The effectiveness of occupational therapy supporting return to work for people who sustain serious injuries or develop long-term (physical or mental) health conditions: A systematic review

Abstract

Introduction:

People with long-term conditions or recovering from serious injuries can struggle to return to work. The evidence for occupational therapy supporting return to work is limited. We aimed to identify and explain how occupational therapy interventions work.

Methods:

Systematic review. Seven databases were searched between 1 January 1980 and 15 June 2022. Studies measuring work-related outcomes among individuals receiving occupational therapy during absence from paid work were included. Multiple reviewers independently contributed to screening, quality appraisal and data extraction processes. Data were analysed as a narrative.

Results:

Twenty studies with 3866 participants were included; 17 were assessed as having high risk of bias. Occupational therapy was inconsistently acknowledged affecting study identification and occupational therapy components were poorly described. Meta-analysis was unfeasible due to outcome heterogeneity. Individually tailored occupational therapy focused on return to work in musculoskeletal conditions indicated the most promising outcomes. Key intervention components included vocational assessment, goal setting and self-management. Key mechanisms of action included early intervention, individualised support and being responsive to needs.

Conclusion:

Occupational therapists’ contributions supporting return to work should be clearly attributed. Future effectiveness research should standardise the measurement of work outcomes to support meta-analysis. Developing a taxonomy for occupational therapy supporting return to work could facilitate comparisons across studies, highlighting occupational therapists’ roles and facilitating training and benefits to patients.

Introduction

Supporting people who have long-term health conditions or sustain serious injuries in returning to work is a core occupational therapy (OT) business (Royal College of Occupational Therapists, 2020). Supporting people returning to work is a major responsibility for healthcare professionals in the United Kingdom (UK) and a key outcome of National Health Service interventions (Department of Health, 2010). It is also important to the UK economy, and individuals’ physical, mental and financial well-being (Royal College of Occupational Therapists, 2018).

One-third of people admitted to hospital following serious injury do not return to work (RTW) within 12 months (David et al., 2022). Those that do, may suffer from physical and mental health problems, including pain, fatigue, anxiety, depression and post-traumatic stress disorder, which threaten work stability (Cancelliere et al., 2016; Kendrick et al., 2017; Vardon-Bounes et al., 2021; Vitturi et al., 2022) and increase health resource use. Each year, around 300,000 people with mental health conditions fall out of work in the UK, and to date, there is limited evidence on how best to support them (Carol Black, 2008; Department for Work and Pensions, 2016).

The effectiveness of OT interventions for supporting RTW following illness or injury remains unclear. A 2011 review only reported findings in favour of using OT as part of a multidisciplinary team (MDT); however, no meta-analysis was conducted, possibly due to study heterogeneity (Désiron et al., 2011). Since then, the development of reporting guidelines (e.g. CONSORT) may have led to improved quality of reporting.

Thus, there is a need to systematically review the contemporary evidence available on OT RTW interventions, to address the following research questions:

1. What RTW interventions are being delivered as part of OT to working-aged people with serious injuries or long-term physical/mental health conditions?

2. Are the OT interventions effective?

3. What are the components and mechanisms of action of OT interventions that facilitate RTW?

Methods

The protocol for this systematic review was registered on PROSPERO CRD42020211670.

Inclusion and exclusion criteria

Research articles were included if (1) the study described an OT intervention including a work-related outcome. A wide range of OT interventions were included such as workplace interventions, vocational rehabilitation (VR) interventions, service coordination interventions, work hardening and multi-component interventions. There were no limitations on the number, format, methods, intensity or duration of treatments. RTW interventions could be delivered as a stand-alone OT intervention or as multidisciplinary rehabilitation in different settings, including community-based services in the public, private and third sectors (e.g. charities, voluntary and community organisations); (2) the study (randomised controlled trials (RCTs) and cohort studies) included a comparator such as control or another non-OT active intervention; (3) the primary outcome of the intervention was work status, which we defined as relating to work disability, that is, sickness absence, total time loss and time until RTW, but also limitations in meeting work demands to stay at work. Work status could be expressed dichotomously for RTW (yes/no) or work status (working/not working) and (4) participants were adults (+16) in paid employment who were absent from work due to an injury or long-term physical or mental health condition.

Studies were excluded if (1) the intervention did not include an OT component; (2) the OT did not include a RTW outcome and (3) studies included participants with congenital health conditions.

Literature search

The search strategy was constructed and piloted using Ovid Medline by the research team that included a librarian. The strategy was adapted to fit the requirements of each database. The databases (Ovid Medline, Ovid Embase, Ovid PsycINFO, Cochrane Library, ClinicalTrials.Gov, CINAHL and ProQuest Theses & Dissertations) were searched for studies published between 1 January 1980 and 15 June 2022 (see Supplemental Appendix 1 for search strategy).

Seven reviewers conducted the selection process to increase the screening reliability. Records from the search were uploaded to Covidence systematic review software (Veritas Health Innovation, 2016) to manage the selection process. Duplicate studies were removed, and the titles and abstracts of each record were independently screened by two researchers. Discrepancies were resolved through discussion with a third reviewer. The full texts of all relevant studies were screened in the same way. Reference lists of included papers were reviewed to identify studies not found in the searches.

