Job assessments and the anticipated retention of behavioral health clinicians working in U.S. Health Professional Shortage Areas

Abstract

Background

A shortage of behavioral health clinicians impedes access to mental health services nationwide in the U.S., with shortages most acute in federally designated Mental Health Professional Shortage Areas (mHPSAs). Retaining behavioral health clinicians currently working in mHPSAs is thus critical. This study sought to identify behavioral health clinicians’ assessments of various aspects of their work and jobs that are associated with their anticipated retention within mHPSA practices.

Methods

Data for this cross-sectional study were drawn from an annual feedback survey of clinicians when they complete federal education loan repayment support contracts for their work within mHPSAs, from 2016 to 2023. Clinicians’ assessments of various aspects of their work and jobs were measured with validated survey items with Likert-scaled response options, with most combined into scales for analyses. Bivariate and then adjusted associations with 5-year anticipated retention were assessed for clinicians’ assessments of various aspects of their work and jobs controlling for demographic, professional, and community characteristics.

Results

The 2,587 respondent behavioral health clinicians (67.5% response rate) included 42% licensed clinical social workers, 39% licensed professional counselors, 12% psychologists, and 7% licensed marriage and family therapists. Two-thirds of these clinicians worked in either community mental health centers or federally qualified health centers. 42% anticipated they would remain in their practices at least another five years. Five-year anticipated retention rates were nearly three times higher for clinicians who indicated satisfaction on global work and practice assessment measures than for clinicians neutral or dissatisfied on these measures. Five-year anticipated retention rates were also higher for clinicians who reported they had an effective and supportive administration, felt well and fairly compensated, had jobs that permitted a good work-life balance, and had jobs that allowed them to practice the full range of services they desired.

Conclusions

How behavioral health clinicians view their jobs within U.S. mHPSAs is important to their anticipated retention. Based on study findings, to promote their retention practice administrators should provide fair and adequate compensation, foster work-life balance, permit them to deliver the services they wish to provide, and value them and their input and maintain good relationships with them.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12913-025-12698-6.

Introduction

The shortage of behavioral health clinicians in the U.S. is most acute for the 122 million people living in the geographic areas and communities designated as federal Mental Health Professional Shortage Areas (mHPSA) [1]. To address the clinician needs of mHPSAs, federal and state programs principally aim to recruit more behavioral health clinicians through education loan repayment incentives and by exposing learners to the opportunities and rewards of careers in these settings [2–5].

But with the typical 15 to 40% annual job turnover rates for behavioral health clinicians in many U.S. settings, recruiting more clinicians into mHPSAs addresses only half of the shortage problem. It is equally important for the facilities where behavioral health clinicians work within mHPSAs—community mental health centers, substance use treatment facilities, federally qualified health centers, and other safety net practices providing care to economically at-risk populations—to provide jobs and work environments that meet clinicians’ needs so they will not promptly leave for more favorable employment elsewhere [6–9]. This study seeks to identify the things about behavioral health clinicians’ jobs and work within mHPSAs that affect their retention and could be addressed by their safety net practices.

The richest and most relevant empiric literature to anticipate how jobs and work might affect retention for behavioral health clinicians are the studies of the retention and, conversely, turnover of the broader range of providers, nurses and administrative staff of community mental health facilities, hospital psychiatric units, and other health and human service agencies in the U.S. Based on this wider literature, employers of behavioral health clinicians within mHPSAs would provide competitive salaries, create a positive and collegial work environment, and structure jobs to permit a healthy work-life balance [5–7, 10–13]. Employers would allow clinicians to work at the top of their license utilizing the full range of their professional skills and provide opportunities for career advancement (Hartsell A, Noecker A: Quantifying the cost of advanced practice provider turnover, unpublished), [5, 10, 14]. Employers would also support a sense of personal fulfillment and joy through work by aligning the practice’s mission with clinicians’ values, helping clinicians feel their work is important and valued, and helping them maintain strong clinician-client relationships [7, 10, 12]. And most critically, employers would build strong, collegial administrative leadership that is viewed as fair and that listens to and supports clinicians [7, 10–13, 15, 16].

It is not known which of these recommendations based on a rather broad literature in terms of disciplines, levels of training, work roles, and work settings apply specifically to behavioral health clinicians with master’s and doctoral level training whose principal role is counseling or psychotherapy: licensed clinical social workers, licensed professional counselors, psychologists, and licensed marriage and family therapists [17]. Furthermore, it is not known which recommendations apply to these disciplines when working in the principally outpatient safety net practices of mHPSAs.

