Study design and setting
The study has a two-armed cluster randomized controlled design with a primary follow-up period of 1 year (see Fig. 1). The study is conducted in collaboration with three OHSs in Sweden. One OHS has nationwide coverage while two are regional OHSs. Each participating OHS consultant is considered as one cluster. The participating OHS consultants consist of occupational nurses, physicians, physiotherapists, ergonomists, occupational health counselors and psychologists.
Ethical approval
Ethical approval has been granted by the Ethical Review Board in Stockholm (registration number 2015/549–31/1). The study complies with the ethical guidelines regarding voluntary participation, privacy and the handling of personal data in accordance with Sweden’s Personal Data Act and Secrecy Act. Informed consent will be obtained from all participants and will include, amongst other things, information about the right to opt-out from the study at any time without giving a reason and without any consequence for the help received from the OHS (see also Recruitment of participants).
Recruitment of participants
The OHS consultants provide preliminary information about the study to employees seeking care at the OHS for work-related stress symptoms or CMDs which fulfil the inclusion criteria. They then ask whether the employee’s contact information may be sent to the research group at Karolinska Institutet (KI). Written consent to transfer this information to KI is obtained from those who are preliminarily interested in participating in the study. Subsequently, members of the research group contact the employee and give both verbal and written information (sent by email) about the study and what participation would entail. Final consent is obtained and study inclusion achieved when the employee has submitted the baseline questionnaire and confirmed having taken part of the information about the study by ticking a box at the end of the baseline questionnaire. Paper-and-pencil questionnaires are offered to those who prefer this option. Recruitment of participants began in September 2015 and will be completed during the spring 2017.
Inclusion criteria
Inclusion criteria are met by employees fulfilling the following criteria:
-
The employee seeks help at the OHS for occupational stress affecting their ability to work. If the employee is on sick leave due to adjustment disorders, anxiety or depression, the period of sick leave should not have exceeded 3 months at the time of the first OHS visit.
-
The employee should accept involvement of the supervisor in the intervention.
-
The employee also needs to understand both written and spoken Swedish as the study questionnaires and text messages are in Swedish.
Exclusion criteria
Employees will be excluded from the study if they are pregnant at the time of inclusion, victims of workplace bullying, have post-traumatic stress-disorder, severe mental illness (e.g. psychosis) or any other co-morbidity that could considerably affect their ability to work and/or quality of life.
Randomization
The study uses a two-armed cluster randomized design in which the participating consultants are randomized using computer-generated random numbers into either giving the intervention or performing care as usual. Hence, each OHS consultant has to continuously apply the same treatment within the context of this study. The randomization is stratified by OHS unit as well as by occupational category, as those two factors are assumed to have a potential impact on the results and a balance within strata is necessary.
It was not feasible to randomize employees between OHS consultants trained or not trained in the intervention, because different employers have contracts with different OHS consultants, that is, a certain employee is linked to a certain OHS consultant. An alternative may have been to train all OHS consultants in the intervention, however, this should have increased the risk of contamination.
Blinding
The employees are not given information about the possibility of receiving another type of treatment within the trial. This is because the OHS consultants have been pre-randomized into delivering one of the two interventions. The researchers will not know about treatment allocation during the analyses. However, the OHS consultants are not blinded with regard to which treatment is given.
Matched control group
An additional comparison group for those on sick leave at baseline will be obtained by means of register-based data from the SSIA. The material will be matched on the basis of diagnosis, age, sex and sick leave during the previous 3 months. This data allows for a better understanding of the development patterns of sick leave for the study participants on sick leave in both the control and the intervention groups.
The individual and workplace-focused intervention
The OHS consultants delivering the experimental condition receive general information about the study and information about the recruitment process. After this, a 1-day training course is given by members of the research group and a clinical psychologist. Individual supervision and follow-up is thereafter provided on a continuous basis by the KI research group by email and scheduled phone calls. The OHS consultants are also offered further guidance for the intervention at 1–2 follow-up meetings after the initial training (on demand). They also receive detailed work sheets.
The theoretical basis for the intervention stems from PST [22] and the “mismatch” model concerning the match between the employee and the work environment [23]. This mismatch model emphasizes six aspects of the work situation which are addressed during the meetings: workload, control, reward, community, fairness and values. A participatory approach is applied by which the supervisor and the employee are guided by the OHS consultant and encouraged to actively take part in problem solving concerning the work situation [14, 24].
As outlined in Fig. 2, the intervention comprises two interviews, one of which is held with the employer (usually the nearest supervisor) and the other with the employee, and a third session at which both the employer and the employee are present [14, 24]. This third dialogue aims to facilitate a closer match between the employee’s expectations and abilities and the job situation. The order in which the two separate interviews are held depends on the internal logistics at the OHS.