Data extraction and synthesis

A data extraction form was developed, piloted, and modified by the review team using Covidence (Veritas Health Innovation, 2016). Data on interventions were extracted using the template for intervention description and replication (TIDieR) checklist (Hoffmann et al., 2014) and the Rehabilitation Treatment Specification System (RTSS) (van Stan et al., 2019) to explain the intervention components and mechanisms of action. Data were extracted independently by two reviewers; any discrepancies were resolved through discussion with a third reviewer.

The following information was extracted from the studies: author, country, study type, inclusion criteria, participant’s characteristics and intervention description following the TIDieR checklist (Hoffmann et al., 2014). The intervention description included a section reporting on the OT intervention. When occupational therapists deliver RTW support alone or in a team, this is often labelled as VR, which can be defined as, ‘a multi-professional evidence-based approach that is provided in different settings, services, and activities to working-age individuals with health-related impairments, limitations, or restrictions with work functioning, and whose primary aim is to optimize work participation’ (Escorpizo et al., 2011). Drawing on previous work revealed whether a study’s intervention could be identified as VR using existing descriptions (Cullen et al., 2018) and VR intervention terminology by Hart (2006) and Cullen et al. (2018). A glossary of terms can be seen in Supplemental Appendix 2.

The data extraction form also included a section regarding the intervention outcomes. Our primary outcome was work status, which we defined as relating to work disability, that is, sickness absence, total time loss and time until RTW, but also limitations in meeting work demands to stay at work. This could be expressed dichotomously for RTW (yes/no) or work status (working/not working) at a time point within study groups. Continuous outcome measures included the number of hours at work, number and/or duration of sickness absences, total duration of sick leave over a given period and recurrences of sick leave/work absence (by self-report or collected from organisational or system records measurements). RTW was assessed as the rate of RTW amongst a group after their allocation to RTW OT intervention or control conditions. Because work outcomes are usually reported in a multitude of ways, we used an existing framework by Wasiak et al. (2007). We identified any work status outcome and categorised it using the framework that encompasses four phases: off work, work reintegration, work maintenance and advancement. Secondary outcomes included all other outcomes reported such as functional ability, mood, and quality of life. Due to study and outcome heterogeneity, the results are presented as a narrative synthesis (Popay et al., 2006), which refers to the process of synthesising the identified studies using descriptions (a narrative) of the studies and findings, as opposed to statistical synthesis.

Assessment of methodological quality

We assessed the risk of bias using the Cochrane tool from the Cochrane Handbook for Systematic Reviews of Interventions (Higgins et al., 2011). This tool assesses six main domains of bias (performance bias, detection bias, attrition bias, reporting bias and other biases) (Higgins et al., 2011). Each study was assessed by two independent reviewers; a third was consulted to resolve discrepancies.

Results

Study selection

In total 6633 studies were identified, and 307 were removed as duplicates, resulting in 6326 studies screened by title and abstract. We excluded 6017 based on the title and abstract and assessed the full text of 309 studies for eligibility. Of these 289 studies were excluded. The reasons for exclusion were: not OT intervention (n = 112), no RTW outcome (n = 75), excluded study design (n = 66), not in English (n = 11), participants were not in work or did not return to their original job (n = 7), unable to obtain full text (n = 7), ongoing studies (n = 6), or participants were at work, unemployed or not returning to work (n = 5). Twenty studies were included in the review. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 flowchart (Page et al., 2021) depicts the study selection process (Figure 1).

Figure 1.

PRISMA 2020 flowchart.

Study characteristics

Among the 20 included studies (Table 1), there were 18 which were RCTs, one cohort study, and one non-randomised study spanning 24 years (1995–2019) of literature from 10 countries including Sweden n = 4 (Berglund et al., 2018; Carlsson et al., 2013; Eklund et al., 2013; Johansson et al., 1998), France n = 1 (Jousset et al., 2004), Netherlands n = 4 (Hees et al., 2013; Lambeek et al., 2010; Schene et al., 2007; van Vilsteren et al., 2017), Denmark n = 3 (Bendix et al., 1995, 2000; Stapelfeldt et al., 2011), Canada n = 2 (Park et al., 2018; Sullivan et al., 2006), UK n = 2 (Hammond et al., 2017; Macedo et al., 2009), United States of America n = 1 (Keysor et al., 2018), Switzerland n = 1 (Kool et al., 2007), Germany n = 1 (Fauser et al., 2019) and Australia n = 1 (Wu et al., 2017) (see Supplemental Appendix 3 for theories underlying the development of the interventions).

Table 1.

Summary of studies included.

