experiencing a gap in clinical care even when accounting for retention with RPNs. Numerous studies have demonstrated that lack of insurance is associated with poor access to MOUD [1, 2, 5]. Even once engaged with treatment, those without insurance are at a greater risk of discontinuing care, [35] which is concordant with our study findings. Patients lacking insurance are likely experiencing multiple socioeconomic barriers to care and interventions that directly address those barriers such as direct provision of medications, offering transportation assistance, onsite childcare services, flexible program hours, and implementing mobile outreach are likely needed to adequately engage this population in treatment [34].

RPNs in clinics for opioid use disorder are a promising intervention to improve outcomes for individuals at risk for gaps in care. Their work in the clinic provides patient-centered, culturally sensitive support that we show benefits populations that have otherwise been excluded from treatment. Dedicated trials are needed to demonstrate the benefit of RPNs and substantiate a widespread recommendation for their implementation in outpatient OUD clinics. We also call for continued healthcare reforms and the development of a safety-net OUD treatment system, especially in the context of Medicaid unwinding, which is expected to leave millions of Americans without access to health insurance [36]. As more people lose access to public health insurance, we expect general utilization and reliance on low-barrier care to increase, which will place more demand on already stretched clinics. Anticipating this change in need for services, funding for addiction treatment should be comprehensive and include a safety net that ensures those that need treatment have access irrespective of insurance status.

Limitations

The findings from this study should be interpreted in the context of the study’s limitations. The results are from a single treatment location, the overall sample size was small, and the samples between the pre- and post-RPN periods were unbalanced. The representation of NHW, Hispanic, and other racial/ethnic groups were small, so analyses of RPN impact based on sociodemographic characteristics could not be completed. The small sample size also impacted the precision of the multivariable logistic regression for NHB. The proportion of Hispanic patients in our study was incongruent to that of the city’s population. Possible explanations for this include insufficient outreach to the local Hispanic community, limited provider- or staff-patient cultural congruence, or reduced perception of SUD treatment need in the population. Another limitation of the study was the absence of urinary toxicology results and prescription fill records in the assessment of successful MOUD initiation and treatment adherence. Lastly, the specific interventions used by the RPNs, or their role in linking patients with the clinic were not abstracted from the chart and, therefore, conclusions about most effective practices could not be made.

Generalizability

These findings are promising for other clinics serving a similar demographic. RPNs show great benefit in minority communities where those with OUD also have multiple factors contributing to socioeconomic vulnerability.

Conclusions

In our study, RPNs improved retention in OUD care across all racial/ethnic groups.

Specifically, RPN interaction was associated with increased treatment retention for socioeconomically vulnerable individuals prone to discontinuation of care. More studies are needed in large racially/ethnically diverse patient populations to continue to evaluate the impact of RPNs on within group treatment retention outcomes and to assess the components of engagement with RPNs that are correlated with retention in care.

Funding Statement

Research reported in this publication was supported by the National Institute on Drug Abuse of the National Institutes of Health under Grant K23DA053989. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Authors contributions

AG contributed to data acquisition and interpretation and manuscript preparation and revisions. PS contributed to data acquisition and interpretation and manuscript revisions. EZ contributed to study conception and design and manuscript revisions. ANN contributed to the study conception and design, data analysis and interpretation, and manuscript development and revisions. All authors read and approved the final manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The datasets used and/or analyzed during the current study are not publicly available due to the presence of protected health information but are available from the corresponding author on reasonable request.