The peer to career pipeline: An observational study of peer worker trainee characteristics and training completion likelihood

Abstract

Introduction

Peer recovery support services (PRSS) for substance use disorder (SUD) are a flexible and evidence-based intervention employed across multiple settings and for a variety of populations. These services have expanded over the past two decades, but there is little research on recruitment and training of prospective peer workers – the peer to career pipeline. This study observed training outcomes for applicants to a peer worker scholarship program in Texas.

Methods

A total of 448 participants provided baseline personal history information, and a subset of participants (n=239) completed optional psychosocial surveys. Logistic regression analysis tested associations of personal history and psychosocial variables with three training stage completion outcomes: classroom training completion, placement at an internship site, and full certification.

Results

The greatest decline in advancement between stages occurred in the transition between classroom training (78.1% of participants completed) and internship placement (43.3% of participants completed). Participants were diverse in terms of race/ethnicity and life experiences salient to the peer worker role, but Hispanic/Latinx peer workers were under-represented. Past work with a SUD peer worker, age, and having a bachelor’s degree were positively associated with training stage completion across multiple models, while having basic technological access, being a woman and veteran status were each positively associated with training stage completion in only one model. Years since recovery initiation date, non-monosexual orientation, and quality of life were negatively associated with training stage completion in only one model.

Conclusions

The existing peer workforce may be a key source of recruitment for new peer workers; thus retention of existing workers is key to ensuring continued expansion of these services. Additional support may be required to recruit and retain younger peer worker trainees, men trainees, Hispanic/Latinx trainees, trainees who lack basic technological access, or trainees without bachelor’s degrees. Unanswered questions about the peer workforce remain and must be addressed to ensure that an appropriately diverse workforce is recruited, that disparities in training outcomes are minimized or prevented, and that existing peer workers are well-supported.

1. Introduction

Peer recovery support services (PRSS) for substance use disorder (SUD) are a flexible, evidence-based intervention employed across a variety of settings, at different levels of SUD acuity, and over varying lengths of time (Eddie et al., 2019; Gaiser et al., 2021). The use of PRSS is particularly promising as a means to expand the SUD workforce efficiently and relatively rapidly compared to the multi-year credentialing process of training and licensing SUD clinicians (Beck et al., 2018; Videka et al., 2019). Substantial variation exists between U.S. states in certification requirements for peer workers delivering PRSS, ranging from 6 to 126 training hours (median 46), and 72 to 2000 experiential training hours across states (median 500; Beck et al., 2018). The provision of PRSS is covered by Medicaid in a majority of U.S. states, for both individual and group-delivered PRSS (United States Government Accountability Office, 2020).

In spite of the widespread adoption of state-level certification for peer workers and approval of Medicaid reimbursement in most states, little is known about potential barriers or facilitators to recruitment, training and retention of SUD peer workers. This gap in the evidence base represents a potential barrier to further expansion of the peer workforce, and to ensuring that the peer workforce is diverse and representative of the life experiences upon which the authenticity of the peer worker-to-participant relationship are built. In addition to the central shared experience of SUD recovery, the relationship between peer workers and those they serve may also be strengthened by concordance in racial/ethnic or other identities, as is the case in other health professions (Ma et al., 2019; Shen et al., 2018). To that end, identifying potential disparities in recruitment and successful training of peer workers across axes of race, ethnicity, gender, sexual orientation, and key life experiences such as criminal legal system involvement and veteran status is key. These individual-level factors upon which the sense of authenticity and trusting relationships may be built between peer workers and their participants are likely not the only factors influencing participant outcomes and peer workforce development, but are a critical starting point in understanding how to best recruit, train and ultimately retain the valuable peer workforce.

The shortage of behavioral health providers in Texas is higher than the national average, at 35% compared to 26% nationwide (Kaiser Family Foundation, 2020), and expansion of the peer workforce has been of particular interest in the state. While PRSS are distinct from other forms of behavioral health services, their expansion is complementary to and may help fill gaps in the current system of care by extending existing services after acute treatment and across varied settings (Gaiser et al., 2021). But unanswered questions about recruitment and retention through training pose barriers to this expansion. Demographic and life history experiences may function either as barriers or facilitators of recruitment and retention of peer workers, as may resource-based barriers, such as access to high-speed internet and a computer, income, and caring for dependents. Similarly, features of a trainee’s recovery and wellbeing may serve as barriers to training, including whether they are in very early recovery or more established recovery, their recovery capital (Groshkova et al., 2013; White & Cloud, 2008), quality of life (Kelly et al., 2018), sense of self-stigma or shame about their recovery (Luoma et al., 2008, 2013), or other psychosocial states. Other socioecological levels may also play a role in recruitment and training outcomes for peer workers, such as having friends, family and significant others who are supportive of a trainee’s choice to pursue certification (Tate et al., 2022), but potential organizational, community, and society/policy-level influences are not yet well understood.

Identification of potential individual-level barriers and facilitators to recruitment and retention in training are critical to establishing best practices in supporting an expanded peer workforce, both in Texas and in the U.S. as a whole. If PRSS is to help fill shortages in the behavioral health workforce across the U.S., then best practices for recruiting, supporting and retaining a diverse peer workforce must develop concurrently with the expansion of this workforce. Vulnerabilities in the peer to career pipeline for the recruitment, training, and employment of peer workers may have been exacerbated by the COVID-19 pandemic, thus allowing for a unique opportunity to observe a particularly stressed system.

To begin to address this gap in the literature and advance the evidence base for the SUD peer workforce, this observational study assessed whether any demographic, life history, recovery history or psychosocial variables were significantly associated with greater odds of training stage completion among SUD peer worker trainees in Texas during the COVID-19 pandemic. The study addressed this question in the context of a peer worker training scholarship program, wherein all classroom (approximately $500) and experiential training fees (variable, up to $2500) were covered, removing a substantial financial barrier to training entry. The scholarship did not cover fees associated with background checks (approximately $60), the fee to file the certification application (approximately $60), travel for training or time off from work to attend training-related activities. The Committee for the Protection of Human Subjects at the University of Texas Health Science Center Houston declared this study exempt (IRB No. HSC-SPH-19–0712) and approved the dissertation research project using these data under IRB No. HSC-SPH-21–0768.

