Impact of an osteopathic peer recovery coaching model on treatment outcomes in high-risk men entering residential treatment for substance use disorders

Methods

The team at Rowan University School of Osteopathic Medicine received approval from the Rowan University IRB committee (Pro. no. 2020001111). The data evaluation team at Rutgers University School of Social Work’s Center for Prevention Science (hereafter “Rutgers CPS,” New Brunswick, NJ) received approval from the Rutgers Arts and Sciences IRB committee (Pro. no. 2019002500). This study was funded by a discretionary grant from the Substance Abuse and Mental Health Services Administration (SAMHSA, #6H79TI082453). Participants were read a consent script that included the risks, requirements, and compensation, and asked the participant for verbal consent prior to intake interviews. Rutgers personnel provided reminder calls and texts to participants in advance of 6-month follow-up interview dates. Participants received a $30 gift card for completion of the follow-up interview. A waiver of informed consent was granted for completion of follow-up Government Performance and Results Act (GPRA) interviews over the phone.

Study setting, population, and design

Research coordinators at Rowan University School of Osteopathic Medicine (Stratford, NJ) engaged and recruited participants in partnership with staff at Maryville Addiction Treatment Centers across New Jersey beginning in September 2019. Participants had to meet the following criteria: 1) report being a man; 2) be over the age of 18; 3) be diagnosed with an SUD; 4) live with or be at high risk for HIV; and 5) live in the Newark, Atlantic City, or Camden metropolitan areas. Due to the disproportionate impact of the opioid crisis on racially minoritized populations, patients were also given the choice to self-report race and ethnicity as part of the intake questionnaire. There were 28 recruitment goals for the first and second years of the program, with racially diverse samples of 45 and 70 participants respectively, with the goal of recruiting 300 men over a 5-year period. Many participants were already enrolled in a treatment program at a Maryville center. This observational cohort study collected data at intake and 6-month time points. All men who enrolled in the Minority Aids Initiative study during the first two fiscal years of the grant (9/30/2019–9/29/2021) were included in the overall sample (n=115). Similarly, analyses that required intake and 6-month data only included men who had completed their 6-month interview OR were beyond the 6-month data collection window (i.e., 8 months postintake) at the end of Year 2 (i.e., 9/29/2021).

Clinical staff at Maryville assessed participants’ SUD at enrollment. Participants presenting with co-occurring substance use/mental health disorders were referred to a psychiatric nurse practitioner to determine severity, establish stability, and prescribe psychotherapeutic medication if necessary.

Participants were assigned a peer recovery coach at Maryville at enrollment. Coaches had received prior training in medication-assisted treatment (MAT) and medications for opioid use disorder (MOUD). The meeting goals were for coaches to meet with participants at least once a week during the first 6 months of the program and at least once a month thereafter. Coaches administered SAMHSA’s GPRA tool, including items from the Addiction Severity Index (ASI), at intake. Sessions were guided utilizing the Manual for Recovery Coaching, and coaches worked with participants to develop weekly action plans containing steps to address their long-term goals. see Figure 1 for an illustrated patient pathway.

Figure 1:

Participant flowchart.

Results

As of the end of the second year of the Minority Aids Initiative (9/29/21), 117 men were engaged in the recovery coaching treatment program. Among them, 115 opted to enroll in the study and completed GPRA intake interviews. Demographic breakdowns are displayed in Table 1. A total of 114 (99.1%) men identified as cisgender and 1 (0.9%) as transgender. Sixty (52.2%) identified as White, 34 (29.6%) as Black or African American, 4 (3.5%) as Multiracial, and 14 (14.8%) as None of the Above. Of the 28 participants who identified their ethnicity as Hispanic, 22 (78.5%) identified as Puerto Rican, two (7.1%) as Dominican, one (3.6%) as Cuban, and four (14.3%) as Other. The mean participant age was 35.4 years old (range, 19–62 years), and the most represented age group were men aged 25–34 (40.9%).

Substance use breakdowns are displayed in Table 1. Among all participants, opioid use disorders were the most common substance use diagnosis, with 80.0% (n=92) of participants presenting. Following were alcohol use disorders (13.9%, n=16), stimulant use disorders (3.5%, n=4), sedative-hypnotic/anxiolytic use disorders (1.7%, n=2), and hallucinogen use disorders (0.8%, n=1).

For participants who completed recovery coaching sessions (n=109), coaches spent a total of 391.7 h across 321 sessions providing evidence-based coaching and strengths-based case management through motivational interviewing techniques. A multitude of services were discussed, referred to, or linked during these sessions (Table 2).