The goals of this study are, for clinicians of these four behavioral health disciplines working within U.S. mHPSAs, to identify how their assessments of various aspects of their jobs and work correlate with whether they anticipate that they will still be working in their practices in five years, when about half would be expected to have left. This study uses anticipated retention as its outcome, as the strongest predictor of actual turnover and therefore the commonly used proxy in job turnover and retention studies, including those of behavioral health clinicians [12, 15, 18].

Methods

Subjects

This cross-sectional study uses questionnaire data from a cohort of behavioral health clinicians working in a variety of types of safety net practices within mHPSAs in the majority of U.S. states. The commonality is that they each received education loan repayment support from the National Health Service Corps (NHSC) while working within the 33 participating states of the Provider Retention and Information System Management (PRISM) Collaborative [3, 19, 20]. The PRISM Collaborative is a group of states that since 2012 has administered questionnaires querying work experiences and soliciting feedback from clinicians receiving education loan repayment support through various state and federal programs. Some states choose not to participate in PRISM, most often when they do not need or prioritize the information it provides, or cannot afford its modest annual participation fee.

The Bureau of Health Workforce of the federal Health Resources and Services Administration regularly provides the Collaborative with roster information on clinicians receiving loan repayment support through the NHSC. Subjects for this study were the behavioral health clinicians completing their first two or three-year term of loan repayment support through the NHSC’s Loan Repayment Program (LRP), Rural Community LRP, and Substance Use Disorder Workforce LRP from January 1, 2016, through December 31, 2023. Eligible behavioral health disciplines were Licensed Clinical Social Workers (LCSW), Licensed Professional Counselors (LPC) (inclusive of Licensed Mental Health Counselors), Psychologists, and Licensed Marriage and Family Therapists (LMFT), disciplines selected because their work principally involves behavioral health counseling and psychotherapy, and their training requires a master’s or doctorate [21, 22]. Clinicians of these disciplines were excluded if they (1) had worked in their current jobs for less than six months and perhaps could not yet knowledgeably report on all aspects of their work, (2) worked in solo private practices which would not fit questionnaire items written for employed clinicians working with professional colleagues, and (3) worked fewer than 30 h per week because turnover is generally greater for parttime workers and may occur for different reasons [23].

Survey processes

With PRISM’s ongoing survey processes, emailed requests to complete online feedback questionnaires are sent when clinicians’ loan repayment support terms end. Survey requests typically carry the signature of the director of participating states’ Primary Care Office or another state workforce office. In emailed requests, clinicians are informed that the principal purpose of the questionnaire is for program feedback and improvement, that participation is voluntary, and that responses would not be shared with employers and only disseminated in aggregate. Studies using these data gathered principally for administrative purposes have been deemed exempt from human subjects review by the University of North Carolina at Chapel Hill Office of Human Research Ethics (Study 12–0626). As such, informed consent from clinicians is not required. This study was conducted in accordance with both the U.S. Federal Policy for the Protection of Human Subjects and the Declaration of Helsinki.

Questionnaire contents

Data used in this study were drawn from the questionnaire PRISM routinely administers to clinicians as their loan repayment ends, that is, it is not a questionnaire designed for this study. The questionnaire initially has clinicians confirm information on the name and location of their current practices, or update this information if they have relocated. They answer all subsequent survey questions with respect to this identified current practice.

Key questionnaire information for this study is clinicians’ assessments of various aspects of work and jobs shown to affect turnover of professional and other staff in studies of mental health and other human services agencies, as outlined earlier. Clinicians indicate their agreement or disagreement with a series of statements using a five-point Likert response scale (1, strongly disagree; 2, disagree; 3, neutral; 4, agree; 5, strongly agree). These statements were drawn both from the Physician Worklife Survey and from previous studies of work satisfaction and job turnover for primary care and behavioral health clinicians [24–28]. Examples of these statements include, “My total compensation package, including benefits, is fair,” “My work leaves me enough time for my personal life,” “I have a good relationship with the practice administrator,” “I feel that I am doing important work,” and “I am able to provide the full range of services for which I was trained and wish to perform.” Using the same five-point Likert response scale, clinicians also indicate their level of agreement with three global work assessment statements, including “I would recommend my practice to others of my discipline.” Clinicians also report the number of hours they work each week total and in administrative roles.