2. Material and Methods

2.1. Participant characteristics

The present study is an observational prospective cohort study of 448 participants in a SUD peer worker training program in Texas, enrolled between February 2020 and November 1, 2021. The study observed participant outcomes through September 1, 2022. Prospective participants provided detailed demographic, life history, and recovery history information at baseline as part of their application to the training scholarship. All applicants who self-reported eligibility for peer worker training in Texas on their applications (were at least 18 years old, had at least a high school diploma or GED, proficiency in English, and met certain criteria for criminal legal history) and consented to participation were included as participants in the study. After selection for participation, they study invited participants to complete psychosocial surveys for a $20 e-gift card incentive, and 239 participants (53.3%) completed this second survey. Participants completed applications and psychosocial surveys online, and the study linked the two measures by unique URL generated for each participant’s unique identifier. Each survey and the scholarship application took approximately 30 to 40 minutes to complete in total.

2.2. Measures

Three of the authors at UTHealth developed the scholarship application and psychosocial survey (SCM, JMW, SAM) in collaboration with a state-wide recovery advocacy organization, RecoveryPeople (author JH). We selected items on the scholarship application to reflect the inclusion/exclusion criteria for Texas peer worker certification, to collect data required by the funder, and to collect data about standard demographics. The study refers to demographics, life history, and recovery history variables collected on the scholarship application as personal history variables. The study selected psychosocial variables as those with potential impact upon completion of training. Two training stages and full certification as a peer worker were the primary outcomes.

2.2.1. Training Outcomes.

The study operationalized training completion as a three-part process. First, trainees in Texas are required to complete a core (8–16 hours) and supplemental training (46 hours), referred to here as classroom training, as this component is structured as a typical classroom learning environment, either in-person or delivered by online video conference. Participants had to complete both core and supplemental training to be considered having completed classroom training, and training entities provided daily training logs to verify classroom training completion. The state approves training entities, and participants could elect to take their training at any approved training entity, either in-person or online. Training entities are typically community-based organizations. Second, Texas requires trainees to complete 250 experiential training hours supervised by a certified peer supervisor, which may be a paid or unpaid internship. Training entities engaged in this study had previously indicated that moving from classroom training to securing and starting experiential training is a substantial barrier for trainees, and so this outcome was operationalized as internship placement: the study considered any trainee who began their experiential training hours to have reached this training outcome, even if they did not complete their full experiential training. Trainees could complete their internships at any site, so long as the work in which they engaged was in line with peer worker roles, and was supervised by a certified peer supervisor. Trainees completing internships at sites that did not already employ a certified peer supervisor could use scholarship funds to pay for an external certified peer supervisor to meet with the trainee independently during the course of the 250 hours. RecoveryPeople provided assistance to trainees seeking internship placement sites or external supervision when requested by trainees, and some training entities also served as internship placement sites depending on availability. Finally, the study considered trainees who completed both classroom and experiential training, and who received full certification as a SUD peer worker in Texas to have achieved full certification. The study verified full certification by matching participants to the database of certified peer workers in Texas, maintained by the Texas Certification Board of Addiction Professionals. Participants who applied for a scholarship, were eligible to become a peer worker in Texas, but did not begin classroom training were considered to have dropped out, as were those who began, but did not complete classroom training. The study recorded each stage outcome as a binary, yes/no variable, and assumed participants who completed a more advanced stage (e.g., full certification) had completed all previous stages (e.g., classroom training and internship placement).

2.2.2. Personal History Variables.

The initial scholarship application collected measures of race, ethnicity, age, present gender identity, sexual orientation, education level, disability status, veteran status, household income, number of dependents, and access to technology. Participants could select multiple race categories, and these were operationalized as independent binary variables. Ethnicity was operationalized as Hispanic/Latinx or non-Hispanic/not Latinx. The present gender identity item provided multiple response options capturing an array of gender identities, as well as an open response option; however, transgender, non-binary, and other gender diverse identities were under-represented in this sample, and the study ultimately excluded them from the analysis. The study consolidated sexual orientation response options indicating interest in more than one gender into a single non-monosexual category, and converted asexual to missing, following the methods in Wilkerson et al. (2020). Highest education levels were high school diploma/GED, associate’s degree, bachelor’s degree, and master’s degree or above, with other responses converted to missing. The study measured household income and number of dependents as continuous variables. The study operationalized access to technology as a list of types of technology to which a participant had regular access, and dichotomized participants into “no access” and “basic technology access.”

Recovery history variables included the date of the participant’s own recovery initiation, and whether they had used a recovery community center (RCC), received support from an SUD peer worker, or attended recovery mutual aid meetings such as SMART, Alcoholics Anonymous, etc. The study subtracted the date of the participant’s own recovery initiation from the application date to create a continuous variable representing years since recovery initiation. The study used this measure instead of continuous abstinent time to better reflect variations in self-defined recovery pathway. The study recorded Use of RCCs, SUD peer workers, and recovery mutual aid groups as binary variables.

Life history variables included criminal legal involvement and personal history of trauma, measured as binary variables, with 0 indicating no history and 1 indicating a history of either experience.

2.2.3. Psychosocial variables.

The psychosocial surveys included four measures with good reliability and validity: the Assessment of Recovery Capital (ARC; Groshkova et al., 2013), the WHOQOL-BREF (Skevington et al., 2004), the Substance Abuse Self-Stigma Scale (SASSS; Luoma et al., 2013), and the Patient Health Questionnaire-4 (PHQ-4; Kroenke et al., 2009). The study used total scores for the ARC, SASSS and PHQ-4 in the analysis, and the single item measure of quality of life (item 1) from the WHOQOL-BREF.

