Extent and predictors of work-related distress in community correction officers: a systematic review

Abstract

Research indicates that indirect exposure to trauma can have a detrimental psychological impact on professionals working within, and interfacing with, law enforcement and the criminal justice system. This systematic review aimed to explore the extent and predictors of work-related distress amongst community corrections personnel. A search of five databases identified 19 papers eligible for inclusion; 16 addressed burnout, and the remainder investigated secondary trauma, vicarious trauma and compassion fatigue. Synthesis revealed that community corrections personnel reported burnout at levels akin to those of other professions working in forensic contexts, though reports of secondary trauma appeared higher. Predictive factors encompassed personal, role-based and organisational factors. Research reporting work-related distress in correctional officers is focused on burnout but uses divergent models of stress, reveals methodological weaknesses, and to date has little examined responses to indirect trauma. The limitations of this review are discussed, alongside clinical implications and areas for future research.

Key words:

• burnout
• community correction
• literature review
• occupational distress
• parole
• probation
• secondary trauma
• vicarious trauma
• work distress

Community corrections personnel

Community correction personnel, encompassing both parole and probation officers, primarily supervise offenders within the community. Parole officers’ duties include supervision of offenders either after completing a prison sentence or once released on license. Probation officers, by contrast, largely work with individuals who have been sentenced to non-custodial sanctions within the community or who are awaiting sentencing. These distinct roles can vary across locations and often overlap dependent on policies and provisions applied in differing global contexts and legislatures.

Common to both roles are the supervision and management of offenders within the community. This requires ensuring that offenders fulfil the conditions of their sanction and rehabilitative activity, supporting them to adjust to community life and social and behavioural changes (Hsieh et al., 2015). This latter role has parallels with social work (Ohlin et al., 1956; Raynor & Vanstone, 2016), reliant on officers building a therapeutic and facilitative relationship with their clients (J. Miller, 2015; Spiess & Johnson, 1980).

In such multi-faceted roles, exposure to emotionally charged materials is inevitable; assessments of offenders’ criminal and social backgrounds necessitate reading of police reports and victim "statements. Given extensive trauma in offenders’ histories (e.g. Wilson et al., 2013), and its association with criminal behaviour (e.g. Kirk & Hardy, 2014), community correction officers (CCOs) may regularly be exposed to details of traumatic and violent crimes committed either by or towards their client. Home visits and establishing relationships with family members of the offenders expose CCOs to the wider impact of offenders’ crimes (diZerega & Verdone, 2011; Lewis et al., 2013), as well as intergenerationally transmitted trauma on offenders themselves (Halsey, 2018; Will et al., 2016). Such consistent indirect contact with trauma may render CCOs vulnerable to experiencing work-related distress (WRD), impeding their ability to carry out their roles effectively.

Work-related distress

Substantial literature over the last three decades has revealed that trauma’s impact extends beyond direct exposure (e.g. Beck, 2011; Greinacher et al., 2019). Burgeoning numbers of studies highlight that indirect experiences of trauma can also have adverse effects, suggesting that those in contact with a traumatised individual are at risk of developing significant emotional and psychological difficulties (Cocker & Joss, 2016; Elwood et al., 2011; Sinclair et al., 2017).

The impact of indirect exposure to trauma on professionals has been operationalised through a number of concepts, the most notable of which are described here. ‘Burnout’ communicates the physical, emotional and psychological impacts of chronic exposure to work stress from working with others during intensely emotive situations (e.g. Pines & Aronson, 1988). Symptoms include emotional exhaustion, depersonalisation, and reduced perceived personal accomplishment (Maslach & Jackson, 1981). Subsequent conceptions of WRD include ‘secondary traumatic stress’ (STS), a trauma response arising through engagement with another’s trauma and suffering (Figley, 1995). The symptoms echo those of post-traumatic stress disorder (Figley, 2002), including hypervigilance, flashbacks and nightmares (American Psychiatric Association, 2013). Similar to STS, the term ‘vicarious trauma’ (VT) describes a response to prolonged empathic engagement with a trauma survivor, altering cognitive schemas regarding the self, others and the world (Pearlman & Mac Ian, 1995). Finally, ‘compassion fatigue’ (CF), often used interchangeably with STS and VT, encapsulates emotional and physical fatigue as a consequence of chronic use of empathy when working with trauma survivors (Figley, 2002).

Whilst the proliferation of concepts describing and explaining WRD has led to increased awareness of the impact of indirect exposure to trauma, they are critiqued as often applied interchangeably, creating confusion about their distinctive character and their precise phenomenology. CF and STS have been viewed as substitutable, yet other research describes CF as comprising characteristics of both STS and burnout (Adams et al., 2008). Construct validity has also been criticised (Sabin-Farrell & Turpin, 2003; Sinclair et al., 2017), with some researchers suggesting that CF is ill-defined and ambiguous (Fernando & Consedine, 2014; Ledoux, 2015). Given the range in terminology used within the literature, and the absence of a broad-ranging review of correctional officer stress, the current review will adopt the use of ‘work-related distress’ (WRD) to encompass all forms of consequences of exposure to another’s traumatic material at work, such as CF, STS, burnout and VT, to ensure a comprehensive analysis of the literature. Use of WRD also moves away from implying that the constructs are unique risks to specific occupations, such as the ‘caring’ professions (Newell & MacNeil, 2010).

Evidence of work-related distress

Regardless of the terminology used, professionals working with trauma survivors are susceptible to developing WRD, particularly in the caring professions across diverse healthcare settings (Huggard et al., 2017; Sorenson et al., 2016; Williams et al., 2019), first-responders (Greinacher et al., 2019), and emergency service personnel including firefighters (Jo et al., 2018), ambulance officers (Regehr et al., 2002) and police officers (Sherwood et al., 2019). Moreover, the prevalence of WRD has been evidenced by non-clinical staff groups within the criminal justice system, notably judges (Lustig et al., 2008), lawyers (Levin et al., 2011) and jurors (Lonergan et al., 2016). Given CCOs’ exposure to offenders, victims of crime and a context of trauma, it seems timely to examine how they might be affected.