Information used as control variables included clinicians’ responses to three questions querying how they and their families are faring in their communities: “I/We enjoy the activities the community offers,” “I/We have access to most of the things we like to do,” and “I/We feel safe in the community.” Responses to these three items were combined into a scale of clinician-community fit (Cronbach’s Alpha, 0.79). Demographic information, the month and year clinicians began working in their current practice, the year they completed their clinical professional degree, and the amount of their current remaining education debt are also obtained from questionnaires.

The outcome variable in analyses was derived from responses to the open-ended question “Looking ahead from now, about how many years do you anticipate remaining in your current practice?” This statement of anticipated retention has been shown to predict closely primary care physicians’ subsequent actual job retention through at least five years and has been used previously with behavioral health clinicians [27, 29]. In this study’s analyses, clinicians’ anticipated retention durations were dichotomized to four or fewer years versus five years or more.

Behavioral health clinicians’ discipline and birth month and year were identified from NHSC records at the time they began receiving loan repayment support, as were the types of practices where they worked. Clinicians who had changed practices during their loan repayment terms indicated its type on questionnaires from a list of options.

The zip codes of clinicians’ practices drawn from NHSC records were mapped to counties and then to 2013 Urban Influence Codes through the Area Health Resource File of the Health Resources and Services Administration [30, 31]. UIC’s were collapsed into a rural-urban dichotomous indicator based on codes of 1 or 2 (urban) and 3 through 12 (rural).

Analyses

We used factor analysis with Varimax rotation to collapse clinicians’ Likert-scaled reported levels of agreement with 12 statements about their practices, jobs and work into their fewer underlying concepts for these clinicians, which also minimized collinearity. A best, four-factor matrix solution retained 10 of the 12 items, explaining 75.1% of variance. These factors were: Factor 1, Having good administration (3 items: The administrator of my practice/organization is effective. I have a good relationship with the practice administrator. I have real input into administrative decisions. Alpha, 0.85); Factor 2, Finding meaning in work (3 items: I feel that I am doing important work. I feel a strong personal connection with my patients. I fully value the mission of my practice. Alpha, 0.76); Factor 3, Having a good work-life balance (2 items: Work rarely encroaches on my personal time. My work leaves me enough time for my personal life. Alpha, 0.82); Factor 4, Feeling well and fairly compensated (2 items: I am not well compensated given my training and experience (reverse coded). My total compensation package, including benefits, is fair. Alpha, 0.61). The remaining two survey items were incorporated individually in analyses: I am able to provide the full range of services for which I was trained and wish to perform; staff in my practice are a major source of personal support.

All variables were first characterized with descriptive statistics. Bivariate associations with 5-year anticipated retention for all clinician, community and job and work variables and scales were assessed with chi-square tests. Binary logistic regression was then used to test for independent associations with 5-year anticipated retention for each of the six scale and individual item assessments of clinicians’ jobs and work. Regression models included as control variables all reported demographic, professional, community fit, and job characteristics, with several exceptions. Clinician race and ethnicity were not included as recommended when there are no direct postulated race and ethnicity relationships [32]. Information on the types of practices where clinicians work, how long they had worked there, whether they worked more than 44 h per week and their assessments of the amount of flexibility in work hours were not included in regression models as potential confounding and intervening variables in the relationships between clinicians’ assessments of various aspects of their work and their anticipated retention. The number of years since clinicians completed their professional degrees was not included because of collinearity with clinician age. Gender was treated as a binary variable in regression models by excluding the very few subjects reporting a third, non-binary gender because of small group size. To adjust for possible changes in anticipated retention and its correlates during the COVID- 19 pandemic (2000–2001) and after pandemic (2002–2003), two dummy variables for these periods were included in models [33].

The 2.5% of otherwise eligible respondents who did not respond to the questionnaire item querying their anticipated retention, the study outcome variable, were omitted from analyses. Variables included in the regression model had up to 3.6% missing values: missing values were not imputed. A 0.05 false discovery rate-adjusted p value of 0.039 was set for statistical significance using the Benjamini-Hochberg procedure to account for the 6 job assessment variables of this study’s final regression of anticipated retention in addition to three chi-square tests of association with anticipated retention for three global satisfaction indicators [34]. All analyses were carried out on SPSS Version 29.0 (IBM Corporation, Chicago).

Results

A total of 3,110 (67.5%) of 4,608 surveyed clinicians responded. Response rates varied by discipline: LPCs, 70.9%; LCSWs, 66.5%; psychologists, 65.7%; LMFTs, 60.1% (p <.001). Of the 3,110 total respondents, 2,587 met all eligibility criteria, including working more than 30 h per week, and were included in analyses.