2.3. Data Analysis

The study used logistic regression in STATA version 17 (StataCorp, 2021) to assess changes in the likelihood of completing each training stage based on each independent variable, first in a set of simple, univariable logistic regression models, regressing single independent variables against each training outcome (the single independent variable [IV] model). The study grouped personal history variables into a multivariable logistic regression model to test for significant associations with training stage completion outcomes (the personal history model). Similarly, the study also tested psychosocial variables against training stage completion outcomes as a separate multivariable model (the psychosocial model). Finally, the study combined the full set of both personal history and psychosocial variables into a full model for each training outcome (the full model). When testing each training stage completion outcome, participants who had not completed the previous stage were removed from the analysis (e.g., when testing the outcome of full certification, only participants who had achieved an internship placement were included in the analysis). The study selected an arbitrary small category cutoff size of 3% of the analysis sample, as different sizes of analysis samples were used in different stages of the analysis (e.g., the full participant sample is 448, but the participants who completed the classroom training stage and are thus included in the internship placement analysis is 350), thus, while arbitrary, the cutoff is responsive to the size of the analysis group. The study omitted two categorical variables with less than 3% of the total sample from the analysis: the present gender identity response option for transgender, gender non-conforming, non-binary, or other; and Asian American/Pacific Islander race category. The study set statistical significance to p < 0.05.

3. Results

3.1. Descriptive Statistics

Tables 1 and 2 display the descriptive statistics for personal history and psychosocial variables, as well as training completion outcomes for all participants. Figure 1 presents the number of completers and percent of the overall participant pool that achieved each training stage. A majority of trainees were White (72.3%), non-Hispanic/non-Latinx (82.1%) and identified as women (58.6%). Most earned $30,000 or less (68.6%), and a majority had a high school diploma or GED as the highest level of education completed (66.0%). Gay and lesbian (7.8%) and non-monosexual (8.1%) individuals comprised 15.88% of the participants, and gender non-conforming, transgender, or other gender identities comprised a total of 1.1% of the participants. More than half of participants had a history of criminal legal system involvement (53.2%). More than three-quarters of the participants (78.1%) completed the classroom training, and just over one-third advanced to full certification by the end of the observation period (33.7%). The greatest decline in advancement between stages occurred in the transition between classroom training and internship placement (43.3%).

Table 1.

Descriptive statistics for demographic characteristics of participants.

Training Stage Completed	Full Sample n = 448		Classroom Training n = 350		Internship Placement n = 194		Full Certification n = 151
Demographic Characteristics	n	%	n	%	n	%	n	%
Race (participants may select multiple)
Asian American/Pacific Islander	7	1.6	5	1.4	1	0.5	1	0.7
Black or African American	105	23.4	78	22.3	47	24.2	34	22.5
American Indian/Alaska Native [AI/AN]	14	3.1	12	3.4	8	4.1	5	3.3
White	324	72.3	257	73.4	134	69.1	108	71.5
Other, or Not Reported	22	4.9	16	4.6	8	4.1	7	4.6
Hispanic/Latinx	80	17.9	65	18.6	36	18.6	27	17.9
Non-Hispanic/Non-Latinx	368	82.1	285	81.4	158	81.4	124	82.1
Age
18 to 25	25	5.6	18	5.1	6	3.1	5	3.3
26 to 44	212	47.3	168	48.0	87	44.9	65	43.1
45 to 64	193	43.1	151	43.1	91	46.9	72	47.7
65 or older	18	4.0	13	3.7	10	5.2	9	6.0
Present Gender Identity
Man	181	41.4	135	39.0	75	38.7	60	39.7
Woman	256	58.6	205	59.2	116	59.8	90	59.6
Transgender, non-binary, or other	5	1.1	6	1.7	3	1.5	1	0.7
Sexual Orientation
Straight	355	84.1	281	85.4	161	88.0	124	82.1
Gay or Lesbian	33	7.8	23	7.0	13	7.1	11	7.3
Non-Monosexual	34	8.1	25	7.6	9	4.9	6	4.0
Education level
High school diploma/GED	293	66.0	234	67.6	123	63.4	93	61.6
Associate’s degree	76	17.1	58	16.8	34	17.5	26	17.2
Bachelor’s degree or higher	73	16.4	52	15.0	36	18.6	31	20.5
Has a disability	101	22.5	76	21.7	42	21.7	30	19.9
Is a veteran or active military	48	10.7	35	10.0	20	10.3	19	12.6
Number of dependents
0	266	60.1	204	59.0	105	54.1	82	54.3
1	100	22.6	81	23.4	49	25.3	36	23.8
2 or more	77	17.4	61	17.6	39	20.1	32	21.2
Household income
≤$15,000	158	35.3	122	34.9	65	33.5	46	30.5
$15,001 – $30,000	149	33.3	118	33.7	66	34.0	53	35.1
$30,001 – $45,000	76	17.0	58	16.6	34	17.5	30	19.9
$45,001 – $60,000	27	6.0	23	6.6	12	6.2	8	5.3
>$60,000	38	8.5	29	8.3	17	8.8	14	9.3
Technological access
Has basic technological access	368	82.1	278	79.4	163	84.0	127	84.1
Lacks basic technological access	19	4.2	11	3.1	6	3.1	4	2.6
Missing or no response	61	13.6	61	17.4	25	12.9	20	13.2

Table 2.

Descriptive statistics for recovery history, life history, psychosocial, and training outcomes.

