**1. Introduction**

Common mental illness and chronic pain are the two most frequent reasons for long-term sick leave in many countries, including Sweden [1,2]. In addition to the individual suffering related to mental illness and chronic pain, there is a major public/societal burden related to productivity losses [3]. In 2008/2009, the social insurance system in Sweden was reformed and maximum time sick-leave reimbursements were introduced; people on sick leave were transferred to the Swedish Public Employment Service (SPES) to have their work ability assessed [4]. During 2010–2012, about 40,000 people reached the maximum time for sick-leave compensation, and were among the first to have their work ability assessed by SPES, and thus be available for work in the labor market.

People on long-term sick leave often have a long-lasting process back to work [5], where the individuals may be in multiple and recurrent states; i.e., receiving different social security benefits or working, and over time they may shift between these states. This includes shifts between different social security benefits, programs, and part-time and full-time work [5]. Even if individuals do not necessarily move directly to return-to-work (RTW), they might come closer to, or further apart from, the labor market. Longer periods of sick leave are also known as a risk factor for not RTW [6,7]. Other factors that are associated with less RTW are female gender, age, pain, disability, depression, high work demands, previous sick leave, unemployment, and activity limitations [8].

Several strategies have been proposed to prevent work absence and facilitate RTW, and these interventions can be broadly classified as unimodal (for example individual psychotherapy), multimodal programs (team-based assessments and synchronized treatment by several health professionals), and interventions that target a structural level [9]. A recent systematic review of workplace interventions for RTW for musculoskeletal, pain-related, and mental health conditions concluded that there is strong evidence that multimodal intervention encompassing at least two of the three domains shortens duration away from work [10]. The current intervention project, which was carried out between 2010–2012, targeted people on long-term sick leave and at risk of losing their reimbursement from social insurance. Previous studies have foremost targeted people having short-term or medium-term sickness absence [10]. It is still not fully explored how people on very long-term sick leave RTW after experiencing mental illness or chronic pain. The aim of the study was to investigate the effects on RTW or changes in employability for people on long-term sick leave. The objective was to analyze the effects of two vocational rehabilitation interventions: a multidisciplinary team assessment and individualized treatment and/or unimodal psychotherapy with ACT intervention in patients on sick leave due to common mental disorders or chronic pain. To study an empirically relevant intervention outcome that reflects a stepwise rehabilitation perspective, individuals shifting between sources of income was used as an indicator of increased or decreased employability and RTW.

#### **2. Materials and Methods**

This study was conducted as a randomized controlled trial (RCT) and was implemented in two phases. In the first phase, only female participants were allocated to the multidisciplinary treatment (MDT) intervention group, the acceptance and commitment therapy (ACT) intervention group, or to the control group. In the second phase, both women and men participants were allocated either to the MDT group or the control group. The study sample from phase 1 has being used in previous studies [11,12]. The present study evaluates pooled data from both phases.

#### *2.1. Subsection*

Participants eligible for the study were women and men (men only in phase 2) on long-term sick leave or a temporary disability pension due to a mental illness and/or pain-related diagnosis in Uppsala County, Sweden. Mental illnesses included F-diagnoses (with the exceptions of the exclusion criteria diagnosis), and pain diagnoses included M-diagnoses and R-diagnoses defined in the International Statistical Classification of Diseases and Related Health Problems 10th revision (ICD-10). The office of the Swedish Social Insurance Agency (SSIA) identified 1331 individuals on sick leave, with these problems expected to reach their maximum time of sick leave between 2010–2012. After first inclusion, the individuals' sickness certificates were screened by a physician and an occupational

therapist or psychologist to determine fulfilment of the inclusion criteria, and to ensure that they did not fulfil the exclusion criteria. The inclusion criteria were: (1) on sick leave for mental illness and/or chronic pain; (2) aged between 20–64 years. The exclusion criteria were: (1) at high risk for suicide; (2) ongoing alcohol/substance abuse; (3) major mental illness (schizophrenia, bipolar disorder type I, severe social dysfunction/personality disorder); (4) participation in psychotherapy or another vocational rehabilitation program. The SSIA initially identified 1331 people as eligible for the project; after screening, 418 were excluded from the project due to not meeting the inclusion criteria or meeting the exclusion criteria. The remaining 913 individuals were then contacted by mail with information about the project and invited to participate. Out of these, 473 did not respond or declined to participate. A further 13 people were excluded from the research part of the study (but received care in accordance with their allocation) due to being contacted before the formal approval by the ethics committee. The remaining 427 gave their consent to participate, and were randomized by the SSIA into the multidisciplinary team group (n = 178), ACT group (n = 102), or control group (n = 147). Out of these, 282 participants answered the outcome measure (see Figure 1). The inclusion, randomization, and allocation were performed consecutively during the project, about three to four months ahead of the date to match when participants were expected to transition to the SPES. Initially, the participants had an equal chance of being allocated into either the MDT, ACT, or control groups. In phase 2 of the study, two-thirds were randomized to the MDT group and one-third was allocated to the control group. The second phase also included men; the reason for also including men was an increased need among men for this type of intervention.

