The Readiness for Return to Work Scale; Does it Help in Evaluation of Return to Work?

This study applied a prospective cohort design with 12 months of follow-up. Participants took part in one of three randomized controlled trials evaluating the effects of occupational rehabilitation on sickness absence [8, 9]. Two randomized trials compared different inpatient occupational rehabilitation programs with outpatient acceptance and commitment therapy [8]. The third trial compared inpatient programs with or without an added workplace intervention [9]. Several articles have been published from these projects, so the methods therefore overlap somewhat [10‐18]. The studies were approved by the Regional Committee for Medical and Health Research Ethics in Central Norway (No. 2012/1241 and 2014/2279).

Common eligibility criteria for the trials were individuals aged 18 to 60 years who had been sick-listed for 2 to 12 months (at least 50% if on part-time sick leave) with a diagnosis within the musculoskeletal (L), psychological (P), or general and unspecified (A) chapters of ICPC-2 (International Classification of Primary Care, Second Edition). In addition, potential participants in the third trial had to be employed at least 20%, have an employer, and anticipate at least four more weeks of sick leave. Common exclusion criteria were: (a) alcohol or drug abuse; (b) serious somatic (e.g., cancer and unstable heart disease) or psychiatric disorders (e.g., high suicide risk, psychosis, and ongoing manic episodes); (c) disorders requiring specialized treatment; (d) pregnancy; (e) current participation in another treatment or rehabilitation program; (f) insufficient oral or written Norwegian language skills to participate in group sessions and fill out questionnaires; (g) scheduled for surgery within the next 6 months; and (h) serious problems with functioning in a group setting.

The inpatient occupational rehabilitation programs were multicomponent programs designed to target three areas of rehabilitation: mental training, physical training, and work-related problem-solving [8, 9]. The inpatient programs consisted of acceptance and commitment therapy, individual and group-based physical exercise training, mindfulness, psychoeducation on various topics, and individual meetings with a coordinator to solve work-related problems and create an RTW plan. Acceptance and commitment therapy is a form of cognitive behavioral therapy, where both negative and positive experiences are accepted [19]. The individual’s values should guide their actions towards their goals, and the aim is to increase psychological flexibility through mindfulness techniques, values, and committed action [18‐20]. In the first two trials the program lasted 3.5 weeks and 4 + 4 days (with 2 weeks at home in between). In the third trial, the inpatient programs lasted 2 + 1 weeks with 1 week at home in between. The content of the programs was similar to the program described above. In addition, during the week at home, one of the groups received a workplace intervention. The workplace intervention consisted of (a) a preparatory part; (b) a workplace meeting involving the sick-listed worker, the employer, and the primary rehabilitation therapist at the rehabilitation center; and (c) follow-up work related to the meeting. In contrast to the inpatient programs, the outpatient program targeted mainly the mental aspect of the RTW process. The program consisted of outpatient acceptance and commitment therapy, and participants were offered 2.5-h group sessions once weekly for 6 weeks. In addition, one group session provided psychoeducation on physical activity, two individual sessions featured a social worker to clarify personal values and work-related issues, and a short individual closing session was attended by the social worker and the group therapist (a psychologist or a medical doctor). In the last meeting a summary letter was written to the participant’s general practitioner. Primary outcomes in the randomized controlled trials were the number of sickness absence days and the time to sustainable RTW. The 3.5-week program reduced sickness absence compared with the outpatient program [12], while there was no evidence for an effect of the 4 + 4 days program [10] or the workplace intervention [11].

All legal residents are included in the Norwegian public insurance system. Medically certified sick leave is compensated with 100% coverage for the first 12 months (with some limitations based on salary). The first 16 days are covered by the employer, while the rest is covered by the Norwegian Labor and Welfare Administration (NAV). After 12 months it is possible to apply for more long-term medical benefits, which cover approximately 66% of one’s income.

The RRTW scale [4] consists of two parts: Part A features 22 questions to be answered by participants who are 100% sick-listed, whereas working participants (including those on part-time sick leave) answer Part B (12 questions). In the revised versions of the RRTW scale, fewer number of questions are used to define the stages of Precontemplation (items A1, A4, and A22), Contemplation (A15, A20, and A21), Prepared for action—self-evaluative (A9, A12R [reversed item scale], A13, and A18), and Prepared for action—behavioral (A6, A10, and A11) in the not working sample, and the stages of Uncertain (B8, B9, B10, B11R, and B12) and Proactive maintenance (B2, B5, B6, and B7) in the working sample (Tables 2 and 3). In the Norwegian version applied in the present study, the questions are rephrased with proper nouns (first person) [21]. The wording of two questions are also changed: from pain and injury in the original scale to health complaints in the Norwegian cross-cultural adaption of the scale [21]. All 34 items are scored on a 5-point Likert scale, from 1 (strongly disagree) to 5 (strongly agree). For 25 of the 34 items, strongly disagree indicates that returning or maintaining work is unlikely, while for the remaining nine items it indicates that doing so is likely (i.e., reversed). All participants completed the RRTW scale via internet-based questionnaires before they started the rehabilitation programs (baseline).