First author, year of publication, country and study design	Type of intervention, target period and health condition	Participants total (intervention arm (females)) and inclusion criteria	Staffing and Intervention attributes	Components ( Delivered by occupational therapists )	Control group	Primary and secondary outcomes
Long-term physical health
Bendix et al. (1995) Denmark RCT	Health-Focused Intervention MDT Early (1–6 m) LBP	N = 132 (40 (n = 30)) • >6 months of disabling low back pain. • Risk of job loss. • Aged 18–59 • Able to read and write in Danish	OTT, PT, psychologist 6 wks (3 wks, 39 h/wk; 3 wks 1 day/wk) In person: individual, group Clinic	• Vocational assessment • Vocational counselling/education • Work preparation •  Work hardening • Behavioural/interpersonal interventions • Self-responsibility and self-management • Graded activity/exercise • Therapeutic recreation	Control 1: Active physical training Control 2: Psychological pain management and active physical training	4 m follow-up Sick leave (days) Back, leg pain Perception of disability
Bendix et al. (2000) Denmark RCT	Health Focused Intervention MDT Target period unclear LBP	N = 138 (64 (n = 39)) Person in precarious work situation because of low back pain	OTT, PT, psychologist 6 wks (3 wks, 39 h/wk; 3 wks 1 day/wk) In person: individual, group Clinic	• Vocational assessment • Vocational counselling/education • Work preparation •  Work hardening • Behavioural/interpersonal interventions • Self-responsibility and self-management • Graded activity/exercise • Therapeutic recreation	Outpatient physical training	12 m follow-up Sick leave (days) Back, leg pain
Johansson et al. (1998) Sweden RC Study	Multi-domain Intervention MDT Long-term (12+ m) Chronic musculoskeletal pain	N = 42 (21 (n = 36)) • Chronic musculoskeletal pain • No further medical or surgical treatment appropriate for pain • No psychotic illness present	OTT, clinical psychologist, PT, physical education teacher, vocational counsellor, physician, nurse 4 wk: 5 d/wk. 2 m later + 2 booster sessions + ad hoc sessions thereafter. In person; individual, group Inpatient ward	• Goal setting • RTW planning and coordination • Work hardening • Ergonomics • Self-responsibility and self-management • Graded activity/exercise	Waiting list control	1, 2 and 12 m follow-up Sick leave Activity levels Catastrophising and pain behaviours
Jousset et al. (2004) France RCT	Health Focused Intervention Unidisciplinary Long-term (12+ m) LBP	N = 86 (44 (n = 13)) • Aged 18-50 • Living within three counties of the west of France • Engaged in a non-limited work contract • Risk of job loss by chronic LBP • LBP not relieved by medical or surgical interventions	OTT, PT 5 wks In person, groups Clinic	• Physical/OT • Work hardening	Active individual therapy (3 h/week)	6 m follow-up Sick leave (days) Pain QoL Functional status Mood
Kool et al. (2007) Switzerland RCT	Health-focussed intervention MDT Target period unclear Chronic low back and leg pain	N = 174 (87 (n = 18)) • Aged 20–55 • Primary diagnosis of non-acute LBP • >6 wks sick leave in the last 6 m	OTT, rheumatologist, PT, sport therapist, social worker, nurse, psychologist 3 wks; 4 h/d, 6 d/wk In person: individual Clinic	• Vocational assessment • Job analysis • Work hardening • Self-responsibility and self-management • Graded activity/exercise	Pain centred treatment (primary goal pain reduction)	3 and 12 m follow-up Workdays Disability compensation Self-efficacy Strength Pain Mobility
Lambeek et al. (2010) Netherlands RCT	Multi-domain intervention MDT Early (1–6 m) LBP	N = 134 (66 (n = 29)) • Aged 18-65 • LBP for more than 12 wks • Paid employment at least 8 h/wk • Absent/ partially absent from work	OTT, OHP, medical specialist, PT 3 m In person: individual + group Clinic, workplace	• Vocational assessment •  Job analysis • Case management • RTW planning and coordination • Work hardening • Graded exercise •  Ergonomics • Emotional/adjustment interventions	Usual care – from specialist occupational physician, general practitioner, and/or allied health professionals.	3, 6, 9 and 12 m follow-ups Days to sustainable RTW Sick leave Functional status Pain intensity
Stapelfeldt et al. (2011) Denmark RCT	Case and Service Coordination MDT Early (1–6 m) LBP	N = 351 (176 (n = 95)) • Age 16–60 • Partially or fully sick-listed from work for 4–12 wks because of LBP	Case manager (OTT, social medicine specialist, social worker), physician, PT Duration/frequency unclear In person: individual Clinic	• Vocational assessment • Goal setting • Vocational counselling/education • Case management/ advocacy • RTW planning and coordination • Emotional/adjustment interventions • Self-responsibility and self-management	Brief clinical Intervention – advice, physical exercise with physiotherapy follow up after 2 wks.	12 m follow-up Sickness absence Time to RTW Sub-group analysis of group with low and high job satisfaction and job control
Cancer
Fauser et al. (2019) Germany Cluster RCT	Multidomain Intervention MDT Target period unclear Cancer	N = 484 (229 (n = 163)) • Aged 18-60 • Completed initial cancer treatment • No active disease • Risk of not returning to work • Employability for at least 3 h a day	OTT, psychologist, physician, PT, social worker 25 d: 1 h assessment, 6 h + 4 h + 3 h intervention In person; individual, group Inpatient ward, clinic classroom	•  Vocational assessment •  Job analysis •  Goal setting • Vocational counselling/education • Case management/advocacy • RTW planning and coordination •  Work hardening • Workplace adjustments • Ergonomics • Cognitive remediation • Self-responsibility and self-management • Group peer support	Conventional medical rehabilitation – no detail	3 m follow-up Sick leave Disability days off Employment status Work ability Quality of life Fatigue Coping skills
Injury-related conditions
Park et al. (2018) Canada Cluster RCT	Health Focused Intervention MDT Early (1–6 m) Musculoskeletal disorder (whiplash)	N = 728 (367 (n = 123)) • Active workers’ compensation claims for musculoskeletal disorder • Participating in RTW programme	OTT and exercise therapist Duration/frequency unclear In person: individual Clinic	• RTW planning and coordination • Work hardening • Self-responsibility and self-management • Interventions addressing motivation	Interdisciplinary approach focused on improving physical and functional abilities, RTW planning, individual counselling, and educational workshops	At discharge from intervention RTW (yes/no)
Sullivan et al. (2006) Canada Cohort	Case and Service Coordination MDT Early (1–6 m) Whiplash injury	N = 130 (70 (n = 32)) • Employed prior to whiplash injury • Attending research rehabilitation clinic	OTT, PT 10 wks; 1 h/wk In person: individual Community-based	• Goal setting • Emotional/adjustment interventions • Self-responsibility and self-management • Psychosocial targets • Graded activity/exercise	Functional restoration physical therapy intervention	4 wks follow-up RTW Pain disability