Training Stage Completed	Full Sample n = 448		Classroom Training n = 350		Internship Placement n = 194		Full Certification n = 151
Life & Recovery History	n	%	n	%	n	%	n	%
Time since recovery initiation date
Less than 1 year	4	0.9	3	0.9	2	1.0	1	0.7
1 to < 2 years	29	6.5	26	7.4	8	4.1	7	4.6
2 to < 3 years	81	18.1	68	19.4	33	17.0	22	14.6
3 to < 4 years	67	15.0	46	13.1	26	13.4	20	13.3
4 to < 5 years	39	8.7	35	10.0	23	11.9	17	11.3
5 to < 10 years	107	23.9	84	24.0	46	23.7	36	23.8
10 or more years	121	27.0	88	25.1	56	28.9	48	31.8
Used a recovery community center (RCC)	100	25.8	81	23.1	43	22.1	32	24.4
Received support from SUD peer worker	133	34.4	100	31.4	61	31.4	45	34.4
Used recovery mutual aid groups	221	57.1	162	46.3	92	47.4	75	57.3
Has a history of criminal justice involvement	206	53.2	161	46.0	98	50.5	80	61.1
Has a personal history of trauma	292	75.4	219	62.6	128	66.0	100	76.3
Number of optional psychosocial surveys by training stage completed	Full Sample n = 239		Classroom Training n = 177		Internship Placement n = 109		Full Certification n = 89
Psychosocial Variables	Mean	SD	Mean	SD	Mean	SD	Mean	SD
Assessment of Recovery Capital total score (max. 50)	40.8	9.6	40.9	9.8	40.9	9.8	40.5	9.9
General quality of life (WHOQOL-BREF item 1, max. 5)	4.3	0.6	4.3	0.6	4.3	0.6	4.3	0.6
Substance Abuse Self-Stigma Scale total score (max. 200)	70.6	18.3	71.4	17.8	72.0	18.4	72.1	18.6
PHQ-4 total score (max. 12)	2.3	2.7	2.2	2.4	2.2	2.3	2.1	2.1

Figure 1.

Percentage of participants from the original participant pool that completed each training stage.

3.2. Single IV Models

Full single IV model logistic regression results are in table 3. The number of years since recovery initiation was weakly negatively associated with classroom training completion (OR = 0.98, 95% CI [0.96, 0.99]), and past receipt of support from an SUD peer worker was positively associated with classroom training completion (OR = 2.00, 95% CI [1.19, 3.38]). Age was positively associated with internship placement (OR = 1.03, 95% CI [1.01, 1.05]) as was having a bachelor’s degree compared to having a high school diploma or GED (OR = 2.35, 95% CI [1.08, 5.08]). Non-monosexual orientation was negatively associated with internship placement (OR = 0.42, 95% CI [0.18, 0.98]) as was White race (OR = 0.60, 95% CI [0.37, 0.98]). No other variables in the single IV models were associated with training outcomes, including no significant associations with full certification.

Table 3.

Logistic regression results for single IV regression models. Statistically significant (p < 0.05) results are in bolded text.

	Classroom Training	Internship Placement	Full Certification
Total included in analysis	n = 448	n = 350	n = 194
Variable	OR [95% CI]	OR [95% CI]	OR [95% CI]
Race and Ethnicity *
Black/African-American	0.75 [0.45, 1.26]	1.29 [0.77, 2.15]	0.67 [0.32, 1.43]
AI/AN	1.70 [0.38, 7.75]	1.63 [0.48, 5.53]	0.46 [0.11, 1.99]
White	1.28 [0.79, 2.08]	0.60 [0.37, 0.98]	1.64 [0.81, 3.33]
Other, Race Not Reported	0.74 [0.28, 1.93]	0.80 [0.29, 2.17]	2.04 [0.24, 17.07]
Hispanic or Latinx	1.26 [0.68, 2.33]	0.99 [0.58, 1.72]	0.82 [0.35, 1.91]
Age	0.99 [0.98, 1.02]	1.03 [1.01, 1.05]	1.02 [0.99, 1.05]
Present Gender Identity (ref. = man)	1.37 [0.87, 2.16]	1.04 [0.67, 1.62]	0.87 [0.42, 1.77]
Sexual Orientation (ref. = heterosexual)
Gay or Lesbian	0.61 [0.28, 1.33]	0.97 [0.41, 2.28]	1.64 [0.35, 7.74]
Non-monosexual	0.73 [0.33, 1.63]	0.42 [0.18, 0.98]	0.60 [0.14, 2.50]
Education (ref. = High School Diploma/GED)
Associate’s degree	0.81 [0.45, 1.48]	1.28 [0.71, 2.29]	1.05 [0.43, 2.56]
Bachelor’s degree	0.61 [0.31, 1.18]	2.35 [1.08, 5.08]	2.47 [0.69, 8.82]
Master’s degree or higher	0.67 [0.25, 1.79]	1.50 [0.53, 4.27]	1.29 [0.26, 6.41]
Disability status (ref. = no)	0.81 [0.48, 1.36]	0.99 [0.60, 1.65]	0.64 [0.29, 1.39]
Veteran status (ref. = no)	0.73 [0.37, 1.43]	1.08 [0.53, 2.19]	6.05 [0.79, 46.52]
Number of Dependents	1.01 [0.82, 1.25]	1.19 [0.97, 1.48]	1.06 [0.76, 1.47]
Household Income (per $1k)	1.00 [0.99, 1.01]	0.99 [0.99, 1.00]	1.01 [0.99, 1.02]
Has basic technological access (ref. = no access)	2.25 [0.88, 5.76]	1.18 [0.35, 3.96]	1.76 [0.31, 10.02]
Years since recovery initiation date	0.98 [0.96, 0.99]	1.02 [0.99, 1.05]	1.02 [0.98, 1.07]
Used a RCC (ref. = no)	1.62 [0.92, 2.84]	0.74 [0.44, 1.24]	0.79 [0.35, 1.78]
Received support from SUD peer worker (ref. = no)	2.00 [1.19, 3.38]	0.82 [0.51, 1.32]	0.72 [0.34, 1.51]
Used recovery mutual aid groups (ref. = no)	0.08 [0.53, 1.34]	0.85 [0.53, 1.37]	1.65 [0.80, 3.42]
Criminal legal system history (ref. = no)	1.48 [0.94, 2.35]	1.25 [0.78, 2.00]	1.74 [0.84, 3.61]
History of trauma (ref. = no)	1.07 [0.63, 1.82]	0.99 [0.58, 1.72]	1.15 [0.50, 2.63]
Assessment of Recovery Capital total score	1.01 [0.98, 1.04]	0.99 [0.97, 1.03]	0.98 [0.93, 1.04]
WHOQOL-BREF item 1 score	0.90 [0.56, 1.45]	1.01 [0.62, 1.67]	1.91 [0.83, 4.40]
Substance Abuse Self-Stigma Scale total score	1.01 [0.99, 1.03]	1.01 [0.99, 1.02]	1.00 [0.97, 1.03]
PHQ-4 total score	0.94 [0.85, 1.05]	0.99 [0.87, 1.12]	0.96 [0.78, 1.18]