Through CCOs’ exposure to offenders, victims of crime and a context of trauma, any WRD that CCOs experience has significant implications for the criminal justice system. For instance, WRD is associated with increased turnover and staff sickness (Kim et al., 2020); recent turnover rates amongst UK probation officers were at 6.3% (Ministry of Justice, MOJ, 2019b), an increase of around 50% since 2016, which has the potential to compromise recruitment and retention. Increased sickness absence at 10.7 days is 67% higher than the UK average (Murphy, 2020), conferring significant societal impacts. If remaining staff have higher caseloads, increased care burden and reduced time for client supervision, their own compromised effectiveness may undermine offender rehabilitation and reduce public safety. It therefore seems key to systematically examine WRD amongst this staff group, understand its extent and determinants, and consider how best to address distress.

Predictor variables

As extent of WRD for professionals’ exposure to traumatogenic material becomes more evident, factors that might increase risk of distress have come under greater scrutiny, notably previous and unresolved trauma history, empathic engagement with clients and client narratives focused on childhood trauma (Figley, 1995). Additionally, six key organisational domains have also been advanced: control; rewards; workload; community; fairness; and values (Leiter & Maslach, 2004), as well as managerial responses to trauma and social support predictive of WRD in mental health professionals (Turgoose & Maddox, 2017), midwives (Suleiman-Martos et al., 2020), first-responders (Greinacher et al., 2019) and lawyers (Maguire & Byrne, 2017).

Previous literature reviews

Despite community corrections officers arguably engaging in more frequent and sustained contact with clients than other professionals within the criminal justice system (Slate et al., 2003), there appears no published systematic reviews examining WRD amongst this population. A previous narrative review examining generic stress amongst probation officers described workplace stress as associated with personal (including financial concerns and family matters) and organisational factors (including unnecessary paperwork and lack of time; Slate et al., 2000). However, the review neither adequately defined WRD nor was systematic in approach (offering no clarity regarding search process, inclusion criteria or quality appraisal).

Two reviews have more systematically examined correlates of stress and burnout in correctional officers, but only within closed correctional settings. Schaufeli and Peeters (2000) reported elevated levels of stress and burnout, somewhat higher than other professional groups, with high workload, lack of role autonomy and role ambiguity predicting stress. Similarly, Finney et al. (2013) reviewed the relationship between organisational stressors, stress and burnout amongst correctional officers working in adult facilities, finding that an absence of administrative and organisational support was correlated with higher levels of stress and burnout. Neither review examined work-related responses to client trauma, highlighting that a more recent and wider review of impacts within community settings is needed.

The current review

Given the potential vulnerabilities of this key professional group and absence of any previous systematic appraisal and synthesis, the current review aims to systematically review existing published literature examining the extent of WRD in CCOs, as well as predictive psychosocial factors. In doing so, this review seeks to increase understanding of the factors that might potentiate vulnerability, to both screen for and determine potential preventative or supportive strategies to mitigate distress at work. Given the relative infancy of this research field and an initial scoping review that suggested a circumscribed literature base, the review aims to explore the following questions:

1. What is the extent of WRD in community corrections personnel?

2. What psychosocial factors appear to predict WRD?

Method

Protocol

The present systematic review was undertaken following guidance by the Centre for Reviews and Dissemination (CRD, 2009) for conducting reviews in health care and the PRISMA framework for improving transparency in systematic reviews (Moher et al., 2009). PROSPERO was consulted to avoid duplication, with no salient reviews recorded, though the present review was not registered. The search strategy was developed using the PICOS statement (Santos et al., 2007), as presented in Table 1.

Table 1. PICOS statement.

PICOS	Category	Application as defined in the present review
P	Population	Corrections officers working within the community
I	Intervention/Exposure	Indirect exposure to, and effects of, traumatogenic material
C	Comparison	None
O	Outcome	The extent and predictive factors of WRD.
S	Study Design	Quantitative studies using validated measures of WRD

Note: WRD = work-related distress.

Search strategy

An initial scoping search was undertaken on Google Scholar, with a senior dedicated University librarian, to determine relevant literature reviews and provide an overview of the quantity of existing literature. This supported the focus of the current review and enabled identification of search terms typically used in the research area. On 2 September 2019, the following electronic bibliographic databases were systematically searched using both University of Leicester and NHS Athens access: PsycINFO, Scopus, Medline, Web of Science, PILOTS and Lexis. The databases were selected as they cover a broad range of research from multidisciplinary professions, including psychology, medicine and law. The following search terms were employed in each database: (‘burnout’ OR ‘burn-out’ OR ‘second* trauma*’ OR ‘compassion fatigue’ OR ‘vicarious trauma*’ OR ‘PTSD’ OR ‘post-trauma*’ OR ‘trauma’ OR ‘distress’) AND (‘parole officer*’ OR ‘probation* officer*’ OR ‘community correction*’). A title, abstract and keyword search was undertaken. Filters were applied to limit included papers to those published in the English language and peer-reviewed journals.

Selection criteria of studies

Following the database searches, the remaining articles were transferred to RefWorks, and duplicates were removed. The titles and abstracts of the remaining articles were then examined for relevance by the first author, guided by these eligibility criteria:

Inclusion criteria

• A reported measure of WRD.

• Data concerning the extent of WRD, and similar terminology, amongst CCOs.

• Data concerning factors associated with or predicting WRD, and similar terminology, amongst CCOs.

• Quantitative articles and mixed methods papers with clearly distinct quantitative data, to allow for the identification of data relating to the extent and predictive factors of WRD.

• English language publication.

• Peer-reviewed published articles.

Exclusion criteria

• Qualitative studies.

• Articles investigating primary forms of trauma, including post-traumatic stress disorder (PTSD), following scanning of full papers for relevance. Primary traumatisation occurs through direct first-hand experience of trauma, such as threat to life or directly witnessing threat to a loved one’s life (Branson, 2019). This review specifically examines the effects of indirect, rather than overt, exposure to trauma through others (WRD), which is not anticipated in the job role and is therefore inadvertent and potentially unacknowledged.

• Articles investigating occupational stress, following scanning of full papers for relevance. Occupational stress is defined as a change in an individual’s physiological or psychological state in response to work-related factors that can result in negative consequences, though not necessarily (Richardson & Rothstein, 2008). It is not specific to work-related factors concerning the emotional toll of working with people or being exposed to indirect trauma.

• Articles in which data for community corrections officers are not distinctly recorded.

• Theses, dissertations, conference papers or editorial papers.

• Intervention studies in which no relevant pre-intervention data are presented.