Wu et al. (2017) Australia RCT	Health Focused Intervention MDT Early (1–6 m) Road trauma injury	N = 220 (107 (n = 33)) • Aged over 18 • Sustained road trauma	OTT, PT, rehabilitation physician, nurse (in-reach rehabilitation team) Duration based on patient need: two sessions of physiotherapy and/or OT per day In person: individual Acute hospital	• Medical rehabilitation • Emotional/adjustment interventions • Cognitive remediation	Usual care – ward-based rehabilitation; no detail	RTW Functional independence Mental health Pain
Inflammatory conditions
Hammond et al. (2017) UK Feasibility RCT	Work Modification Intervention Unidisciplinary Long-term (12+ m) Inflammatory arthritis	N = 55 (29 (n = 20)) • Aged over 18 • Diagnosis of RA • In paid work • Able to read, write, and understand English • Willing to receive VR	OTT 2–4 m: 4.5 h + 1.5 h if required In person + remote; individual Clinic, telephone, home, workplace	•  Vocational assessment •  Job analysis •  Goal setting •  Vocational counselling/education •  Case management/advocacy •  Work hardening •  Workplace adjustments •  Ergonomics •  Formal review •  Self-responsibility and self-management •  OT (health condition)	NHS usual care, written self-help work information	6 and 9 m follow-ups Employment status Work self-efficacy Confidence to work Ability to manage arthritis at work
Keysor et al. (2018) USA RCT	Case and Service Coordination MDT Target period unclear Rheumatic or musculoskeletal condition	N = 287 (143 (n = 104)) • Aged 21–65 • Employed (<15 h) • Living or working in Massachusetts • Self-reported or diagnosed rheumatic or musculoskeletal condition	OTT, PT 1.5-h meeting and follow-up support at 3 wks and 3 m In person + remote: individual Clinic, telephone	• Vocational assessment • Goal setting • Vocational counselling/education • Case management/advocacy	Packet of written resources via email	6, 12 and 24 m follow-ups Employment status Functional work limitations Presenteeism
Macedo et al. (2009) UK RCT	Multi-domain intervention Unidisciplinary Long-term (12+ m) RA	N = 32 (16 (n = 15)) • RA diagnosis • Employed • Fluent English • Lived locally • Medium or high work instability	OTT 6 m: 30 min/session ⩽6 sessions or 6 m In person: individual Clinic, home, workplace	•  Vocational assessment •  Goal setting •  Case management/ advocacy •  Work modification, adaptation, adjustment •  Ergonomics •  Behavioural/ interpersonal interventions •  OT (health condition)	Usual care – routine reviews by the rheumatologist	6 m follow-up Function Work productivity Coping RA disease activity
van Vilsteren et al. (2017) Netherlands RCT	Multi-domain intervention MDT Target period unclear RA	N = 150 (75 (n = 63)) • Aged 18–64 • RA diagnosis • Employed >8 h/wk • Minor difficulties at work • <3 m sick leave	OTT, OHP, rheumatologist 12 wks: frequency? In person + remote: individual Clinic, telephone	•  Job analysis • Case management/advocacy (physician-led) •  Work modification, adaptation, adjustment •  Ergonomics •  Formal review (job retention) •  Formal reporting	Usual care – no description	12 m follow-up Working hours Productivity loss Functional work limitations QoL Pain and fatigue
Mental and physical health condition
Berglund et al. (2018) Sweden RCT	Multi-domain Intervention MDT Long-term (12+ m) Mental illness + pain	N = 427 (178 (n = 161)) • on long-term sick leave for mental illness and/or chronic pain • 20–64 years	OTT, psychologist, physician, and social worker ⩽1 year In person: individually Clinic	• Vocational assessment • Goal setting • Case management/advocacy • Emotional/adjustment intervention	Usual care available, if sought, via Swedish Public Employment Service	12 m follow-up RTW (based on increased income) Income Mental health
Carlsson et al. (2013) Sweden RCT	Health Focused Intervention MDT Early (1–6 m) Psychiatric or Musculoskeletal diagnoses	N = 33 (18 (n = 11)) • Sick-listed • Employed • ICD-10 diagnosis • Ongoing sick-leave (max. 28 days)	OTT, PT, and psychotherapist. Duration/frequency unclear. Assessment only In person: individual GP clinic	•  Vocational assessment •  Formal reporting	Usual care – no description	12 m follow-up Sick leave
Mental health conditions
Eklund et al. (2013) Sweden Non-randomised experimental study	Health Focused Intervention Unidisciplinary Early (1–6 m) Stress	N = 84 (42 (n = 42)) • ICD-10 Stress-related diagnosis • Employed • On sick leave	OTT 16 wks: Phase I and II 5 wks each, phase III, job placement 6 wks; 2.5 h sessions. In person; group Clinic	•  Vocational assessment •  Goal setting •  Vocational counselling/education •  RTW planning and coordination •  Work hardening •  Emotional/adjustment interventions •  Self-responsibility and self-management •  Peer support	Usual care – follow-up with the Social Insurance Office and employer. About 50% of reported additional focussed work rehabilitation, e.g. PT, CBT, mindfulness training, pain rehabilitation or work training in an ordinary workplace.	12 m follow-up Sick leave Worker role perception (individual and environment/social) Mental health
Hees et al. (2013) Netherlands RCT	Multi-domain intervention Unidisciplinary Early (1–6 m) Major depression	N = 117 (78 (n = 37)) • Aged 18–65 • Depression for at least 3 m • Absent from work for ⩾25% of contracted hours for ⩾8 wks	OTT 4 m: 18 OT sessions; 9 individual, 8 group, 1 with employer In person + remote; individual, group Clinic, telephone, workplace	•  Vocational assessment •  Job analysis •  Goal setting •  Vocational education •  Case management/advocacy •  RTW planning and coordination •  Formal review •  Self-responsibility and self-management •  Peer support	Usual care – psychiatric residents; clinical management, psychoeducation, supportive therapy and CBT. Pharmacotherapy as required.	6, 12 and 18psychoeducationm follow-ups Work participation Absenteeism (h) Time (d) to RTW Work limitations Self-efficacy Coping skills Health
Schene et al. (2007) Netherlands RCT	Health-focussed intervention Unidisciplinary Target period unclear Work-related depression	N = 62 (30 (n = 15)) • Aged above 18 • Major depressive disorder without psychotic features • No history of psychosis or drug abuse • BDI > 15 • Work reduction of at least 50% because of depression for a minimum of 10 weeks and maximum 2 y	OTT + usual care 48 wks: 2 wk assessment; 24 wk group session ×1/wk(2 h) + 12 individual sessions + ×3 in person follow ups over 20 wk In person: individual, group Outpatient clinic	•  Vocational assessment •  Goal setting •  Vocational counselling/education •  Work preparation •  RTW planning and coordination •  Behavioural/interpersonal interventions	Usual care – out-patient psychiatric treatment; clinical management, psychoeducation, supportive therapy, and CBT. Pharmacotherapy as required.	3, 6, 12 and 42 m follow-ups Hours worked Mental health Work stress