^*= Respondents could select multiple race categories. Race and ethnicity are modeled as separate binary variables with the reference category set to 0, indicating the participant did not select that category.

3.3. Personal History Models

The full results for the personal history variable logistic regression model, which includes all demographic, life history, and recovery history variables, are presented in table 4. Past receipt of support from a SUD peer worker was significantly positively associated with classroom training completion (aOR = 2.22, 95% CI [1.21, 4.06]) when controlling for all other personal history variables. Age was significantly positively associated with full certification (aOR = 1.06, 95% CI [1.01, 1.11]), as was having veteran status compared to not being a veteran (aOR = 11.61, 95% CI [1.02, 131.97]). No personal history variables were associated with internship placement likelihood.

Table 4.

Logistic regression results for personal history variable model. Statistically significant (p < 0.05) results are in bolded text.

	Classroom Training	Internship Placement	Full Certification
Total included in analysis	n = 448	n = 350	n = 194
Variable	aOR [95% CI]	aOR [95% CI]	aOR [95% CI]
Race and Ethnicity *
Black/African-American	0.40 [0.07, 2.28]	0.09 [0.01, 1.50]	2.03 [0.19, 21.96]
AI/AN	1.35 [0.13, 13.93]	0.37 [0.03, 4.77]	^
White	0.43 [0.08, 2.36]	0.06 [0.00, 1.03]	4.54 [0.55, 37.39]
Other, Race Not Reported	0.41 [0.08, 2.23	0.12 [0.01, 2.26]	7.82 [0.44, 140.14]
Hispanic or Latinx	1.31 [0.59, 2.89]	0.98 [0.46, 2.08]	0.72 [0.21, 2.54]
Age	1.01 [0.98, 1.04]	1.03 [0.99, 1.06]	1.06 [1.01, 1.11]
Present Gender Identity (ref. = man)	1.68 [0.96, 2.97]	0.73 [0.40, 1.30]	1.44 [0.54, 3.82]
Sexual Orientation (ref. = heterosexual)
Gay or Lesbian	0.52 [0.22, 1.24]	0.82 [0.30, 2.20]	2.46 [0.25, 24.65]
Non-monosexual	0.77 [0.29, 2.03]	0.67 [0.22, 2.04]	0.21 [0.03, 1.52]
Education (ref. = High School Diploma/GED)
Associate’s degree	0.91 [0.45, 1.83]	1.01 [0.48, 2.15]	0.38 [0.11, 1.33]
Bachelor’s degree	0.80 [0.37, 1.75]	2.29 [0.89, 5.91]	5.27 [0.87, 32.05]
Master’s degree or higher	1.10 [0.35, 3.42]	2.49 [0.68, 9.07]	0.90 [0.14, 5.96]
Disability status (ref. = no)	0.87 [0.46, 1.64]	0.94 [0.47, 1.88]	0.63 [0.20, 1.96]
Veteran status (ref. = no)	0.86 [0.39, 1.94]	0.57 [0.23, 1.44]	11.61 [1.02, 131.97]
Number of Dependents	0.94 [0.73, 1.21]	1.16 [0.88, 1.53]	1.15 [0.71, 1.86]
Household Income (per $1k)	1.01 [0.99, 1.02]	0.99 [0.99, 1.00]	1.00 [0.98, 1.02]
Has basic technological access (ref. = no access)	3.14 [0.96, 10.27]	2.47 [0.54, 11.38]	2.05 [0.23, 18.46]
Years since recovery initiation date	0.98 [0.95, 1.01]	1.01 [0.97, 1.04]	1.03 [0.97, 1.09]
Used a RCC (ref. = no)	1.16 [0.61, 2.23]	0.69 [0.37, 1.29]	0.94 [0.32, 2.76]
Received support from SUD peer worker (ref. = no)	2.22 [1.21, 4.06]	1.13 [0.63, 2.02]	0.91 [0.32, 2.58]
Used recovery mutual aid groups (ref. = no)	0.76 [0.44, 1.31]	0.99 [0.56, 1.73]	2.62 [0.96, 7.14]
Criminal legal system history (ref. = no)	1.49 [0.86, 2.59]	1.48 [0.84, 2.61]	2.15 [0.81, 5.71]
History of trauma (ref. = no)	0.68 [0.34, 1.35]	0.86 [0.43, 1.74]	0.64 [0.20, 2.05]

^*= Respondents could select multiple race categories. Race and ethnicity are modeled as separate binary variables with the reference category set to 0, indicating the participant did not select that category.

^{^}= omitted due to collinearity.

3.4. Psychosocial Models

None of the psychosocial variables tested were associated with significant differences in training stage completion likelihood in the psychosocial model (see table 5), which controlled for other psychosocial variables, but not for personal history variables.

Table 5.

Logistic regression results for psychosocial variable model. No associations were statistically significant (p < 0.05) in this model.