By the end of year two, 65 participants had either completed 6-month interviews or were beyond the sponsor’s end of the data collection window (i.e., 8 months postintake) determined by the funding sponsor. Of those 65 participants, 53 completed the appointment and respective GPRA interviews (follow-up rate of 81.5%). There was a significant improvement on five of the six NOMs from intake to 6-month time points among these participants (Figure 2). Substance use abstinence increased (x ²=20.00, p<0.001), as did the likelihood for participants to report no health-related, behavioral, or social consequences of drug use (x ²=41.00, p<0.001). Housing stability increased (x ²=7.14, p=0.008), as well as the likelihood for participants to report an absence of criminal involvement (x ²=5.00, p=0.025). Employment and education status, defined as being employed or in an educational program, also increased (x ²=19.17, p<0.001). The change in feeling socially connected increased (i.e., from 90.0% to 98.0%), albeit to a nonsignificant degree (x ²=2.67, p=0.103).

Figure 2:

SAMHSA’s national outcome measures. p<0.001***, p<0.01**, p<0.05*.

With regard to mental health outcomes, participants displayed significant improvements on all measures that were powered for analyses. There was a significant reduction in the likelihood of both depression and anxiety (Ps<0.001) and trouble understanding, concentrating, or remembering (p<0.01), and improvement for being prescribed psychotherapeutic medication (p<0.001). Two ASI measures demonstrated floor effects (0 cases at intake) and one had few cases (<5 at both timepoints), necessitating interpretation of a different test statistic (Fisher’s Exact Test). For these reasons, it was not possible or appropriate to test for symmetry across time points. Full statistics are displayed in Table 3.

Discussion

There is little evidence to evaluate the extent of the efficacy of peer recovery coaching as a method of improving outcomes for those with SUD. The goal of this study was to elucidate the effects that peer recovery coaching has on follow-up rates, NOMs (e.g., drug abstinence, housing stability), and mental health outcomes (e.g., anxiety, depression). We believe that positive changes in these areas would lend themselves to improvements in a participant’s recovery.

Among the current studies, we find a higher rate of substance use abstinence among our cohort (73.6% vs 66.8%) and a meaningful difference in the rate of change (final stable housing/initial stable housing) in housing stability (90.4% vs 21.1%) [35]. In Linton et al. [36], homelessness was associated with relapse and experienced by 38% of their participants. Therefore, a decrease in substance misuse coupled with an increase in housing stability is encouraging support for our peer recovery model. Furthermore, we find a large difference in health, behavioral, and social consequences compared to a report compiling these data from other SAMHSA grants (2134.2% vs 23.7%) [37]. The percentage of participants who deny health, behavioral, or social consequences at the 6-month follow up is less than the average found on other discretionary grants (81.1% vs 84.7%) [37]. Therefore, the vast majority of the difference lies in the percentage of those who deny consequences at intake (3.80% vs 68.2% for other grants). Through our grant, participants’ consequences diminished from nearly universal (96.2%) to around the national level (19.9%). We also observe appreciable rates of change for depression and anxiety (−62.0% and −56.2%), which is higher than what McLellan et al. has shown to improve [29] for these same mental health conditions (−14.5% and −21.8%, respectively)

Our follow-up rate of 81.5% exceeds that of a similar project [4], and overall, our 6-month follow-up rate is above the national rate for other SAMHSA Center for Substance Abuse Treatment discretionary grants [37]. Continued exposure to the recovery coach might foster intrinsic motivation of the participant, thereby allowing maximal therapeutic value and a higher chance of long-term sobriety. During sessions, the four services discussed most often were recovery planning (96.33%), building recovery capital (54.13%), transition from treatment to community (39.45%), and mutual support meetings (35.78%). These all focus on important skills and social connections helpful for maintaining sobriety, demonstrating what the participants needed the most help managing. Recovery planning and mutual support meetings were also the most common topics that led to referrals and linkages. Less often discussed, but still common topics during sessions, were transportation (18.35%), legal services (18.35%), and educational services (18.35%). We believe that several of these topics demonstrate the ability recovery coaches have to tailor sessions to the participants’ particular needs, struggles, and aspirations beyond their SUD.