Components in bold and italics indicate they were delivered by occupational therapists.

RCT: randomised controlled trial; RC: randomised controlled; OTT: occupational therapist; OHP: occupational health physician; PT: physiotherapist; VR: vocational rehabilitation; MDT: multidisciplinary team; RTW: return to work; HADS: Hospital Anxiety and Depression Scale; UK: United Kingdom; USA: United States of America; ICD-10: International Classification of Diseases 10; LBP: low back pain; WLQ: work limitations questionnaire; WIS: work instability scale; NHS: National Health Service; RA: rheumatoid arthritis; HAQ: Health Assessment Questionnaire; COPM: Canadian occupational performance measure; DAS: disease activity score; BDI: beck depression inventory; min: minutes; h: hour; wk: week; m: month; y: year; CBT cognitive behavioural therapy; OT: occupational therapy.

Risk of bias

Risk of bias ratings are shown in Figure 2; only three (15%) studies had a low risk of bias (Keysor et al., 2018; Park et al., 2018; Wu et al., 2017). Across the studies, the lowest ratings were given to the ‘blinding of participants and personnel’ domain because VR and OT require participants to actively engage in the treatment process. For the remaining domains, 14 studies (70%) used appropriate sequence generation, 10 studies (50%) used appropriate allocation concealment, 9 studies (45%) used appropriate blinding of outcome assessor and 7 studies (35%) included complete outcome data.

Figure 2.

Risk of bias assessment.