	Classroom Training	Internship Placement	Full Certification
Total included in analysis	n = 448	n = 350	n = 194
Variable	aOR [95% CI]	aOR [95% CI]	aOR [95% CI]
Assessment of Recovery Capital total score	1.02 [0.98, 1.06]	0.98 [0.94, 1.03]	0.97 [0.91, 1.05]
WHOQOL-BREF item 1	0.73 [0.42, 1.25]	1.09 [0.62, 1.93]	2.00 [0.80, 5.01]
Substance Abuse Self-Stigma Scale total score	1.02 [0.99, 1.04]	1.00 [0.98, 1.02]	1.00 [0.97, 1.03]
PHQ-4 total score	0.92 [0.81, 1.04]	0.99 [0.86, 1.15]	0.95 [0.76, 1.19]

3.5. Full Models

The logistic regression results of the full model, which controlled for both personal history and psychosocial variables, are in table 6. Identifying as a woman was significantly associated with more than twice the likelihood of classroom training completion compared to identifying as a man (aOR = 2.40, 95% CI [1.05, 5.53]) in the full model. Having basic technological access was associated with more than six times the likelihood of completing classroom training (aOR = 6.72, 95% CI [1.06, 42.45]). Past receipt of support from a SUD peer worker was associated with three times the likelihood of completion classroom training (aOR = 3.01, 95% CI [1.23, 7.35]). The only psychosocial variable to be significantly associated with any training stages across any of the models was the first item from the WHO-QOL BREF, which is an assessment of overall self-reported quality of life, and was significantly negatively associated with classroom training completion (aOR = 0.48, 95% CI [0.24, 0.96]). Age was associated with significantly higher likelihood of internship placement (aOR = 1.05, 95% CI [1.00, 1.10]), as was having a bachelor’s degree compared to having a high school diploma or GED (aOR = 6.37, 95% CI [1.02, 39.64]). No variables were associated with changes in full certification likelihood in the full model.

Table 6.

Logistic regression results for the full model (including both personal history variables and psychosocial variables). Statistically significant (p < 0.05) results are in bolded text.

	Classroom Training	Internship Placement	Full Certification
Total included in analysis	n = 448	n = 350	n = 194
Variable	aOR [95% CI]	aOR [95% CI]	aOR [95% CI]
Race and Ethnicity *
Black/African-American	0.14 [0.01, 2.66]	0.07 [0.00, 4.09]	0.70 [0.01, 66.50]
AI/AN	^	0.12 [0.00, 5.60]	^
White	0.23 [0.01, 4.21]	0.07 [0.00, 3.34]	1.98 [0.04, 100.29]
Other, Race Not Reported	1.13 [0.07, 17.25]	0.53 [0.02, 12.37]	0.72 [0.01, 40.39]
Hispanic or Latinx	1.18 [0.35, 3.97]	0.62 [0.19, 2.06]	0.42 [0.04, 3.96]
Age	0.99 [0.95, 1.03]	1.05 [1.00, 1.10]	1.06 [0.99, 1.15]
Present Gender Identity (ref. = man)	2.40 [1.05, 5.53]	0.77 [0.31, 1.94]	4.30 [0.82, 22.45]
Sexual Orientation (ref. = heterosexual)
Gay or Lesbian	0.40 [0.13, 1.28]	0.77 [0.17, 3.38]	0.66 [0.05, 9.32]
Non-monosexual	0.35 [0.08, 1.48]	0.65 [0.09, 4.56]	0.12 [0.00, 3.07]
Education (ref. = High School Diploma/GED)
Associate’s degree	1.16 [0.38, 3.56]	0.38 [0.13, 1.14]	0.31 [0.04, 2.50] 1.80 [0.17,
Bachelor’s degree	0.53 [0.18, 1.59]	6.37 [1.02, 39.64]	18.88]
Master’s degree or higher	0.77 [0.20, 2.98]	3.84 [0.49, 30.34]	0.42 [0.03, 6.22]
Disability status (ref. = no)	0.55 [0.21, 1.47]	0.71 [0.23, 2.17]	0.21 [0.03, 1.49]
Veteran status (ref. = no)	1.16 [0.33, 4.12]	0.32 [0.08, 1.26]	^
Number of Dependents	1.38 [0.88, 2.14]	1.17 [0.80, 1.71]	1.00 [0.56, 1.77]
Household Income (per $1k)	1.00 [0.98, 1.01]	0.99 [0.97, 1.01]	1.01 [0.97, 1.05]
Has basic technological access (ref. = no access)	6.72 [1.06, 42.45]	1.38 [0.99, 19.35]	^
Years since recovery initiation date	1.00 [0.95, 1.05]	0.99 [0.94, 1.06]	1.09 [0.96, 1.24]
Used a RCC (ref. = no)	0.76 [0.31, 1.86]	0.56 [0.23, 1.38]	1.08 [0.18, 6.34]
Received support from SUD peer worker (ref. = no)	3.01 [1.23, 7.35]	1.33 [0.54, 3.28]	5.11 [0.76, 34.44]
Used recovery mutual aid groups (ref. = no)	0.73 [0.32, 1.66]	2.07 [0.89, 4.81]	1.83 [0.34, 9.82]
Criminal legal system history (ref. = no)	1.24 [0.54, 2.84]	1.79 [0.75, 4.30]	I. 98 [0.33, II. 76]
History of trauma (ref. = no)	1.10 [0.40, 3.05]	0.38 [0.12, 1.23]	0.26 [0.03, 2.08]
Assessment of Recovery Capital total score	1.01 [0.95, 1.06]	0.95 [0.90, 1.01]	0.96 [0.88, 1.04]
WHOQOL-BREF item 1 score	0.48 [0.24, 0.96]	1.18 [0.57, 2.44]	1.52 [0.45, 5.11]
Substance Abuse Self-Stigma Scale total score	1.02 [0.99, 1.04]	0.99 [0.97, 1.02]	0.98 [0.94, 1.03]
PHQ-4 total score	0.87 [0.74, 1.03]	1.11 [0.90, 1.35]	1.18 [0.83, 1.68]

^*= Respondents could select multiple race categories. Race and ethnicity are modeled as separate binary variables with the reference category set to 0, indicating the participant did not select that category.