**Figure 1.** Flow chart of inclusion and follow-up procedure.

#### *2.2. Interventions*

The interventions started one to three months ahead of each participant's expected transfer to the SPES. Patients randomized to the MDT group met individually with professionals from a multidisciplinary team, including a psychologist (PS), a physician (MD), an occupational therapist (OT), and a social worker (SW). The health professional assessed the participant's strengths and limitations for RTW from each perspective. The team then met without the participant to establish an individualized rehabilitation plan, which was later brought back to the participant by one of the team members. The participants had the choice of accepting all, none, or parts of the rehabilitation plan's suggestions and interventions. The plan could involve further examinations, assessments, meetings, or psychotherapy (which then was delivered by the PS) for up to a period of one year. The team met weekly during the project period to evaluate the situation and synchronize the planned or ongoing activities for each participant. Participants allocated to the ACT group received solely ACT treatment. ACT is psychological therapy that uses cognitive behavioral therapy (CBT) and acceptance and mindfulness strategies, together with behavioral strategies, to increase function and quality of life rather than decreasing symptoms to increase psychological flexibility, which has shown to be of value for a number of long-lasting conditions [13]. Most psychotherapy sessions and meetings, for both the multidisciplinary team and ACT group, took place at the clinic, but there was also a possibility of conducting meetings at the participant's home, work, or elsewhere. Sessions were typically about an hour long.

In the second phase of the project, the possibility of being randomized to sole ACT treatment (ACT group) was omitted. The reason was that during phase 1, it turned out that some participants randomized to the ACT group were not interested in receiving ACT, but were interested in receiving other types of treatments. Participants randomized to the MDT group (in both phases) had the option of receiving ACT if suggested by the psychologist after assessment. In addition to the treatments, all of the participants also had scheduled collaboration meetings with their administrator at the SPES, the SSIA, and their contact person within the project. These meetings aimed to ascertain the rehabilitation plan and its goal for the individual as well as for the participating organizations.

The control group did not receive any intervention organized within the intervention project, but was free to receive the usual assistance and care provided by their regular contact with the SSIA, SPES, and potential contacts with healthcare providers. The control group responded to the same follow-up questionnaires as the intervention groups.

#### *2.3. Questionnaires and Outcome Measures*

Data collected before the intervention included demographics, such as the respondent's gender, age, and educational level (categorized as compulsory school, secondary school or equivalent, or university). Factors related to mental health were assessed using the Hospital Anxiety and Depression Scale (HADS) [14]. The HADS is a 14-item scale; seven of the items relate to anxiety, and seven relate to depression. The General Self-Efficacy scale (GSE) was used to measure the participants' perceived self-efficacy [15]. The GSE consists of 10 statements and is answered on a four-point Likert scale ranging from 1 = "Not at all true" to 4 = "Completely true". The items were summed to give a total score of 10 to 40. A mean value was calculated as the sum of all of the answers divided by the number of statements, as long as no more than three statements were missing [16]. The current version of the GSE has been translated into Swedish and has been validated [17]. There is no definite cut-off score for the self-efficacy scale. In this study, the mean value (2.3) of self-efficacy was used to categorize participants with lower (<2.3) versus higher (≥2.3) self-efficacy. The work-related predictor variables that were used in this study included employment contract status (unemployed or employed), extent of sick leave (full-time or part-time), and years with income replacement, which were collected through SSIA registry data.

To measure intervention outcomes, the participants responded to a mailed follow-up questionnaire 12 months from randomization (see Figure 1).

This study's primary outcome uses self-reported sources of income at baseline and follow-up at 12 months to create an outcome variable with four exclusive categories that were believed to capture both changes in RTW as well as changes in employability. Outcome was measured by using the following question: "How do you provide for yourself?" Participants indicated answers in percentages of income from various sources currently and a year earlier. Income sources included: wages (through employment or own business); sickness compensation from the SSIA; and compensation through the SPES. The answers were then compared. If a person reported having increased their proportion of wage due to more work as source of income at 12 months, then they were categorized as "RTW" (having returned-to-work in full or to some degree). A person was categorized as having an "increased employability" if the compensation change indicated increased availability for work, such as increased compensation from the SPES instead of sick leave compensation from the SSIA, as an indication that the person was (more) eligible for work in the labor market. Outcome was considered "negative" if the (part-time) wage was reduced or stopped, or if there was an indication of decreased employability (increased compensation from the SSIA without increased wage). An unchanged outcome was defined as having no changes in wage, employability, or sickness benefits. The outcome variable was further dichotomized for the binary logistic regression analysis into those having a positive (having RTW or increased employability) or negative (having negative or unchanged outcome) outcome.