• Articles not using validated measures of WRD and similar terminology.

Articles clearly meeting the inclusion criteria were retrieved and read in full, as were papers whose immediate focus was not transparent. Reference mining of the included articles was also utilised to identify further suitable articles. Figure 1 displays the PRISMA flow diagram illustrating the selection process.

Figure 1. PRISMA flow diagram.

Data extraction and synthesis

The authors created a data extraction tool adapted from previous reviews on the prevalence and predictors of WRD in other professional groups (Brooks et al., 2016; Sage et al., 2018). Shortlisted articles were read in full by the first author, and data were then extracted and tabulated using the data extraction tool, encompassing the following variables where reported: publication year, country of study, study design, the type of WRD investigated, sample size, demographics, measures used to assess WRD, relevant statistics (such as reliability coefficients and measures of association) and statistical analysis method. Key information from each article’s main aims, findings and conclusions was also extracted where relevant to the review question.

Assessment of methodological quality of included studies

All shortlisted articles employed a cross-sectional design. Whilst there are a range of validated tools for assessing the quality of other study designs, such as randomised-controlled trials, observational and case-control studies, the number of tools for appraising cross-sectional studies is limited (Sanderson et al., 2007; Zeng et al., 2015). Those employed to assess the quality of cross-sectional studies, such as ROBINS-I and STROBE, have questionable validity and appropriateness, particularly with regard to their specificity, applicability and generalisability (Downes et al., 2016). Thus, the Appraisal Tool for Cross-Sectional Studies (AXIS; Downes et al., 2016) was selected to appraise the quality of all included articles. The quality appraisal was used to provide further insight regarding the research rather than to exclude articles. Four studies were randomly selected and independently assessed by the second author to compare for consistency and ensure consensus. There was agreement in ratings on all four papers.

Appraisal tool for cross-sectional studies

The AXIS is a twenty-item critical appraisal tool, developed in light of an increasing evidence-base employing cross-sectional designs and a flourishing number of systematic literature reviews utilising this study design (Downes et al., 2016). Created using Delphi methodology and following recommended reporting guidelines and assessment tools, AXIS pays particular attention to the methodology and results reported within a study. This provides a comprehensive and applicable tool to gauge the quality of reporting within studies, as well as the design quality and potential bias risk. Its questions cover three key aspects of research (Kiss et al., 2018), focusing on the design quality, reporting quality and introduction of bias.

The AXIS does not provide a numerical scale to assess overall quality of papers (Downes et al., 2016), and such scales have been critiqued for their lack of uniformity and consistency when weighting papers (Greenland & O’Rourke, 2001; Jüni et al., 1999). Each question in the tool was assessed individually as being met (‘yes’) or not being met (‘no’ or ‘not known’), similar to other studies (Kiss et al., 2018; Schoth et al., 2019); Questions 13 and 19 were exceptions to this rule, and reverse scoring was utilised, with a ‘no’ meeting the criteria for quality assessment. Comments were also provided when context was needed. Quality was determined based on the total number of criteria met (total = 20, ≥16 = high quality), in accordance with other studies employing AXIS (Henderson et al., 2019; Wong et al., 2018).

Results

Overview

Following the selection process, 19 papers were elicited that investigated WRD in CCOs. Of these articles, 12 included data regarding the extent of WRD, and 15 examined predictive variables. However, different definitions of WRD were investigated across the papers, as well as different measures and statistical analyses, precluding a meta-analysis of the data. The statistical analysis undertaken was explicitly stated in all but two studies (Lindquist & Whitehead, 1986; Whitehead, 1985). Statistical analyses varied across the articles, dependent on their aims. Table 2 provides a summary of the shortlisted study characteristics.

Table 2. Study characteristics.