^{^}= omitted due to collinearity.

4. Discussion

In this study, the greatest drop-off in participation was in the transition from classroom training to internship placement, despite the scholarship funds and support from project staff available to participants, which may explain why there were mostly null findings for the transition between the latter two stages of training. Structured internship placement infrastructure, such as the Department of Labor’s Registered Apprenticeship Programs (Office of Apprenticeship & U.S. Department of Labor, 2023) for paramedics and other healthcare workers, could provide a model for fortifying this portion of the peer to career pipeline. Other potential barriers in the peer to career pipeline are discussed and contextualized in the following subsections, organized by participant characteristics.

4.1. Race, Ethnicity, Gender, Sexual Orientation, and Age

Overall, those recruited to train as peer workers as part of this study included proportionate or over-representation of groups that are often under-represented in other sectors. The proportion of Black participants in this study (23.4%) was higher than among the U.S. population in recovery (7.92%; Substance Abuse and Mental Health Services Administration [SAMHSA], 2022) and higher than among U.S. adults who had resolved a drug or alcohol problem but did not necessarily identify as in recovery (13.8%; Kelly et al., 2017). The proportion of Hispanic/Latinx participants in this study was higher than national estimates of people in recovery (11.07% of people in recovery; SAMHSA, 2022), but in line with estimates of people who had endorsed resolving an alcohol or other drug problem (17.3%; Kelly et al., 2017). However, Texas has a substantial Hispanic/Latinx population compared to the U.S. as a whole (40.2% and 19.1%, respectively; U.S. Census Bureau, 2020, 2022), thus the Hispanic/Latinx population may be underrepresented among peer worker recruits and targeted recruitment may be warranted. Previous research has found significantly higher perceived stigma toward SUD among the Hispanic/Latinx population compared to non-Hispanic Whites (Keyes et al., 2010; Smith et al., 2010), and this stigma may be a further driver of lower-than-anticipated recruitment of Hispanic/Latinx peer worker trainees. No associations between race/ethnicity and training outcomes were found for any traditionally minoritized or underrepresented racial or ethnic groups.

In this study, most participants were women (58.6%), but more men than women identify as in recovery in the U.S. (59.64% of people in recovery; SAMHSA, 2022). Both the higher prevalence of women and the lower likelihood of men to complete parts of the training in this study are in line with patterns observed in other helping professions; for example, women comprised 70.4% of substance use and behavioral disorder counselors in the U.S. in 2022 (Bureau of Labor Statistics, 2023). To ensure that men with SUD can have greater patient/provider concordance with a peer worker, men seeking to become peer workers may require additional support in completing training.

LGBTQ+ populations are disproportionately impacted by substance use problems (Allen & Mowbray, 2016; Coulter et al., 2018; Gonzales & Henning-Smith, 2017; Wilkerson et al., 2020), and, while gay/lesbian and non-monosexual orientations were relatively well-represented in this study (7.8% and 8.1%, respectively), people with transgender, non-binary, or other gender non-conforming identities were not well-represented among participants (1.1%). However, the prevalence of LGBTQ+ people in SUD recovery in the U.S. is not well-understood, and the representation found in this study may be comparable. Importantly, non-monosexual orientation was associated with significantly lower odds of internship placement – the key transition at which many in the study exited training – thus, additional support may be warranted for non-monosexual prospective peer workers, and additional investigation is needed.

Young adults were also a small proportion of the participants, ranging from three to five percent of each training outcome category, and this was lower than the percentage of people in recovery who are young adults (approximately 7.54% of people in recovery; SAMHSA, 2022). Because young adults are consistently the age category in which the highest prevalence of SUD is observed (SAMHSA, 2022), and because age was significantly positively associated with achieving the key transition point of internship placement, additional investigation and development of strategies to recruit and support young adult peer workers is needed.

4.2. Socioeconomic Factors

Education level, income and access to technology may be the source of additional potential disparities in the peer to career pipeline. Most peer worker certifications across the U.S. require only a high school diploma (Videka et al., 2019), thus the higher likelihood of those with bachelor’s degrees to achieve an internship placement suggests a potential education-based disparity. A study of peer worker trainees in Appalachia found that a peer worker certification was viewed as a steppingstone for career or educational advancement by 28% of participants (Hagaman et al., 2023), which may suggest that at least some trainees in the present study may seek additional educational attainment in the future, with peer worker certification as an initial step. Opportunities for advancement within the field of peer recovery support services should be cultivated so that those steppingstones can lead to promotion within the peer workforce.

The null findings across household income as an independent variable may be due to the removal of the primary financial barrier to peer worker training, as the study was part of a scholarship program that covered all training fees. However, it is important to note that more than one third of participants made less than or equal to $15,000 per year, with an additional one third of participants earning between $15,001 to $30,000 per year. Thus, the typical classroom training fees of approximately $500 may be especially burdensome to potential peer workers. To ensure that the position of peer worker remains an opportunity for financial stability and that the peer to career pipeline includes those with few current resources, funders should continue to make scholarships and vouchers available to cover training costs. Further, 33% of current peer workers in an Appalachian sample perceived that they were financially fragile (probably could not or certainly could not come up with $2,000 in one month; Hagaman et al., 2023), indicating that financial challenges may persist into employment as a peer worker, though whether due to low pay as peer workers or due to continued financial recovery after entering SUD recovery remains unclear.