#### *2.4. Analyses*

Differences in outcome between the two intervention groups and the control group were investigated using Chi-square tests. The intervention's effects were analyzed using binary and multinomial logistic regression models. In randomized controlled trials, confounders are believed to be equally distributed between compared groups. However, it might be of interest to investigate the associations among different factors to the outcome. It might be of further interest to adjust for confounders that are potentially associated with the attrition. To adjust for potential predictors and confounders and other variables that may be associated with the outcome, a stepwise approach was performed using sets of variables. In model 1, the intervention group was adjusted for age and education level. In model 2, additional adjustments were made for HADS, self-efficacy, employment contract, extent of sick leave, and years with income replacement.

Multinomial logistic regressions were used comparing a negative outcome, increased employability, and RTW to those with no changes as a reference category. Multinomial logistic regressions were also performed and adjusted for potential confounders. A complete case analysis was performed in which respondents answered the outcome measure; also, due to a large number of missing values in the outcome variable (34.0%), a sensitivity analysis was performed, in which missing participants were assumed to have made no change in outcome (reference category). All of the tests were two-sided, and a level of *p* < 0.05 was considered statistically significant. The statistical analyses were performed using SPSS statistics (IBM Corp, Armonk, NY, USA), version 22.0.

#### *2.5. Ethical Considerations and Trial Registration*

All of the participants provided written informed consent for the study. The first phase of the study was approved by the Regional Ethical Review Board of Uppsala in 2010 (Reg. no. 2010/088) and the extension (second phase) of the project was approved in 2011 (Reg. no. 2010/088/1). In the reporting of the results of this trial, we have tried to follow the Consolidated Standards of Reporting Trials (CONSORT) checklist as far as possible. The study was registered at the Clinicaltrials.gov Register Platform (ID NCT03343457) on 15 November 2017 (retrospectively registered).

#### **3. Results**

The average age of the study group was 48.9 years (SD 8.3). The study group consisted of 94.7% women and 5.3% men. Secondary school was the most common completed education level. Most of the participants (68.1%) had an employer, and the average time on sick leave was 7.7 years (SD 3.2). About half of the participants were on full-time sick leave, and half were on part-time sick leave.

One-third of the participants were on sick leave for psychiatric disorders, about one-third were on sick leave for pain-related conditions, and about one-third were on sick leave for a combination of psychiatric and pain-related disorders. Common pain-related problems were fibromyalgia and pain in the back, neck, or joints. Common mental disorders were depression (current and recurrent) and stress-related and anxiety disorders.

Participants in the multidisciplinary team group attended on average 4.4 (SD 5.4) sessions delivered by the MD, OT, PT and SW, and 4.7 (SD 6.4) sessions with a psychologist who provided ACT, for a total average of 9.1 (SD 8.4) sessions. Participants in the ACT group attended on average 8.0 (SD 6.0) sessions with a psychologist providing ACT, as shown in Table 1. Of the complete cases in this study, 66% of the participants in the MDT group received at least one meeting with a psychologist providing ACT as one part of their individual plan.


**Table 1.** Characteristics of study participants.

Figures as percentages if not stated otherwise. <sup>a</sup> Hospital Anxiety and Depression Scale, <sup>b</sup> Self-efficacy, <sup>c</sup> Physician, <sup>d</sup> Occupational therapist, <sup>e</sup> Physical therapist, <sup>f</sup> Social worker. ACT: acceptance and commitment therapy.

#### *3.1. Return-to-Work, Increased Employability, Negative Outcome, or No Change*

At the 12-month follow-up, 64 participants (22.7%) had RTW, 69 participants (24.5%) had increased employability, 28 participants (9.9%) had a negative outcome, and 121 participants (42.9%) had not changed in any direction. Among participants in the multidisciplinary team, 31.1% had RTW, 27.0% had increased employability, and 4.3% had a negative outcome. Among participants in the ACT group, 17.9% had RTW, 35.8% had increased employability, and 9.0% had a negative outcome, as shown in Table 2.

**Table 2.** Return to work, increased employability, negative, or unchanged outcome according to different groups.


Pearson Chi-Square test was used for proportions. \*\* *p* ≤ 0.01. RTW: return-to-work.