Author (year), country	Aim	Conception of WRD and measurement	Participants	Extent	Predictors
Allard et al. (2003) Australia	Explores relationship between inter-role conflict, intra-role conflict, role ambiguity and role preferences with burnout.	Burnout MBI	55 CCOs 69.1% female Mage = 37.82 years	Both intra and inter-role conflict significantly independently related to level of EE (p < .05)
Anson and Bloom (1988) USA	Compares occupational stress amongst various professions, including police and correctional officers.	Burnout MBI	43 POs 46% female Mage = 34.6 years	NS difference between police and POs on EE and depersonalisation NS. M score for POs: EE = 22.0 (SD = 11.0), depersonalisation = 12.0 (SD = 6.7) and personal accomplishment = 30.0 (SD = 7.0).
Dir et al. (2019) USA	Examines relationship between burnout, participatory atmosphere and mental health stigma.	Burnout MBI–GS	245 JPOs 67% female Age range = 20–69 (M unreported)	Most JPOs reported mild to moderate levels of burnout.
Gayman and Bradley (2013) USA	Examines the association between depressive symptoms, work stress and work environment (including EE).	EE CSS	826 PPOs 54% female M age = 39.24 years (SD = 9.22)	Officers reported high levels of EE (M = 2.83).	White and female officers reported higher levels of burnout. Job tenure also a predictor: longer employment associated with increased burnout. Increased work stress, role conflict and role overload all predict burnout.
Gayman et al. (2018) USA	Examines the association between caseload with depressive symptoms and EE.	EE CSS	798 PPOs 54.49% female Mage = 39.16 (SD = 9.22)	M score for EE = 2.83, representing ‘agree’ that they felt EE with work environment.	PPOs depressive symptoms, number of people on a caseload with a mental health problem, caseload size and years in current job all positively correlated with EE. Percentage of caseload receiving mental health services support associated with less EE. Perception of training adequacy to supervise people with mental health problems negatively associated with EE.
Holgate and Clegg (1991) Australia	Compares burnout between an older and younger sample of POs. Compares predictive effects of organisational and personality variables on burnout.	Burnout MBI	106 CCOs 61 females (Mage = 34 years), 45 males (Mage = 39 years)		NS difference in burnout scores between older and younger groups. For both groups: depersonalisation, role conflict, role ambiguity and emotionality all correlated positively with EE. Emotionality and role conflict had a significant direct effect upon EE.
Holloway et al. (2019) USA	Examines demographics, attitudes about participation in the workplace and burnout	Burnout MBI–GS	219 JPOs 67% female Majority 30–49 years old (M unreported)	M levels of burnout: Cynicism = 0.77 (SD = 1.02), EE = 1.23 (SD = 1.15) and professional efficacy = 3.06 (SD = 0.93).
Jin et al. (2018) China	Assesses the impact of positive (job autonomy, procedural justice and role clarity) and negative (role conflict, job stress, job dangerousness) job characteristics on burnout	Burnout MBI	225 CCOs 34% female Mage ∼ 39 years		Greater role clarity linked to reduced burnout. Greater role conflict, job stress and job dangerousness were more likely to generate greater burnout. Males more likely to report higher degree of burnout.
Lewis et al. (2013) USA	Measures the effect of negative caseload events on traumatic stress responses	CF/burnout CFST	309 POs, supervisors and administrators 159 female, 127 male, 23 gender not indicated		Supervising those who report violent recidivism involving a child victim or sexual re-offence while on caseload have significantly higher levels of CF and burnout. Those reporting offenders’ threat to self or family, assault on duty or client suicide linked to higher CF and burnout.
Lindquist and Whitehead (1986) USA	Compares burnout amongst Supervised intensive restitution officers, institutional corrections officers and PPOs	Burnout MBI	108 PPOs 20.4% female, Mage = 40.1 years	No significant difference in burnout between groups of workers.	Significant predictors: lower levels of support, greater weekly hours of contact and greater correctional seniority predict greater EE. Younger officers and those who reported greater correctional seniority, greater role conflict and lower levels of social support reported greater depersonalisation.
Merhav et al. (2018) Israel	Explores correlation of attachment styles with disruption in cognitive schemas.	VT TABS	189 adult POs 78% female Mage = 38 years		Secure attachment styles reported significantly fewer disruptions in cognitive schemas of trust than other attachment styles. Secure and dismissive-avoidant attachment styles reported significantly fewer disruptions in cognitive schema of safety than those with preoccupied and fearful avoidance attachment types. Human-induced personal trauma history significantly predicted disruptions in trust schema when controlling for attachment.
Rhineberger- Dunn et al. (2016) USA	Explores predictive factors for ST among PPOs and ROs.	ST STSS	179 PPOs: 47% female Mage = 43.86 years 98 ROs: 56% female Mage = 41.11 years	PPOs reported significantly greater number of secondary trauma symptoms compared to ROs, controlling for the other variables. Secondary trauma M for PPOs = 18.46 (SD = 13.64) For ROs = 14.19 (SD = 12.37)	Better health predicted lower ST symptoms. Reporting adequate training for job associated with lower levels of ST. Greater contact hours associated with increase in ST.
Rhineberger- Dunn et al. (2017) USA	Compares association between background and workplace predictive factors for burnout between PPOs and ROs.	Burnout MBI	179 PPOs: 47% female Mage = 43.86 years 98 ROs: 56% female Mage = 41.11 years	PPOs > ROs were more likely to report symptoms of EE and depersonalisation. NS difference in personal accomplishment. For PPOs: M scores EE = 13.61 (SD = 8.12), depersonalisation = 13.05 (SD = 7.71), personal accomplishment = 28.16 (SD = 5.74) For ROs: M scores EE = 11.09 (SD = 7.84), depersonalisation = 10.88 (SD = 8.36), personal accomplishment = 28.55 (SD = 5.4)	Women and those reporting poorer health reported higher EE. Perceptions of adequate education and job training significantly associated with lower EE. Health, educational training, job training and pay dissatisfaction all significant determinants of depersonalisation. Greater years in the field associated with lower personal accomplishment.
White et al. (2015) USA	Explores prevalence and predictive variables of burnout.	Burnout MBI–GS	245 JPOs 63.7% female Age not reported	EE: 25.5% participant scores in the low range, 35.8% moderate, 31.7% high Cynicism: 36.2% in the low range, 28.9% moderate, 27.6% high Professional efficacy: 51.2% in the low range, 26% moderate, 15.4% high	Job satisfaction was the strongest predictor of burnout: higher job satisfaction predicting lower emotional exhaustion, cynicism and higher professional efficacy. Race: Caucasians reported higher EE and cynicism. Those with high-risk clients on caseload reported higher cynicism.
Whitehead (1985) USA	Compares extent of burnout among PPOs with human service workers sample.	Burnout MBI	968 PPOs 37.4% female Mage = 35.7 years	PPOs reported significantly lower EE, though higher intensity, than human service workers sample. PPOs had higher depersonalisation across frequency and intensity. Personal accomplishment was lower in both frequency and intensity.	Curvilinear relationship between seniority and burnout. More experienced workers reported highest burnout scores, and most experienced workers reported levels of burnout similar to those just starting the job.
Whitehead (1986a) USA	Examines gender differences in job burnout, job satisfaction and role conflict	Burnout MBI	711 POs 41% female Mage = 33.9 years		Gender a significant predictor for depersonalisation. Weekly hours of offender contact predicted EE. Caseload size and age predicted depersonalisation. Job satisfaction and role conflict predicted EE and depersonalisation.
Whitehead (1986b) USA	Examines extent of job burnout and job dissatisfaction among probation managerial employees	Burnout MBI	184 PO managers: 44% supervisors, 56% upper-level administrators including probation directors, chief POs or assistant chiefs 25% female Mage = 41.7 years	Reported experiencing feelings of EE about once a month, feelings of depersonalisation less frequently than once a month and feelings of personal accomplishment about once a week. NS differences between administrators and supervisors. M: scores EE = 2.08, depersonalisation = 1.58, personal accomplishment = 4.06.
Whitehead (1987) USA	Tests two leading theories of job burnout – Maslach’s theory emphasising client contact as central cause of burnout, and Cherniss’ theory highlighting the relationship between organisational factors and burnout	Burnout MBI	387 POs Predominantly middle-aged (M not reported)		Role conflict, job satisfaction and age all had significant direct effects on EE and depersonalisation. Job satisfaction and weekly hours of client contact had significant direct effects on personal accomplishment.
Whitehead and Lindquist (1985) USA	Assesses perceptions of burnout	Burnout MBI	108 PPOs Mage = 40.1 years 20% female	21% PPOs reported feeling emotionally exhausted at least once a week or more. 8% reported depersonalisation at least once a week.	Social support and correctional seniority were significant predictors of EE. Significant predictors for depersonalisation were social support, role conflict, age and correctional seniority. Younger officers and those reporting lower levels of social support, greater role conflict and greater seniority reported greater depersonalisation.