Sick leave was measured using data from the Norwegian Labor and Welfare Service, where all individuals receiving any form of sick leave or disability benefits in Norway are registered. The data comprised all registered medical benefits individually traceable for each participant by their social security number. Two work measures were constructed: (a) the number of sickness absence days, defined as the number of days receiving medical benefits during 12 months of follow-up after inclusion (part-time sick leave was recalculated as whole sick leave days), and (b) sustainable RTW, defined as 1 month without receiving medical benefits during follow-up.

A Rasch analysis [22, 23], was applied to evaluate the measurement properties of the RRTW scale for the working and not working subset at baseline (start of rehabilitation). The Rasch analysis was applied to all 22 items in the not working group, followed by an analysis of the 13 items included in the four subscales, and lastly, a separate analysis of the subscales of Precontemplation (three items), Contemplation (three items), Prepared for action—self-evaluative (four items), and Prepared for action—behavioral (three items). The same procedure was followed for the working group—analyzing all 12 items in the working group, followed by an analysis of the nine items generally constituting the Uncertain (five items) and Proactive maintenance (four items) dimensions, and lastly, a separate analysis of the five items of the Uncertain and the four items of the Proactive maintenance dimensions. All items initially scored on a 5-point scale were analyzed regarding the thresholds between the scoring points/levels. The scoring options were reversed for the negatively framed questions (A1, A4, A12, A16, A22, B8, B9, B10, and B12). Hence, for all items a score of 5 represented the highest likelihood of returning to and maintaining work. If the thresholds were disordered (i.e., the score levels did not separate the level of the underlying construct), the responses were re-scored. Local dependency of the items was evaluated using a correlation analysis of the residuals of the items. A coefficient of above 0.2 or below − 0.2 was chosen to indicate that the responses were dependent on each other or not fitting the Rasch model [24].

Fit to the Rasch model was investigated for the items and the individual participants. The fit of the items was statistically evaluated using standardized residuals and chi-squared statistics according to the weighted maximal likelihood method with residuals <  ± 2.5 and a non-significant chi-squared probability accepted as a fit to the Rasch model. The overall summary fit of the RRTW working and not working scales as well as the subscales was evaluated using chi-squared item-trait interaction statistics (χ²). The probability level of 0.05 was Bonferroni-adjusted according to the number of items included in the analysis. A nonsignificant probability value indicates a fit to the Rasch model [25]. The Person Separation Index (PSI) was calculated for all scales and subscales to estimate the scale’s internal consistency. PSI is the Rasch-based analogue to Cronbach’s alpha [25]. The PSI also indicates of the power of the measure to discriminate among persons with different levels of a trait. A value above 0.8 was deemed to differentiate across at least three groups of subjects. Invariance across age (dichotomized into groups below and above the median age of 46 years) and gender was assessed using a differential item functioning (DIF) analysis. DIF was assessed by an analysis of variance for each item, comparing the scores across each level of gender and age. The targeting of the RRTW scale and subscales fitting the Rasch model was evaluated by examining the hierarchical distribution of the items/statements and their response levels and was compared to the distribution of the patients along the same metric scale. The Rasch analyses were performed in RUMM 2030 (RUMM Laboratory, Perth, Australia). We needed 100 persons to detect stable item and person estimates within ± 0.5 logit with a 95% CI [26]. Generally, 10 persons per response category are recommended and assuming a not perfectly even response distribution across the 5-point scaling, we estimated a need for at least 150 persons [27]. Hence, power was not high enough to run DIF by diagnostic group, but preliminary analyses did indicate invariance across diagnostic groups. Furthermore, exploratory t-tests were performed to explore differences between persons with musculoskeletal and mental health conditions for each subscale.

Linear and logistic regressions with adjusted R² and pseudo R² were used to assess how well Rasch-based scaling (logit person location values) predicted future sickness absence and sustainable RTW. The analyses were performed without adjustments and were adjusted for age, gender, education, rehabilitation program and length of sick leave at inclusion. Age was included as a continuous variable. Education was dichotomized as high (college/university) or low. As the linearity assumption was not satisfied, the analyses were repeated with the variable categorized (≤ 0.0, 0.01–1.00, 1.01–2.00, and ≥ 2.01). The regression analyses were performed using STATA 16 (StataCorp. 2019. Stata Statistical Software: Release 16. College Station, TX: StataCorp LP).

This is the first time the RRTW scale has been evaluated according to modern measurement standards. A main strength of this study is its use of registry data for sick leave measurements, ensuring no missing data or recall bias. Another strength is the evaluation of the RRTW scale in a country other than Canada where it was developed. Some limitations should be addressed. First, some modifications were made to the wording of the original questionnaire. Changing the word injury was necessary because the word is not commonly used to describe unspecific musculoskeletal complaints like back pain in Norway. Furthermore, traditional Norwegian RTW interventions are not diagnosis specific, making it natural to replace the word pain with health complaints. We do not assume that these changes affected the scaling or difficulty of the items, yet we acknowledge the limitations related to the lack of formal validation. Second, the number of sickness absence days was not normally distributed for not working individuals. However, the analyses were also performed for sustainable RTW with similar results. It should also be noted that the interventions the participants received in the trials differed in terms of how they were designed to increase work ability. However, the Rasch analyses were based on data collected before the participants received the interventions.