Having basic technological access may have presented a greater barrier during the study period than during non-pandemic conditions. The online application to become a participant may have led to an under-representation of those who lack technological access, in addition to the shift to primarily online trainings during the height of the COVID-19 pandemic. Thus, a lack of access to high-speed internet and a computer would have posed a substantial barrier to completing the training. A further potential barrier to those without basic technological access is the identification of potential internship sites, as using a search engine to identify potential sites, and finding contact information or submitting online applications would have been hindered. Similarly, the application for initial certification (allowing for a trainee to collect internship hours) and full certification are hosted on the state certification board website. The application materials can be printed and submitted by mail, but no pre-printed copies of these application materials are readily available. Step-by-step instructions for completing the application materials and background checks are also found online. Thus, while representing a small number of the study participants, those without basic technological access would have been at a substantial disadvantage during the study period. A recent study of RCCs found that just under half (46.7%) provide visitors with access to technology (e.g., computers, printers, fax, internet; Kelly et al., 2020), and increased availability of this type of service may help close the technological gap for prospective peer workers, even outside of pandemic conditions.

4.3. Life, Recovery, and Psychosocial Factors

Life experiences that are complementary to the peer worker role were well-represented in this study. One study of a nationally representative sample of people who had resolved a drug or alcohol problem found that 50.5% had some history of criminal legal system involvement (Kelly et al., 2017), which is similar to the present study. The percent of individuals with criminal legal system histories increased from the classroom training completion group (46%) to the full certification group (61.1%), but there was not a significant association with likelihood at any stage. This study did not test interactions between variables, as described further in the limitations section of this paper, but, if interaction between criminal legal system involvement and another variable such as race and ethnicity were tested, it is possible that significant impacts could have been detected. Veterans recruited as peer worker trainees in this study (10.7%) also may be well-represented, but the prevalence of veteran status among people in recovery is unclear. It is also unclear why this study found a positive association between veteran status and full certification likelihood, but the availability of veteran-specific supports, including peer supports (e.g., Blonigen et al., 2021; Chinman et al., 2015; Ellison et al., 2016) may play a role, especially because past support from a peer worker was associated with higher training completion in this study.

Existing peer workers may be an important avenue for recruitment of new peer workers, especially from populations that are underrepresented among trainees as described in the preceding sections. However, additional support may be needed to ensure that trainees advance through the remaining stages because past support from a peer worker was only associated with greater classroom training completion, but not later stages of training completion in this study. The negative association between years since recovery initiation and classroom training completion may point to potential biases within more traditional recovery communities (Andraka-Christou et al., 2022), as peer worker training is typically encouraging of multiple pathways to recovery, including working with participants who may desire non-traditional, non-abstinence-based, or medication-assisted recovery pathways. Newer forms of recovery communities, such as RCCs, have markedly more favorable attitudes toward medications for opioid use disorder (Hoffman et al., 2021) and this dynamic may help explain the negative relationship between recovery time and training stage completion. Finally, it is unclear why – contrary to the expected positive relationship between quality of life and training outcomes – general quality of life would be negatively associated with completing classroom training in the full model, and a future investigation of interactions between variables could provide greater clarity.

4.4. Limitations

The present study examined individual-level factors only and did not examine interactions between variables. Both the specific psychosocial variables selected for this analysis and the incomplete response rate (53%) for the psychosocial surveys may also have contributed to prevent the detection of significant effects on training stage completion likelihood. While the absolute numbers of participants in each variable category reduced with each training outcome category, all variables included in the analysis remained above the 3% threshold set prior to analysis, and their listwise deletion would have impacted other results. The inclusion of small categories above the 3% threshold may account for the large confidence intervals for some findings. Associations that are present in only one model, or that involve small absolute numbers of individuals have reduced generalizability but were still important to include. This study had limited statistical power for the full certification and internship placement groups compared to the classroom training groups, because these later training stages included only participants who had completed the previous training stage, and this decrease in statistical power may account for null findings in table 6. Finally, this study is limited geographically to Texas and limited by the unique circumstances of the COVID-19 pandemic.

4.5. Future Directions

Future research is needed on how to best retain the current peer workforce, especially as they emerged as a key asset to help trainees through part of the peer to career pipeline. Peer workers involved in peer support for veterans have shown promising results (Blonigen et al., 2021; Schutt et al., 2021), and continuing to cultivate this subgroup of peer workers is key. Similarly, efforts to recruit other demographics and life experiences, especially those of Hispanic/Latinx ethnicity and gender diverse individuals, is important to ensure people with SUD can find peer workers who can relate to their experiences and serve as role models from a place of experiential authority. Including education and employment intentions after training in future studies may also shed light on whether peer worker certification is considered a temporary steppingstone to move into other roles, and investigation into mobility within the peer workforce may aid in retaining peer workers who do view the role as a steppingstone.

Understanding potential influences on recruitment and training outcomes of peer workers beyond the individual level, as well as interactions between variables is also needed. A study of mental health peer workers found that those living in counties with a higher local unemployment rate had a lower likelihood of currently working in mental health peer services (Ostrow et al., 2022), but this potential community-level dynamic was not examined in this individual-level study. Similarly, potential interactions between individual-level factors, or between individual and higher-level factors should also be examined. Many unanswered questions about the peer workforce remain and must be addressed to ensure that an appropriately diverse workforce is recruited, that disparities in training outcomes are minimized or prevented, and that existing peer workers are well-supported.

5. Conclusions

The purpose of this study was to identify potential barriers to full certification as a peer worker, and critical transitions at which trainees are lost in the peer to career pipeline. This study demonstrates that internship placement is the primary hurdle to full certification as a peer worker, but that existing peer workers may be an asset to support trainees. This study was limited by small absolute numbers of participants in some categories as well as being a Texas-only study during COVID-19. Despite limitations, this paper introduces new evidence for recruitment, training retention, and future research priorities to improve the peer to career pipeline.

Highlights.

• Securing an internship is the greatest barrier in the peer to career pipeline.

• Current peer workers may be an important source of training support.

• Several key groups are under-recruited or may need more support during training.