Note: CCOs = community correction officers; CF = compassion fatigue; CFST = Compassion Satisfaction/Fatigue Self-Test for Helpers; CSS = Children’s Services Survey; EE = emotional exhaustion; JPO = juvenile probation officer; M = mean; MBI = Maslach Burnout Inventory; MBI–GS = MBI–General Survey; NS = non-significant; PO = probation officer; PPO = probation/parole officer; RO = residential officer; ST = secondary trauma; STSS = Secondary Traumatic Stress Scale; TABS = Trauma and Attachment Belief Scale; VT = vicarious trauma; WRD = work-related distress.

Selected studies characteristics

Several different terms were utilised to assess evidence of WRD amongst CCOs: burnout; emotional exhaustion; secondary trauma; VT; and CF. The majority of studies, 14 in total, operationalised WRD using the term burnout, whilst one study interchangeably used burnout and CF (Lewis et al., 2013). Two studies examined one specific dimension of burnout, emotional exhaustion (Gayman & Bradley, 2013; Gayman et al., 2018), whereas the remaining two studies investigated secondary trauma (Rhineberger-Dunn et al., 2016) and VT (Merhav et al., 2018).

All studies employed purposive sampling and were recruited through various community correctional agencies and organisations, though none of the articles explicitly stated the sampling method used. Studies were conducted across the USA, Israel, China and Australia. Sample size considerably varied across the studies, ranging from 43 (Anson & Bloom, 1988) to 968 (Whitehead, 1985). Although not explicitly stated, two studies appeared to investigate different data collected from the same sample (Rhineberger-Dunn et al., 2016, 2017), whilst another two studies explicitly stated that the same survey data were utilised (Gayman & Bradley, 2013; Gayman et al., 2018). Therefore although 19 articles were included, the present review is founded on 17 discrete samples. None of the studies described an a priori power calculation, tempering definitive conclusions regarding the generalisability of the data.

Given that studies were conducted across different countries and states, the type of community corrections role investigated also varied: five exclusively examined probation officers; five investigated a mix of probation and parole officers; two studied probation, parole officers and residential officers; three examined juvenile probation officers; one explored probation officer managers; and three referred to participants broadly as CCOs. Furthermore, three studies compared professions within the sample: between probation and parole officers; a sample of institutional correction officers and officers working in a specific ‘Supervised Intensive Restitution’ programme (Lindquist & Whitehead, 1986); between police officers and probation officers (Anson & Bloom, 1988); and between probation officers and a sample group of ‘human service workers’ (Whitehead, 1985).

Survey data collection methods were utilised by all the studies: online, face-to-face or through postal communication. All studies reported response rates, varying from 45.04% (Rhineberger-Dunn et al., 2016, 2017) to 96% (Jin et al., 2018). Most studies (N = 12) included a greater number of females, consonant with probation services proportionately having more female staff (Ministry of Justice, 2019a), though seven studies reported more males within the sample.

All studies reported the measures used to gather data. In line with the present review’s aims, specific attention is paid to measures that assess WRD. All studies utilised self-report questionnaire measures, albeit diverse, partially as a result of the varied terms employed to define WRD. For burnout, the Maslach Burnout Inventory (MBI; Maslach & Jackson, 1981) was most commonly adopted, assessing both the frequency and intensity of three aspects of burnout: emotional exhaustion; depersonalisation; and personal accomplishment. The MBI has excellent internal consistency with Cronbach’s alpha of .90 (Maslach & Jackson, 1981). Two studies (Dir et al., 2019; White et al., 2015) utilised a variation of the original MBI, the MBI–General Survey (MBI–GS; Schaufeli et al., 1996), designed for use with occupational groups other than human services. This version of the inventory assesses emotional exhaustion, cynicism and professional efficacy; research has demonstrated that the inventory has satisfactory internal consistencies (Leiter & Schaufeli, 1996). Contrastingly, Lewis et al. (2013) assessed both compassion fatigue and burnout by employing the Compassion Satisfaction/Fatigue Self-Test for Helpers (CFST; Figley & Stamm, 1996), used to measure compassion satisfaction, compassion fatigue and burnout.

Two studies (Gayman & Bradley, 2013; Gayman et al., 2018) investigating only one dimension of burnout, emotional exhaustion, utilised measures adapted from scales in the Children’s Services Survey (Glisson & James, 2002), which has good reliability (Glisson & Durick, 1988; Glisson & Hemmelgarn, 1998). Merhav et al. (2018) assessed VT using the Trauma and Attachment Belief Scale (TABS; Pearlman, 2003), demonstrated to have good validity and good internal consistency for the total scale (α = .96), though this has ranged across the subscales (α = .67 to .87; Pearlman, 2003). The only study assessing secondary trauma (Rhineberger-Dunn et al., 2016) utilised the Secondary Traumatic Stress Scale (STSS; Bride et al., 2004), demonstrated to have excellent internal consistency with Cronbach’s alpha of .93 (Bride et al., 2004).

Methodological quality of included studies

Across the included studies, quality ranged from 11 to 19 criteria met, with seven papers meeting the high-quality threshold (≥16 criteria met). See Table 3 for a summary of the quality appraisal of included studies.

Table 3. Quality appraisal using AXIS checklist.