Participants

Overall, 3866 participants were included at entry into the studies with 1889 (49%) receiving the experimental intervention. Participants’ age ranged from 16 to 65 years. Over half of the participants 1064 (56.3%) identified as female, with studies including between 21% (Kool et al., 2007) and 100% (Eklund et al., 2013) female participants in the intervention arm. Most studies (n = 18) did not report participant ethnicity but where they did, most participants were identified as White. Other categories available were ‘not white’ (Keysor et al., 2018) or African American, Asian or other (Macedo et al., 2009).

Even though education is a predictor of RTW, the highest level of attainment was not reported in seven studies. Where it was reported, this varied widely. Local educational systems are organised differently across countries, and this resulted in the heterogeneity of reporting and meant meta-analysis was not possible.

Five studies (25%) reported occupation types following a simple categorisation such as unskilled or skilled workers (Hees et al., 2013; Jousset et al., 2004; Kool et al., 2007; Stapelfeldt et al., 2011; van Vilsteren et al., 2017). Three studies (15%) followed a more detailed classification following four broad levels: level 4 (professionals and managerial), level 3 (associated professionals and technical/skilled trades), level 2 (administrative, caring, leisure, sales) and level 1 (elementary occupations) (Eklund et al., 2013; Hammond et al., 2017; Macedo et al., 2009).

Interventions were directed at participants with a range of conditions or diagnoses, with most (n = 11) related to pain or painful conditions. This included low back pain (Bendix et al., 1995, 2000; Jousset et al., 2004; Lambeek et al., 2010; Stapelfeldt et al., 2011), chronic low back and leg pain (Kool et al., 2007), wider spread musculoskeletal pain (Johansson et al., 1998; Park et al., 2018; Sullivan et al., 2006), mental illness and/or pain-related diagnosis (Berglund et al., 2018; Carlsson et al., 2013). Four studies included participants with inflammatory arthritis diagnoses (Hammond et al., 2017; Keysor et al., 2018; Macedo et al., 2009; van Vilsteren et al., 2017). Two included participants with depression (Hees et al., 2013; Schene et al., 2007). The remaining studies included participants with a range of health conditions including serious traumatic injury (Wu et al., 2017), stress-related disorders (Eklund et al., 2013) and cancer (Fauser et al., 2019).

Service providers

Five studies reported a unidisciplinary OT intervention (Eklund et al., 2013; Hammond et al., 2017; Hees et al., 2013; Macedo et al., 2009; Schene et al., 2007), but most (n = 15) reported OT interventions delivered by occupational therapists within a MDT programme. Eleven studies indicated which components were delivered by occupational therapists including the five delivered only by occupational therapists (italicised in Table 1).

The MDTs included up to 10 different professionals working alongside occupational therapists. These included physiotherapists (also referred to as physical therapists) across 12 studies (Bendix et al., 1995, 2000; Carlsson et al., 2013; Fauser et al., 2019; Johansson et al., 1998; Jousset et al., 2004; Keysor et al., 2018; Kool et al., 2007; Lambeek et al., 2010; Stapelfeldt et al., 2011; Sullivan et al., 2006; Wu et al., 2017), medical physicians/specialists in eight (Berglund et al., 2018; Fauser et al., 2019; Johansson et al., 1998; Kool et al., 2007; Lambeek et al., 2010; Stapelfeldt et al., 2011; van Vilsteren et al., 2017; Wu et al., 2017), psychologists in five (Bendix et al., 1995, 2000; Berglund et al., 2018; Fauser et al., 2019; Johansson et al., 1998), nurses in three (Johansson et al., 1998; Kool et al., 2007; Wu et al., 2017), social workers in three (Berglund et al., 2018; Kool et al., 2007; Stapelfeldt et al., 2011), physical education teacher or exercise/sports therapist in three (Johansson et al., 1998; Kool et al., 2007; Park et al., 2018), occupational health physicians in two (Lambeek et al., 2010; van Vilsteren et al., 2017), psychotherapists (Carlsson et al., 2013), vocational counsellors (Johansson et al., 1998) and case managers (Stapelfeldt et al., 2011).

Experimental interventions

Figure 3 presents a summary of RTW interventions delivered by occupational therapists for people with serious injuries and long-term conditions. Interventions were delivered individually (1:1) in half of the studies (n = 10), or a mix of group and individual sessions in five studies or in small groups in four studies (Table 1). All study participants were seen in person except in one study, where the intervention was solely delivered remotely (e.g. telephone or letter) (Lambeek et al., 2010); five interventions delivered some intervention components remotely (e.g. telephone, information pack) (Table 1).

Figure 3.

Summary of RTW OT interventions for people with serious injuries and long-term conditions.

RTW: return to work; OT: occupational therapy.

Twenty-four components were identified across all studies (Table 1) and the most frequently occurring were: vocational assessment (n = 14), goal setting (n = 11), self-responsibility and self-management techniques (n = 11), work hardening (n = 10), vocational counselling/education (n = 9), case management/advocacy (n = 8) and RTW planning and coordination (n = 8).