Description of clinicians and their communities

The 2,587 eligible respondents included 42% LCSWs, 39% LPCs, 12% psychologists, and 7% LMFTs (Table 1). Their mean age was 41.8 years, and they were predominantly female (81.3%), non-Hispanic White (76.7%), and married (60.9%). Half had children living at home. Respondents had completed their professional degrees an average of 9.1 years earlier, and 40% had less than $10,000 remaining in education debt while 20% still had more than $80,000 in debt remaining.

Table 1.

Behavioral health clinician demographics, professional characteristics, current education debt, and rural vs. urban community location (n = 2,587)

	Number*	Percent of total*	% anticipating retention of 5 or more years	p-value
Period when surveyed
Pre-COVID (2016–2019)	995	38.5%	60.3%	0.20
During COVID (2020–2022)	479	18.5%	55.7%
Post-COVID (2023–2024)	1,113	43.0%	57.5%
Demographics
Age (years)
26–33	553	21.4%	54.1%	< 0.001
34–39	820	31.7%	53.5%
40–49	698	27.0%	62.5%
50 +	515	19.9%	64.5%
Gender
Female	2,043	81.3%	61.4%	0.26
Male	466	18.5%	57.4%
Third gender	5	0.2%	80.0%
Race and Ethnicity
White	1,950	76.7%	59.2%	0.04
Multi-Race and Ethnicity	323	12.7%	54.8%
Black/African American	172	6.8%	51.7%
Asian	34	1.3%	47.1%
Native American/Alaska native	26	1.0%	84.6%
Other	19	0.7%	63.2%
Hispanic	16	0.6%	56.3%
Hawaiian or other Pacific Islander	2	0.1%	50.0%
Marital status
Married	1,554	60.9%	59.7%	0.08
Not married	999	39.1%	56.2%
Parenthood
Has children	1,233	49.1%	62.8%	< 0.001
No children	1,280	50.9%	53.9%
Professional characteristics
Discipline
Licensed clinical social worker	1,091	42.2%	55.8%	0.013
Licensed professional counselor	1,001	38.7%	58.0%
Psychologist	311	12.0%	65.9%
Licensed marriage and family therapist	184	7.1%	60.9%
Number of years since completing clinical professional degree (years)
2–5	425	16.7%	56.7%	0.085
6–9	1,241	48.8%	56.8%
10 +	877	34.5%	61.3%
Remaining education debt
$0–9,999	1,023	40.8%	56.3%	0.26
$10,000–34,999	477	19.0%	58.3%
$35,000–79,999	509	20.3%	60.5%
$80,000 or more	498	19.9%	60.8%
Community settings
Has close family living within the state or convenient driving distance
Yes	1,963	76.8%	59.8%	< 0.004
No	593	23.2%	53.1%
Community fit: Clinician and family enjoy community, what it offers, and feel safe (scale)
Strongly agrees	633	25.2%	62.7%	< 0.001
Agrees	1,313	52.3%	60.0%
Neutral to strongly disagree	566	22.5%	49.1%
Practice county urban vs. rural
Urban	1,673	64.8%	54.9%	< 0.001
Rural	907	35.2%	64.6%

*counts total less than 2,587 due to missing data; % figures are for those without missing data

Nearly two-thirds of these behavioral health clinicians worked in urban counties and one-third in rural counties. Three-quarters reported having close family living within convenient driving distance, and three-quarters responded in the “agree” or “strongly agree” range on the three-item composite scale that they and their families enjoyed what their communities offered.

Two-thirds of these behavioral health clinicians worked in either federally qualified health centers and other types of community health centers or in community mental health centers. The remaining third worked in correctional facilities, Indian Health Service and tribal sites, health departments, substance use disorder facilities, and other various safety net settings.

Descriptions and assessments of jobs and work

Behavioral health clinicians had worked in their current practices a mean of 4.9 years (range, 0.5 to 26 years), and they worked a mean of 41.0 h per week (Table 2). In addition to their clinical work, about one in six served administrative roles 10 or more hours per week. Most respondents reported having the flexibility they needed in their work hours, but one-third did not.

Table 2.