Name of first author (year)	Q1	Q2	Q3	Q4	Q5	Q6	Q7	Q8	Q9	Q10	Q11	Q12	Q13	Q14	Q15	Q16	Q17	Q18	Q19	Q20	Total (max 20)
	RQ	SD	SD	RQ	SD	RB	RB	SD	RB	RQ	RQ	RQ	RB	RB	RB	RQ	SD	RQ	SD	SD
Allard et al. (2003)	✓	✓	✓	✓	✓	✓	X	✓	✓	✓	✓	✓	✓	X	✓	X	✓	X	✓	✓	16
Anson and Bloom (1988)	X	✓	X	✓	✓	✓	X	✓	✓	✓	✓	✓	✓	X	✓	✓	✓	X	X	X	13
Dir et al. (2019)	✓	✓	X	✓	✓	✓	X	✓	✓	✓	✓	✓	✓	X	X	✓	✓	✓	✓	✓	16
Gayman and Bradley (2013)	✓	✓	X	✓	✓	✓	X	✓	✓	✓	✓	✓	✓	X	✓	✓	✓	X	✓	✓	16
Gayman et al. (2018)	✓	✓	X	✓	✓	✓	X	✓	✓	✓	✓	✓	X	X	✓	✓	✓	✓	X	✓	15
Holgate and Clegg (1991)	✓	✓	X	✓	✓	✓	X	✓	✓	X	✓	✓	−	X	✓	X	✓	✓	✓	−	13
Holloway (2019)	✓	✓	X	✓	✓	✓	X	✓	✓	✓	✓	✓	✓	X	✓	✓	✓	✓	X	✓	16
Jin et al. (2018)	✓	✓	✓	✓	✓	✓	✓	−	X	✓	✓	✓	✓	X	✓	✓	✓	✓	✓	✓	17
Lewis et al. (2013)	✓	✓	X	✓	✓	✓	X	✓	✓	✓	✓	✓	X	X	✓	✓	✓	✓	X	✓	15
Lindquist and Whitehead (1986)	✓	✓	X	✓	✓	✓	X	✓	✓	✓	X	✓	✓	X	X	✓	X	X	X	X	11
Merhav et al. (2018)	✓	✓	−	✓	✓	✓	X	✓	✓	✓	✓	✓	✓	X	X	✓	X	✓	✓	✓	15
Rhineberger-Dunn et al. (2016)	✓	✓	X	✓	✓	✓	X	✓	✓	✓	✓	✓	X	X	✓	✓	✓	✓	✓	✓	16
Rhineberger-Dunn et al. (2017)	✓	✓	X	✓	✓	✓	X	✓	✓	✓	✓	✓	✓	✓	X	✓	✓	✓	✓	X	16
White et al. (2015)	✓	✓	X	✓	✓	✓	X	✓	✓	✓	✓	✓	✓	X	X	✓	✓	✓	X	✓	15
Whitehead (1985)	✓	✓	X	✓	✓	✓	X	✓	✓	✓	X	✓	X	X	✓	X	✓	X	✓	X	12
Whitehead (1986a)	✓	✓	X	✓	✓	✓	X	✓	✓	✓	X	✓	X	X	✓	✓	✓	X	X	X	12
Whitehead (1986b)	✓	✓	X	✓	✓	✓	X	✓	✓	✓	X	✓	X	X	✓	✓	X	X	✓	−	12
Whitehead (1987)	✓	✓	X	✓	✓	✓	X	✓	✓	✓	✓	✓	✓	X	✓	✓	✓	X	X	✓	15
Whitehead and Lindquist (1985)	✓	✓	X	✓	✓	✓	X	✓	✓	✓	✓	✓	✓	X	✓	X	✓	X	✓	X	14

Note: Question grouping: RQ = reporting quality; SD = study design; RB = risk of bias (Kiss et al. 2018). Ratings: ✓= criteria met; – = don’t know/partially met; X = criteria not met.

Criteria related to reporting quality were most frequently met (range = 4–7, M = 6.1). All studies clearly reported the target population and adequately described the basic data. However, nine studies did not discuss their methodological limitations, making it difficult to establish the validity of the research.

Criteria regarding study design were met less frequently (range = 2–7, M = 5.2). In all studies, the cross-sectional design was appropriate for the stated aims, whilst the outcome variables and risk factors measured were appropriate to these aims. Furthermore, all studies’ samples were taken from an appropriate population base to ensure they were representative of the population of interest. Nevertheless, only three of the studies justified sample size, and none reported on power analysis. Information regarding ethical approval or funding sources was limited.

Across the studies, criteria relating to the risk of bias were rated less well (range = 3–5, M = 3.5). Only Jin et al. (2018) addressed the bias of potential non-responders within the sample, whilst Rhineberger-Dunn et al. (2017) was the only study to provide any information regarding non-responder characteristics. However, across all studies the samples appeared representative of the target population, and all measures used to assess outcome variables had been trialled, piloted or published previously.

Extent of WRD

Twelve studies included information concerning the levels of WRD amongst CCOs. Irrespective of the definition, all of these studies found evidence of WRD amongst CCOs, though limited detail was provided regarding clinical significance.

The three studies exploring burnout amongst juvenile probation officers reported similar findings. Dir et al. (2019), a high-quality study, reported mild to moderate levels of burnout amongst juvenile probation officers, whilst White et al. (2015) reported moderate levels. Holloway et al. (2019), another high-quality study, provided mean scores across the three burnout constructs that were almost identical to those reported by Dir et al. (2019), suggesting similar mild to moderate levels. Whitehead and Lindquist (1985), reporting on burnout, found that the majority of probation and parole officers reported emotional exhaustion and depersonalisation ‘once a month or less’, though 16% and 8%, respectively, reported experiencing them ‘once a week or more’, indicating high levels of burnout for some. In contrast, two studies (one of high quality) focusing specifically on extent of emotional exhaustion found that probation and parole officers reported high levels (Gayman & Bradley, 2013; Gayman et al., 2018).

In addition to reporting levels of WRD in CCOs, three studies also compared levels of WRD with those in other professions, utilising the MBI, though findings appeared equivocal. Anson and Bloom (1988) investigated the difference in mean scores across the subscales of the MBI between probation officers and police officers, finding no significant difference. Similarly, Lindquist and Whitehead (1986) looked at the frequency and intensity of each subscale of burnout and found no significant difference in levels of burnout between probation and parole officers compared with institutional correction officers and those working on a ‘Supervised Intensive Restitution’ programme. In contrast, Whitehead (1985) reported that probation officers experienced each aspect of burnout more frequently and intensely than a sample group of human service workers (p < .05), with the exception of frequency of emotional exhaustion where human service workers reported higher levels (p < .05). Of the three studies comparing professions, Whitehead (1985) was the only one to attempt to interpret their findings, concluding that probation officers were experiencing burnout on an infrequent and ‘less than intense’ level. Notably, sample demographics differed between the professional groups in these studies across gender, years of experience, age and education, and these potential confounds limit the conclusions that can be drawn.

Three studies also explored levels of WRD between different CCO roles, again with equivocal findings. Whitehead’s (1986b) study of probation managers found no significant difference in levels of burnout between administrators and supervisors. Two studies of high quality also compared levels of WRD between probation and parole officers, with residential officers. Rhineberger-Dunn et al. (2017) found that probation and parole officers were significantly more likely to report symptoms of emotional exhaustion (p < .05) and depersonalisation (p < .01) than residential officers, but with no difference in levels of personal accomplishment. Similarly, Rhineberger-Dunn et al. (2016) reported that probation and parole officers reported a significantly greater number of STS symptoms than residential officers (p < .05).