Eleven studies reported intervention components delivered by occupational therapists (italicised in Table 1). These included vocational assessment (seven studies), goal setting (six studies), job analysis (five studies), work hardening (five studies), vocational counselling/education (four studies), ergonomics (four studies), RTW planning and co-ordination (three studies), work modification adaptation and adjustment (three studies), interventions to support self-management (three studies), formal review after RTW (three studies), case management/advocacy (three studies), behavioural interpersonal interventions (two studies), group/peer support (two studies), emotional adjustment intervention (one study) and formal reporting after assessment (one study).

Control interventions

Control groups were included in 11 studies, and mostly received care as usual (Berglund et al., 2018; Carlsson et al., 2013; Eklund et al., 2013; Fauser et al., 2019; Hammond et al., 2017; Hees et al., 2013; Lambeek et al., 2010; Macedo et al., 2009; Schene et al., 2007; van Vilsteren et al., 2017; Wu et al., 2017), or were on a waiting list (Johansson et al., 1998). Participants in the remaining eight studies received an alternate intervention that was not identified as usual care including active physical training (Bendix et al., 2000; Jousset et al., 2004; Sullivan et al., 2006), written self-management materials (Keysor et al., 2018), pain-centred treatment (Kool et al., 2007), interdisciplinary rehabilitation with counselling and educational workshops (Park et al., 2018) or a brief clinical intervention (Stapelfeldt et al., 2011). A single study delivered two additional interventions: (1) active physical training and (2) psychological pain management combined with active physical training (Bendix et al., 1995); and both were treated as comparators for this review.

Impact on RTW outcomes

Eight studies described positive results on RTW favouring those in the intervention group (Bendix et al., 1995, 2000; Berglund et al., 2018; Hammond et al., 2017; Kool et al., 2007; Lambeek et al., 2010; Park et al., 2018; Sullivan et al., 2006). For participants with long-term physical health conditions, the interventions that led to favourable RTW outcomes included intensive MDT functional restoration (Bendix et al., 1995, 2000), function-centred rehabilitation (Kool et al., 2007), and integrated multidisciplinary case management (Lambeek et al., 2010). For those with injury-related conditions, the interventions included a multidisciplinary progressive goal attainment programme for participants with whiplash injury (Sullivan et al., 2006), and an interdisciplinary functional rehabilitation programme plus motivational interviewing for participants with musculoskeletal disorders (Park et al., 2018). One study showed a positive RTW outcome for patients with inflammatory conditions (rheumatoid arthritis (RA)), following a job retention programme (Hammond et al., 2017). Finally, Berglund et al. (2018) tested two interventions (MDT VR and acceptance commitment therapy) to usual care; the intervention increased employability in patients on long-term sick leave due to common mental illness and/or chronic pain.

Four showed mixed results (Hees et al., 2013; Johansson et al., 1998; Macedo et al., 2009; Schene et al., 2007). Macedo et al. (2009) compared case coordination plus targeted OT to usual care for participants with RA. At 6 months follow-up, there was significantly greater work stability in the OT group than in the usual care group. But there were no significant differences between the two groups for workdays missed per month or percentage of days missed per month. In a cognitive behavioural inpatient pain management programme for people with chronic musculoskeletal pain by Johansson et al. (1998) the intervention increased occupational activity but did not decrease the amount of sick leave at 1-month follow-up. Hees et al. (2013) did not find a significant difference in work participation between an adjuvant OT intervention and usual care for people with depression. However, those in the intervention group showed greater improvement in depression symptoms and an increased probability of long-term RTW in good health. Schene et al. (2007) tested the effect of OT compared to usual care for people with major depression. Over the first 18 months, those receiving OT worked significantly more than usual care, but this was not sustained in the longer term (months 19–42).

In two studies participants in the intervention group returned to work sooner or had lower rates of permanent job loss, but the outcomes did not reach statistical significance (Jousset et al., 2004; Keysor et al., 2018).

Finally, in three studies, both the intervention and control groups improved (Carlsson et al., 2013; Eklund et al., 2013; Fauser et al., 2019); and three other studies reported that participants in the control group returned to work sooner or had less sickness absence than those in the intervention group (Stapelfeldt et al., 2011; van Vilsteren et al., 2017; Wu et al., 2017).

Discussion and implications

This systematic review investigated the effectiveness and mechanisms of action of OT RTW interventions for working-aged people with serious injuries or long-term physical/mental health conditions. To our knowledge, this is the first such systematic review since Désiron et al. (2011) to do so.

Overall, the effectiveness of OT interventions for supporting RTW following illness or injury showed varying results. Studies which included a more individualised, person-centred, solely work-focused approach with vocational assessment, goal setting and job analysis appeared to be linked to a better RTW outcome.

However, in the studies where OT was delivered as part of a multidisciplinary intervention the components delivered by occupational therapists were often not clearly defined. Studies reporting RTW rates for people experiencing low back pain, musculoskeletal conditions and arthritis appeared to show more promising results than those focused on other conditions. However, these results need to be considered alongside the assessed risk of bias as a meta-analysis was not possible to conduct because of heterogeneity in the measurement of RTW outcomes and data collection points.