Characteristics and assessments of behavioral health clinicians’ practices, jobs and work and bivariate associations with their 5-year anticipated retention (n = 2,587)

	Number1	Percent	% anticipating retention of 5 or more years	p-value
Type of practice
FQHC and other community health centers	886	34.3%	51.9%	< 0.001
Community mental health facility	848	32.8%	59.8%
American Indian health facility	165	6.4%	62.4%
Private practice	165	6.4%	59.4%
Correctional facility	93	3.6%	78.5%
Health department	55	2.1%	66.0%
Substance use disorder facility	28	1.1%	60.7%
Other practice settings2	342	13.2%	62.0%
Clinician tenure in current practice (years)
0.5 to 3.49 years	1,075	41.6%	54.3%	< 0.001
3.5 to 6.99 years	1,002	38.7%	58.9%
7 or more years	510	19.7%	65.3%
Total work hours per week
30–39	185	7.2%	57.8%	0.57
40–44	2,173	84.0%	57.9%
45 or more	229	8.9%	61.6%
Having needed flexibility in work hours
Strongly Agree	419	16.4%	68.7%	< 0.001
Agree	1,191	46.7%	61.5%
Neutral	425	16.7%	55.8%
Disagree	388	15.2%	47.4%
Strongly Disagree	127	5.0%	32.3%
Administrative roles 10 or more hours per week
Yes	420	16.2%	65.5%	0.001
No	2,167	83.8%	56.9%
Overall, I am satisfied with my current practice
Strongly agree	605	23.8%	80.2%	< 0.001
Agree	1,284	50.6%	63.2%
Neutral, disagree or strongly disagree	651	25.6%	27.8%
Overall, I am pleased with my work
Strongly agree	706	27.6%	73.8%	< 0.001
Agree	1,487	58.2%	57.7%
Neutral, disagree or strongly disagree	361	14.1%	29.4%
I would recommend my practice to others of my discipline
Definitely	1,269	49.7%	73.4%	< 0.001
Probably	867	33.9%	50.4%
Unsure, probably not or definitely not	418	16.4%	28.5%
Scale 1: Having good administration3
Strongly Agree	427	16.8%	78.0%	< 0.001
Agree	1,018	40.0%	66.3%
Neutral	759	29.8%	48.1%
Disagree	274	10.8%	34.3%
Strongly Disagree	68	2.7%	25.0%
Scale 2: Finding meaning in work4
Strongly Agree	1,505	59.2%	62.9%	< 0.001
Agree	970	38.2%	52.4%
Neutral	56	2.2%	39.3%
Disagree	5	0.2%	20.0%
Strongly Disagree	5	0.2%	60.0%
Scale 3: Having good work-life balance5
Strongly Agree	260	10.2%	70.0%	< 0.001
Agree	692	27.1%	64.2%
Neutral	777	30.5%	57.3%
Disagree	651	25.5%	53.6%
Strongly Disagree	171	6.7%	37.4%
Scale 4: Feeling well and fairly compensated6
Strongly Agree	271	10.6%	73.4%	< 0.001
Agree	580	22.8%	66.9%
Neutral	1,070	42.0%	57.7%
Disagree	449	17.6%	46.8%
Strongly Disagree	177	6.9%	38.4%
Feeling personally supported by staff
Strongly Agree	704	27.6	67.3%	< 0.001
Agree	1,014	39.7	59.0%
Neutral	541	21.2	55.1%
Disagree	219	8.6	42.5%
Strongly Disagree	74	2.9	29.7%
Being able to provide the full range of services desired
Strongly Agree	996	39.0%	69.1%	< 0.001
Agree	1,161	45.5%	56.8%
Neutral	222	8.7%	37.8%
Disagree	143	5.6%	32.2%
Strongly Disagree	31	1.2%	29.0%

¹counts total less than 2,587 due to missing data

²includes designated Rural Health Clinics, school-based clinics, mobile units and other types of practice

³Scale 1: Response average to: The administrator of my practice/organization is effective; I have real input into administrative decisions; I have a good relationship with the practice administrator

⁴Scale 2: Response average to: I feel that I am doing important work; I feel a strong personal connection with my patients; I fully value the mission of my practice

⁵Scale 3: Response average to: Work rarely encroaches on my personal time; My work leaves me enough time for my personal life

⁶Scale 4: Response average to: I am not well compensated given my training and experience (reverse coded); My total compensation package, including benefits, is fair

Three-quarters of clinicians agreed or strongly agreed with statements that they were overall satisfied with their practice and overall pleased with their work. A total of 83.6% indicated that they would recommend their practices to others within their discipline.

Behavioral health clinicians’ assessments of various aspects of their jobs and work varied from strongly positive to somewhat negative. Nearly all clinicians (97.4%) found meaning in their work—recognizing its importance, having strong personal connections with patients, and valuing the mission of their practices. A strong majority (84.5%) agreed that they were able to provide the full range of services they wanted, and two-thirds (67.3%) indicated that they felt personally supported by the staff of their practices. More than half (56.7%) felt their practice had a good and supportive administration, but just over one-third (37.3%) indicated they had a good work-life balance and one-third (33.4%) felt well and fairly compensated.