Predictive factors of WRD

Fifteen studies reported at least one predictive factor of WRD, with similar findings across many of the papers. Only one study (Whitehead, 1985) did not use regression analysis or path analysis to find predictors of WRD, and seven studies also analysed correlations to demonstrate relationships between WRD and other variables.

Eleven studies identified personal characteristics; gender was found to predict WRD in three high-quality studies, though they reported conflicting findings. Both Gayman and Bradley (2013) and Rhineberger-Dunn et al. (2017) found that female gender predicted higher levels of emotional exhaustion, whereas Jin et al. (2018) found that male gender predicted higher levels of burnout. Race was also reported as a predictor, with findings suggesting that Caucasians were likely to report higher levels of burnout than other ethnic groups (Gayman & Bradley, 2013; White et al., 2015). Age was another widely identified factor, with a younger age predictive of increased emotional exhaustion (Whitehead, 1987) and depersonalisation (Lindquist & Whitehead, 1986; Whitehead, 1987; Whitehead & Lindquist, 1985).

Respondents’ health also appeared predictive. Regarding mental health, higher reported current depression was significantly associated with higher levels of emotional exhaustion (Gayman & Bradley, 2013; Gayman et al., 2018), and lower physical health was a significant predictor of elevated burnout (Rhineberger-Dunn et al., 2017) and secondary trauma (Rhineberger-Dunn et al., 2016) in high-quality studies. Only one study reported on personality type (Holgate & Clegg, 1991), with higher levels of emotionality, linked to neuroticism, predicting higher levels of emotional exhaustion. Insecure attachment style and previous human-induced trauma history were also found to significantly predict higher levels of VT (Merhav et al., 2018).

Organisational predictive factors were identified in 10 studies. Greater number of hours of direct contact with clients and increase in length of time in role were significantly associated with higher levels of emotional exhaustion and secondary trauma (Gayman & Bradley, 2013; Gayman et al., 2018; Lindquist & Whitehead, 1986; Rhineberger-Dunn et al., 2016; Whitehead, 1985, 1986a). Yet two studies reported that greater client contact had no relationship with emotional exhaustion but was associated with sense of personal accomplishment; Whitehead (1987) found that an increase in tenure was associated with higher feelings of personal accomplishment and lower levels of burnout, whereas Rhineberger-Dunn et al. (2017), a high-quality study, found that greater years in the field significantly predicted a lower sense of personal accomplishment. A further predictor, lesser job satisfaction, predicted increased burnout (White et al., 2015; Whitehead, 1986a, 1987).

Caseload size was explored by two studies (Gayman et al., 2018; Whitehead, 1986a), suggesting that higher caseloads were predictive of elevated burnout. Additionally, Gayman et al. (2018) found that caseloads with increased numbers of those with mental health difficulties predicted higher levels of emotional exhaustion amongst probation and parole officers; yet when clients received mental health service support, staff reported decreased emotional exhaustion. Lewis et al. (2013) investigated the relationship between WRD and client recidivism in caseloads, noting that those who supervised individuals who re-offended against children or committed sexual offences experienced significantly higher levels of compassion fatigue and burnout. Moreover, CCOs who received threats of harm were also likely to report significantly higher levels of fatigue and burnout.

Eight studies also evidenced role-based predictive factors, with all identifying role conflict (Allard et al., 2003; Gayman & Bradley, 2013; Holgate & Clegg, 1991; Jin et al., 2018; Lindquist & Whitehead, 1986; Whitehead, 1986a, 1987; Whitehead & Lindquist, 1985), and those experiencing higher levels of role conflict reporting higher levels of burnout. Examination of role ambiguity revealed equivocal findings. Whilst Gayman and Bradley (2013), a high-quality study, reported increased role confusion related to reduced emotional exhaustion, Holgate and Clegg (1991) found the obverse, that increased role ambiguity was correlated significantly with increased emotional exhaustion and reduced personal accomplishment.

Discussion

The aims of the current review were to systematically interrogate published literature on the extent and predictors of WRD amongst CCOs. Nineteen articles met inclusion criteria utilising disparate conceptualisations of WRD, with the majority of papers examining burnout and proportionately few examining trauma-related responses.

All but three studies investigated burnout (Lewis et al., 2013; Merhav et al., 2018; Rhineberger-Dunn et al., 2016), an understandable but narrow focus when considering WRD from a perspective of sustained, intense work with a substantial client caseload. A consistent finding across studies demonstrated burnout at mild to moderate levels and commensurate with other professionals working with offenders, notably institutional corrections officers and police officers (Anson & Bloom, 1988; Lindquist & Whitehead, 1986). However, across most papers it was unclear how the extent of burnout was determined, and explicit comparison to population norms would contextualise the extent of distress that CCOs report. That burnout as an outcome is assessed in 16 papers suggests a significant and growing awareness of the deleterious impacts of distress engendered through work with offenders, and such a body of research offers a substantial basis on which to consider and develop supportive interventions.

Regarding dimensions of burnout, high levels of emotional exhaustion (Gayman & Bradley, 2013; Gayman et al., 2018) were prominent, mirroring levels widely reported by social workers (Morse et al., 2012; Paris & Hoge, 2010). Aspects of CCO roles are akin to social work, relying on effective therapeutic rapport building, and involving regular exposure to emotionally complex situations. The findings suggest that the CCO role takes an emotional toll, adversely affecting mental health, and potentially increasing sickness rates and reducing quality of support to offenders. Therefore, greater consideration should be afforded to systematic supports for CCOs, potentially through peer support assuming familiarity with roles, and their inherent complexities.

Yet whilst burnout is privileged in research reviewed herein, and can capture the cumulative impacts of work as a CCO, the relative paucity of papers examining secondary trauma may reflect a more limited awareness of the traumatogenic exposure inherent in the work. In the one study to explicitly investigate prevalence of secondary trauma (Rhineberger-Dunn et al., 2016), CCOs were found to report greater secondary trauma than did residential officers. Significant levels of secondary trauma are reported in other legal professionals (James, 2020), police (Brady, 2017) and social workers (Bride, 2007), and are associated with reduced work performance and higher rates of absenteeism (Ludick & Figley, 2017), and interventions are now advanced to tackle this distress (James, 2020). Thus further research seems warranted to assess whether CCOs are as vulnerable to this form of distress as similar professional groups, and develop strategies to mitigate, and retain staff and their expertise.