Clinical heterogeneity across individual studies makes it difficult to draw conclusions on the effectiveness of OT interventions on RTW. The differences across the studies were expected due to the diverse ways in which occupational therapists work with different populations to support work needs and the interacting components of VR interventions. We also expected that where occupational therapists delivered interventions solely focused on work, this would positively impact work status outcomes. However, this was not always borne out, possibly due to outcome sensitivity issues or challenges measuring work outcomes. For instance, a positive outcome such as reduced sickness absence might be attributable to an OT intervention; however, sickness absence was measured in different ways by half of the studies included in this review. The differences can be accounted for in part by local policies related to employee remuneration and/or state welfare payments. Eklund et al. (2013: 87) explains, ‘In Sweden, a person can be on 100% sick leave or partial sick leave at 75, 50 or 25%, depending on his or her current workability’, compared with the UK where a person’s sickness absence is typically measured in days.

Heterogeneity in research and clinical practice seeking to measure change is not a new issue and is linked to the nature of complex interventions that encompass OT and VR (Skivington et al., 2021) and the continuing complexity of measuring work status outcomes (Wasiak et al., 2007; Watkin et al., 2020). Selecting standardised outcome measures in research requires consideration and stakeholder involvement (Skivington et al., 2021) and in future could support meaningful meta-analyses for OT and VR interventions.

The interventions included in this review usually delivered OT as part of an MDT for several health conditions. OT as a stand-alone discipline was delivered in two interventions for inflammatory arthritis (Hammond et al., 2017; Macedo et al., 2009), and one intervention for mental health conditions (Eklund et al., 2013). In the MDT interventions it was not always possible to determine which components were delivered by the occupational therapists, though their involvement was sometimes referred to as reporting to stakeholders (e.g. employers, family). This limits the understanding of the impact of occupational therapists supporting RTW. We recommend that researchers carefully consider how best to describe MDT and discipline-specific VR intervention components, such as those espoused in the RTSS (van Stan et al., 2019). Better descriptions will highlight the role of occupational therapists and it would potentially increase the occupational therapists’ knowledge and understanding of their roles and responsibilities in the delivery of VR.

MDT support has been recommended (Désiron et al., 2011); while occupational therapists may lead support around RTW, the complexity of interventions providing support to returning to work require the expertise of other professionals to meet all the needs of the patients. This may lead to professionals overlapping support without clear differentiation between professionals. However, in research, it is important to know who is doing what and to recognise the unique role of each professional.

Data extraction for this review was guided by TIDieR (Hoffmann et al., 2014), plus an extensive, pre-defined glossary of VR terms. Unfortunately, none of the included studies followed reporting guidelines to enhance intervention description, resulting in little clarity regarding the interventions or their components. This hampered study comparison and a deeper understanding of the theories underlying the interventions. The glossary of VR terms proved beneficial for synthesising the included studies. Developing a taxonomy of VR intervention components could standardise intervention descriptions, and by extension further knowledge in the area by facilitating cross-study comparisons. Ultimately this could contribute to improved intervention design, reduce research waste, and increase the likelihood of clinical effectiveness being observed (Hoffmann et al., 2014; Skivington et al., 2021)

Strengths and limitations

We followed guidelines that helped us produce a robust narrative synthesis (Popay et al., 2006). This included assessing the methodological quality of studies by applying the Cochrane risk of bias tool (Higgins et al., 2011) and ensuring that studies were critically appraised when synthesising their findings. Inter-rater reliability was improved through multiple reviewers being involved in all the review processes. Additionally, the search strategy was peer-reviewed by a librarian who specialised in systematic reviews.

Using published literature to frame our data extraction method (TIDieR and RTSS) added robustness to the synthesis. Our research team also included two occupational therapists experienced in VR who reviewed the intervention descriptions to improve the accuracy of the classification of intervention components.

One potential limitation is that the review may have missed some studies because OT was not named within the title and abstract. Systematic reviewers often limit initial search strategies to the title and abstract levels. We strongly recommend that researchers reporting primary research into the effectiveness of OT interventions to support people to work, refer to OT in the title to aid study identification and that they use recognised work status outcomes, the range of which is well documented (Wasiak et al., 2007).

Other limitations are that we only included studies in English, unpublished studies were excluded, and the evidence identified has a substantial risk of bias; therefore, we have interpreted our results cautiously. Finally, even though several interventions reported the same outcome (e.g. sick leave), because of methodological differences in measuring the outcome, and variability in the follow-up time points, it was not possible to conduct a meta-analysis combining data from different studies.

Conclusion

This systematic review investigated the effectiveness of OT interventions supporting RTW for people with long-term physical/mental health conditions or serious injuries. Future research should aim to harmonise intervention descriptions and outcomes and attribute correctly the support delivered by each professional to improve the understanding of what practices are most beneficial to support RTW. Additionally, methods such as realist synthesis may improve understanding of the underlying intervention mechanisms leading to a successful RTW.

Key findings

• Individualised interventions focused on RTW resulted in better work outcomes.

• OT involvement is not always reported accurately.

• Heterogeneity in work status outcomes hampers the interpretation of findings.

What the study has added

This systematic review has highlighted the need to standardise descriptions of work outcomes, intervention components, and occupational therapists’ work to correctly evaluate the effectiveness and mechanisms underlying RTW interventions.