Because the Minority Aids Initiative is interested in engaging racially minoritized men, we aimed to recruit a sample that is predominantly Black/African American and Hispanic/Latino men. Enrollment of men reporting Black/African American or mixed race and/or Hispanic ethnicity has been as follows across the project: Year 1–40.0% (n=18 of 45); Year 2–66.7% (n=46 of 69); Year 3–73.7% (n=14 of 19); and across the project to date (9/30/2019 to 3/31/2022)−58.7% (n=78 of 133). Although we started with a smaller proportion of Black/African American and Hispanic/Latino men, we have been successful with engaging a higher proportion of minority men in subsequent years. We hope to continue this trend to better understand how to address the health inequities and lack of efficacy in substance use treatment for racially and ethnically minoritized communities.

One limitation of this paper is the reliance on self-reporting from the participants. Given the relationship of trust between the coach and participant, this was intentional. However, this could have led to biased answers. Next, we recognize that the designations “Black” and “African American,” and “Hispanic” and “Latino,” are not synonymous with each other [28]. This is a limitation in the design of our self-reporting, and we hope to rectify this in future iterations of our client intake surveys. Furthermore, we believe that the definition of stable housing could be more insightful if it expanded the criteria for what is considered to be ‘stable housing’ and factored in the duration of time spent in one residence. The GPRA currently only recognizes ‘stable housing’ as “own/rent apartment, room, or house” based on SAMHSA’s definition of stable housing. This may exclude participants who are living with someone else but are otherwise in ‘stable’ housing. If a participant has moved to 10 different houses in the past 30 days but is renting an apartment at the time of the interview, the GPRA would not be able to reflect this as ‘unstable’ housing. Although our use of NOMs followed established definitions for each outcome, it is important to note that some response options contributed more than others. For example, participants were less likely to report abstinence for drugs compared to alcohol, although both are required for the abstinence NOM. Although changes in employment/education were significant, the changes were exclusively related to employment, because no participants reported education (school or job training) activities at intake or 6-month time points. Similarly, the social connectedness outcome was primarily influenced by participants endorsing interactions with family and/or friends who were supportive of their recovery. Finally, we realize that having a control group to which we could directly compare our data would be beneficial in establishing the direct impact that recovery coaching specifically has on participants’ outcomes.

Future iterations of this research could include disparity analyses between groups to see if the NOMs change with Black/African American and Hispanic/Latino men differently compared to White men. We also recommend more robust evaluations of recovery coaching engagement and quality. Although we were able to find significance in a smaller sample size (n=53), we were only powered to detect large effects (26 cases required). Thus, analyzing these relationships with a larger sample is optimal for exploratory analyses (e.g., to account for potential confounding factors) or the detection of medium effect sizes (87 cases required, 107 required when including one control/confounding factor in McNemar’s tests) (see Cohen, 1992). Future research should examine whether other factors (e.g., comorbid psychiatric disorders and infectious diseases, treatment status for those comorbidities) confound the relationships we tested as the sample size grows. Furthermore, the interpretation of ASI outcome changes that were not plausible to analyze (due to an absence of men or few men with symptoms at intake) will benefit from analysis in the future if more men endorse symptoms at intake. Similarly, the analysis of the social connectedness NOM showed a directional but nonsignificant improvement, and future analyses with a larger sample size may identify changes like this achieving statistical significance. We hope to continue recruiting a higher proportion of racially minoritized men for the Minority Aids Initiative in subsequent years to better meet the goals of the grant. To elucidate whether the peer recovery coaching program has a different relationship to outcomes based on participant characteristics, we could examine whether clinical characteristics and outcomes differ by demographic characteristics in future analyses.

We believe that this research has the potential to demonstrate the power behind an essential pillar of osteopathic medicine: the idea that we all are comprised of mind, body, and spirit. When one of these mainstays is ignored, it can be quite difficult for an individual to heal. We strongly feel that this project, in addressing all three of these pillars, will help attenuate the potency of the opioid crisis and bring the philosophy of osteopathic medicine to the forefront of recovery coaching.

Conclusions

In the last few years, we have seen a spike in the number of SUD, especially among Black/African American and Hispanic/Latino communities [31]. Peer recovery coaching is a method of treatment that can be implemented in addiction and drug abstinence, yet it has not been extensively studied. We found that: a participant paired with a recovery coach extensively trained in MOUD can achieve a higher drug abstinence percentage. In addition, we found: a significant increase in the rate of change of housing stability; lower health, behavioral, and social consequences; lower depression and anxiety evaluations; longer participant-recovery coach exposure time; and higher follow-up rates of participants in the community by their 6-month follow up.