Synthesis of studies in this review revealed predictors across personal, organisational and role-based domains. Of those investigating burnout, nine reported personal variables including age and gender (Gayman & Bradley, 2013; Jin et al., 2018; Rhineberger-Dunn et al., 2017). Fewer papers, eight in total, investigated organisational variables, concentrating narrowly on hours of contact, caseloads and years in role. The three papers investigating alternative forms of WRD also reported predictive factors across personal and organisational-based domains, demonstrating that attachment style and previous trauma history were predictive of VT (Merhav et al., 2018), whilst caseload size and type significantly predicted secondary trauma (Rhineberger-Dunn et al., 2016) and CF (Lewis et al., 2013), respectively. Across all studies there was an absence of exploration into organisational factors known to impact on levels of WRD amongst other professions, such as workplace and co-worker support (Hensel et al., 2015; A. Miller et al., 2017). The findings suggest that future research should examine organisational predictive factors recognised in other professions (McFadden et al., 2015), and promote organisational strategies to enhance resilience rather than taking a solely individual narrative, potentially blaming and shaming for staff. Increased focus on the sustained challenges for CCOs may promote their organisational value, particularly during times of austerity with limited resources and a lack of psychological safety: factors associated with increasing WRD, as reported in similar professions (Alarcon, 2011; Grootegoed & Smith, 2018).

That studies differed in extent of distress reported and predictive factors may be attributed to diversity in how studies conceived and measured WRD. Even when studies utilised the same construct, notably burnout in 14 studies, they applied them differently. Some studies combined scores of frequency and intensity from the MBI across the three dimensions of burnout, rather than keeping them separate, likely contributing to disparities in extent of burnout reported. Additionally, Jin et al. (2018) used a modified version of the MBI measure in which a single scale of burnout was calculated rather than three separate scales, despite recommendation that they are not combined (Panagioti et al., 2017).

Implications

Synthesised findings across the majority of studies indicate that CCOs experience burnout as a result of their work, at levels similar to those of other professionals working in criminal justice settings. In the very few studies that examined trauma responses, levels of secondary trauma were elevated, and indicate a need to examine not only the extent and manifestations of trauma-presenting WRD, but strategies and processes to identify and manage distress, at both individual and organisational levels, as well as to build resilience.

Findings in this review highlighted predictive organisational factors for burnout, and interventions to mitigate WRD should address these to complement those targeting the CCOs individually. Organisational structures should be made aware of the detrimental impact of excessive contact hours and large caseloads to direct sufficient available work resources. Furthermore, as role conflict was frequently cited as a predictive factor, organisations should be advised to consider the impact of the range of roles that CCOs have to fulfil. Clinical interventions at an organisational level could involve psychoeducation for all staff regarding the impact of chronic work distress with the aim of breaking down cultural barriers and attitudes towards mental health difficulties. Organisational approaches to wellbeing are advised, perhaps adopting employee assistance programmes, designed to support staff to manage difficulties that may impact on work, or similar to the Blue Light Wellbeing Framework (Coleman et al., 2018) applied nationally in UK police organisations, promoting regular organisational self-assessments.

Strengths and limitations

The present review was rigorously systematic in its approach and offers an up-to-date synthesis reporting on extent of WRD amongst CCOs. However, the findings of the review were limited to articles reported in English, with most papers specific to service delivery in Westernised countries, particularly the USA. Community corrections systems vary across countries and cultures, limiting the generalisability of the findings. Most noticeably, not one study included a sample from the UK, indicating a clear need for future research to address this gap.

There were various methodological frailties evident. All designs were cross-sectional, precluding establishing firm conclusions regarding causality (both independent and dependent variables were assessed at the same time, giving no evidence of a temporal relationship between them; Bowen & Wiersema, 1999), and longitudinal designs are needed to explore causal relationships further. All 19 included studies utilised self-report data, which are prone to bias and can increase the risk of overstated associations between factors (Theorell & Hasselhorn, 2005). Furthermore, papers were excluded from this review investigating PTSD and other forms of direct traumatisation, given that the focus was to explore inadvertent WRD that is not anticipated in the job role and therefore is potentially unacknowledged as a form of risk to wellbeing at work. However, it can be difficult to distinguish between responses to primary and secondary trauma, and all but one paper (Merhav et al., 2018) did not consider participants’ own trauma histories during investigations; it is therefore possible that findings were influenced by participants’ experiences of direct trauma. It is recommended that a review evaluates the literature concerning direct trauma in CCOs, to increase greater awareness in this area.

Study quality varied; whilst seven papers met the criteria threshold to be considered high quality, the majority were at best moderate with common flaws evident. Only more recent studies attempted to address or define characteristics of non-responders (Jin et al., 2018; Rhineberger-Dunn et al., 2017), perhaps reflecting a key difficulty with cross-sectional designs and opportunistic sampling. Non-responders may have had higher WRD and felt unable to engage in the study as a result of distress. Future research could address this through requesting organisations to provide basic characteristics of all employees to aid the consideration of non-response bias. Furthermore, sample size was only justified in two studies (Allard et al., 2003; Jin et al., 2018), whilst none reported power analysis or effect sizes. Further research in this area should address this issue by completing a priori calculations and ensuring these are reported, to enable transparency of methodology.

Conclusions

The systematic review of WRD identified evidence of mild to moderate levels of burnout amongst CCOs, akin to those in other professions within law enforcement. A number of predictive factors were found for this population ranging across personal, organisational and role-based domains. The review also revealed circumscribed evidence of elevated secondary trauma comparable to that in similar professions. This warrants further research on vicariously transmitted distress for these staff groups, given the sustained engagement with vulnerable clients. Given predictors of WRD in both individual and organisational domains, clinical interventions can target both, with regard to identification of distress and interventions to mitigate.

Ethical standards

Declaration of conflicts of interests

Jessica Page has declared no conflicts of interest

Noelle Robertson has declared no conflicts of interest

Ethical approval

This article does not contain any studies with human participants or animals performed by any of the authors.