Reporting a good work-life balance was more common for those who reported having flexibility in their work hours than those without flexibility (45.9% vs. 22.8%, respectively, p <.001) and for clinicians who worked fewer than 44 h per week than those who worked more (39.0% vs. 20.4%, p <.001).

Bivariate correlates of anticipated retention

Based on the reported number of years into the future these behavioral health clinicians estimated they would remain in their current safety net practices, 76.7% foresaw remaining at least another year, 66.8% at least another two years, and 42.2% another five years or more (Fig. 1).

Fig. 1.

Percentage of behavioral health clinicians who anticipate remaining in their current safety net practices over the years ahead

About three-quarters of clinicians who strongly agreed with each positive global assessment statement about their practice and work, e.g., “Overall, I am satisfied with my current practice,” anticipated they would remain in their practices at least another five years. Conversely, only about one-quarter of respondents who disagreed with or were neutral on these global work assessment statements saw themselves still working in their practices in five years.

Without controlling for other factors, the percentage that anticipated remaining five or more years was greater for older than younger clinicians, for clinicians who had children than for those who did not, and for psychologists than the other disciplines (Table 1, right columns). Five-year anticipated retention was also statistically greater for clinicians working in rural areas than urban areas, those who reported close family within a convenient driving distance than those without family nearby, and those who enjoyed their community, what it offered, and felt safe there than those felt otherwise. Five-year anticipated retention did not differ by gender, marital status, number of years since clinicians completed their clinical professional degree, amount of remaining education debt, or whether clinicians reported their anticipated retention in pre-COVID, COVID, and post-COVID years.

Among job and work characteristics and without accounting for other factors, five-year anticipated retention percentages were lowest for clinicians working in federally qualified health centers and other types of community health centers, and highest for those working in correctional facilities (Table 2, right columns). Anticipated retention rates were higher for clinicians who had worked within their practices longer than those who were newer to their practices, and those who served administrative roles 10 or more hours per week versus those who had no or smaller administrative roles. Five-year anticipated retention percentages were also longer for clinicians who rated their practices and jobs higher in each of the six assessed areas, specifically in: (1) having good administration, (2) finding meaning in work, (3) having a good work-life balance, (4) feeling well and fairly compensated, (5) feeling personally supported by staff, and (6) being able to practice the full range of services desired.

Correlates of five-year anticipated retention accounting for other factors

Using logistic regression to account for demographic, professional and community factors, as well as the calendar period when clinicians completed surveys, anticipated retention remained associated with four of the six queried aspects of their work and job: having good and supportive administration (odds ratio (OR), 1.56, p <.001), having good work-life balance (OR, 1.15, p =.008), feeling well and fairly compensated (OR, 1.27, p <.001), and being able to provide the full range of services desired (OR 1.27, p <.001) (Table 3). The two areas of job and work assessments were no longer associated with anticipated retention after controlling for other factors: finding meaning in work (OR 1.05, p =.63) and feeling personally supported by other staff (OR, 1.07, p =.20). An additional file provides the results of the full regression model (see Additional file 1).

Table 3.

Associations of behavioral health clinicians’ assessments of various aspects of their jobs and work with their 5-year anticipated retention after controlling for other factors¹: results of logistic regression (n=2,587)

	Anticipating retention 5 years or longer
	Exp(B)	p
Work scale 1: Having good administration (5-point scale)	1.56	<.001
Work scale 2: Finding meaning in work (5-value scale)	1.05	.63
Work scale 3: Having good work-life balance (5-value scale)	1.15	.008
Work scale 4: Feeling well and fairly compensated (5-value scale)	1.27	<.001
Feeling personally supported by staff (5-value Likert response)	1.07	.20
Being able to provide full range of services desired (5-value Likert response)	1.27	<.001
Model Chi-square	350.87	<.001
Nagelkerke R square	.194	-

¹Fully adjusted models include clinician gender (F/M), marital status (married/not married), children at home (Y/N), age (26-33, 34-39, 40-49, 50+), current education debt (10,000-34,999; 80,000+), discipline (LCSW, LPC, LMFT, Psychologist), serving administrative roles 10+ hours per week, having close family within convenient driving distance (y/n), clinician-community fit scale (1 to 5 continuous), county practice location (rural/urban); pre-COVID period vs. COVID period vs. post-COVID period

Discussion

This study sought to understand how behavioral health clinicians’ assessments of their jobs and work within U.S. mHPSAs correlate with their anticipation that they will still be working in their practices in five years. It found that four out of ten of its 2,587 licensed clinical social workers, licensed professional counselors, psychologists and licensed marriage and family therapists thought they would still be working in their practices in five years. As evidence of the overall importance of their jobs and work situations to their thoughts of remaining in or leaving their practices, 5-year anticipated retention rates were nearly three times higher for clinicians who indicated satisfaction on global work and practice assessment measures than clinicians who were neutral or dissatisfied on these measures.

After controlling for other factors, four of six tested assessments of various aspects of jobs and work drawn from the broader retention literature were found to be independently associated with 5-year anticipated retention for these mHPSA behavioral health clinicians. Specifically, anticipated retention rates were higher for clinicians who reported a good and supportive administration, felt well and fairly compensated, had a good work-life balance, and were able to provide the full range of services they desired.

Previous studies of clinicians and staff in behavioral health settings have demonstrated the central importance to employee retention of administrative leadership that is effective and understands, respects, supports, and empowers clinicians [15, 35]. This fact reflects a fundamental truism in management: employees leave managers, not jobs. The field of nursing has long recognized the importance of skilled and effective unit and organization-level administration in preventing nurses’ stress, burnout, and turnover, and numerous interventions to strengthen nursing administration have been described and assessed [10, 21, 36, 37]. In the behavioral health field, the negative effect on retention has been demonstrated for a rigid organizational culture that, among other things, excludes clinicians from management decisions [38]. Other than this, there has been little inquiry into the role of administration in behavioral health clinician retention, and to our knowledge, no published interventions to upskill administrators and improve administrative practices, whether in mHPSAs or other communities.

The importance of salary dollar amounts to job turnover and retention has also been previously demonstrated for employees in behavioral health care settings, including clinicians [7, 39, 40]. This study demonstrated, relatedly, that behavioral health clinicians’ sense that their compensation is fair is linked to their anticipated retention. Studies for other health care disciplines and fields suggest that perceived fairness in pay predicts anticipated turnover better than actual compensation levels [41, 42]. This study’s participants received loan repayment, which may affect perceptions of their compensation, that is, compensation at a given level but in a job that qualifies for loan repayment may be more often viewed as fair. As a whole, the behavioral health field struggles with low salary compared to rising student loan debt amounts.

As also previously demonstrated for clinicians of other health care professions and settings, this study’s behavioral health clinicians working in mHPSAs in jobs that permitted a good work-life balance more often anticipated remaining in their jobs [43, 44]. In our study, clinicians who reported a good work-life balance more often reported they had the flexibility they needed in their work hours and less often reported working more than 44 h per week, pointing to some contributors to work-life imbalance as well as potential solutions for practices. Permitting clinicians to individually elect the frequency, scheduling and location of their use of telehealth services may be other ways for organizations to promote work-life balance while demonstrating to clinicians the administration’s flexibility and prioritization of their needs [45].

Among nurses and nurse practitioners, jobs that do not allow one to practice to the full scope of one’s license experience lower satisfaction and higher job turnover [14, 46]. Similarly, our study finds that anticipated turnover was greater for behavioral health clinicians who felt their jobs did not allow them to provide all of the services they wanted, although fewer than one in six respondents reported such limitations to their work. To address this issue, employers can include clinicians in discussions when new services and care modalities are considered. Employers can also update job roles in states where behavioral health clinicians’ independent practice and prescription authorities have been expanded, and where insurers broadly cover behavioral health services and allow independent billing [47].

We were initially surprised that finding meaning in one’s work was statistically unrelated to the anticipated retention of these behavioral health clinicians who chose careers caring for economically marginalized communities. But the reason for this, we believe, is that almost all (97.4%) of the behavioral health clinicians of this study found their work meaningful. Thus, nearly all jobs within mHPSA safety net practices offer this important positive work motivator to behavioral health clinicians, and the meaningfulness of work is thereby acting as a near-universal force maintaining retention in this setting.

The second aspect of this study’s behavioral health clinicians’ jobs and work found not to be independently related to anticipated retention was feeling personally supported by other staff of their practices. The importance to retention of feelings of connectedness, support and trust among coworkers has been repeatedly demonstrated in the nursing literature [48–50]. The behavioral health clinician literature describes the importance of well-functioning multi-disciplinary teams to patient care [51, 52]. More research is needed on the importance of the affective aspects of behavioral health team members’ interactions, including how these do or do not contribute to clinician satisfaction and retention, and if they do, how they can be fostered by organizations.

Limitations

Anticipated retention and anticipated turnover are strong but not perfect indicators of clinicians who actually will remain in or leave their practices, and factors correlated with these two outcomes can differ [18, 53]. This study did not have data to determine the contributions of behavioral health clinicians’ assessments of other aspects of their jobs and work to their anticipated retention, including client caseloads, organizational culture, team functioning, use of onboarding programs, and use of telemedicine [11, 35, 43].

A strength of this study’s mHPSA behavioral health clinician cohort is that it is large and broad in its disciplines and in the types of practices and states where they work. Nevertheless, the cohort may not proportionately represent all clinicians of its disciplines within mHPSAs throughout the U.S. We note that during the study period the PRISM Collaborative included no states in the Deep South region of the U.S. Further, this study did not assess if job and work retention factors differ across the four behavioral health disciplines or for clinicians working in rural versus urban communities. This study also did not assess job and work factors that affect the retention for behavioral health disciplines trained at the bachelor’s and certificate levels.

Conclusion

This study found that behavioral health clinicians’ assessments of their work and jobs within mHPSA practices are strongly associated with the likelihood they expect to still work in their practices in five years. Anticipated retention was more likely with favorable assessments of their work and jobs overall, and then specifically with perceiving they had an effective and supportive administration, had a good work-life balance, felt well and fairly compensated, and could provide the full range of services they wished. An important area of future inquiry for the behavioral health field is to identify the administrator skills, management styles, and ways of shaping the work, schedules, and jobs that best support behavioral health clinicians and enhance their retention.

Nearly all behavioral health clinicians felt their practice’s mission was concordant with their own values and that they were doing important work; hence, finding meaning in one’s work was already a force nearly universally favoring retention within mHPSAs. Future studies, however, should assess how finding meaning in one’s work affects the retention of behavioral health clinicians working in other settings, where clinicians’ values and their practices’ missions may conflict.

Demand for behavioral health services has grown as population rates of depression and anxiety have increased since the COVID- 19 pandemic [54]. Higher rates of substance use have further increased service needs and thereby the demand and competition for clinicians in many practices offering behavioral health services in the U.S. These trends, along with chronic underfunding of public mental health services, challenge safety net practices’ capacity to compete for and then retain clinicians with competitive salaries, manageable client panel sizes, a tolerable pace of work, and sustainable work-life balance. Nevertheless, our study identifies changes in administrative approaches that mHPSA safety net practices can make to potentially stem the high turnover—and perhaps also bolster recruitment—of behavioral health clinicians, at little added cost. This is by creating an administration that values clinicians, seeks their input in decisions, and strives to meet their job and professional needs. This approach recognizes that a health care organization’s successes for its patients are greater when it succeeds with its clinicians. Co-prioritizing clinicians’ needs, listening, empowering, and administrative flexibility are key.

Abbreviations

LCSW

Licensed clinical social worker

LMFT

Licensed marriage and family therapist

LPC

Licensed professional counselor

mHPSA

Mental Health Professional Shortage Area

NHSC

National Health Service Corps

PRISM

Provider Retention and Information System Management Collaborative

Authors’ contributions

DEP initiated the study, conducted its analyses, and led manuscript writing. All authors (DEP, LdSZ, TRK, ABS, JNH, JF, BML) helped refine the study foci, contributed to interpretation of the findings, and reviewed and contributed to sequential drafts of the manuscript, including the final draft. JF, ABS and JNH further contributed to data collection, and TRK, LdSZ, and BML contributed to analytic strategies.

Funding

This project is supported by the Substance Abuse and Mental Health Services Administration (SAMHSA) and the Health Resources and Services Administration (HRSA) of the U.S. Department of Health and Human Services (HHS) under grant number U81HP46529 - 01‐01 Cooperative Agreement for a Regional Center for Health Workforce Studies for $1,121,875. This information or content and conclusions are those of the author and should not be construed as the official position or policy of, nor should any endorsements be inferred by SAMHSA, HRSA, HHS or the U.S. Government.

Data availability

The datasets analyzed in the current study are not publicly available, as they are administrative data that include personal and sensitive survey data, but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Studies using these data gathered principally for administrative purposes have been deemed exempt from human subjects review by the University of North Carolina at Chapel Hill Office of Human Research Ethics (Study 12–0626). As such, informed consent from clinicians is not required. This study was conducted in accordance with both the U.S. Federal Policy for the Protection of Human Subjects and the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.