The interview with the employee’s supervisor covers any earlier measures taken by the employer for the employee and the supervisor’s view of the “causes” of the employee’s stress or poor mental health. The supervisor is also asked about what he/she considers to be the most important problem to address to promote the employee’s work ability or RTW. For practical reasons this meeting usually takes place over the phone and lasts about 15 to 30 min. The same themes are discussed during the interview which takes place in person between the OHS consultant and the employee. However, this session is more comprehensive and also includes an initial section concerning the private situation of the employee, work-family balance, and a survey of possible mismatches between the employee and the job situation [23]. A problem-solving approach is also initiated. This interview takes approximately 90 min.
The problem-solving approach is guided by the following steps (1) goal setting and problem definition – the problem is specified in detail including when and where it happens as well as how it is expressed, (2) generation of possible solutions, (3) decision making about solutions, including potential consequences of different solutions, (4) setting up an action plan, including implementation of prioritized solutions.
During a third meeting the supervisor, employee and the OHS consultant meet face-to-face and focus on problem definition and possible solutions for promoting work ability or RTW. At this meeting the supervisor’s and the employee’s views of the problem are discussed with the aim of achieving a convergence of views. Relevant external parties such as an OHS physician or a representative of the Social Insurance Agency may also participate in this meeting if deemed necessary. The main task of the OHS consultant during the third meeting is to help the employee and his/her supervisor to reach specific problem definitions and feasible solution(s) to be implemented at the workplace to improve work ability or facilitate RTW. The suggested solutions and their follow-up are registered in an action plan by the OHS consultant. It is crucial that both the employee and the supervisor are encouraged to understand, and to be active and participatory in, the problem-solving process and that consensus between the parties is reached on possible solutions to facilitate RTW or manage stress at the workplace.
The OHS consultant is recommended to conduct regular follow-ups of the supervisor and the employee by phone calls or meetings in accordance with an agreed schedule. At least three follow-ups during the 3 months following the third session are recommended. These follow-ups focus on the action plan and whether planned measures have been undertaken; progress and possible new problem solving. In addition to the planned intervention, the OHS may also refer the employee to other services or measures in accordance with routine practice.
Care as usual
The OHS consultants delivering the control condition receive information about the recruitment process followed by a general introduction to research into psychosocial factors and mental health at work. In total this takes 2 h.
The OHS consultants then continue to deliver the routine care throughout the study period. At the three OHSs where the current study is being performed, care as usual in case of stress-related symptoms and CMDs usually involves both the employee and the supervisor. However, these practices do not follow the same structure as in the experimental intervention nor do they use the mismatch model as guidance for defining problem areas or the described problem-solving approach. At one of the OHSs the intervention is more focused on further medical and lifestyle factors, even though the work situation is also addressed and the employer is involved in the intervention. There is also some variance among the OHS consultants as to how this intervention is applied. One of the OHSs does not usually hold an individual interview with the employee; in other words, the joint meeting between the employee, the employer and the OHS consultant is often held directly after a short interview with the supervisor. A third OHS is about to implement the Swedish guidelines for OHSs concerning the management of mental health at work. These guidelines recommend the involvement of the employer and problem-solving techniques. However, despite the similarities with the experimental intervention the personnel delivering the care as usual have not received the same training in these methods as the consultants delivering the active intervention. A survey of the OHS consultants regarding more specific details of the contents and routines of care as usual will also be conducted as part of the study’s process analysis.
Co-interventions
As it is difficult to avoid the possibility of co-interventions, the questionnaires sent out to the employees ask for information about supplementary treatment from the OHS or other agencies. The questionnaire assesses the type of intervention or help that was sought as well as the number of times it was received and whether it was administered by the OHS or others.
Measurements and procedure
Data are gathered by use of registers, questionnaires and SMS text messages. Register data are collected from the SSIA. These data are based on register information, including sick-listing and sickness and activity compensation. Valid questionnaires in electronic format are administered at baseline and after six and 12 months (paper versions are available on demand). Furthermore, monthly (every fourth week) text messages asking five questions about sick leave, stress and work ability obtained from the employee are sent to the participants over a 1-year period [25]. The use of text-messages is motivated since this measurement also includes short spells of sick leave (<15 days) which are not covered by the SSIA. A 3-year follow-up of register-based data on sick leave is also planned.
Primary outcome
Registered sickness absenteeism
The primary outcome is the total number of days of sickness absence during the 12-month follow-up period (including sickness benefit and disability pension). Sickness benefit and disability pension will also be analyzed separately.
Secondary outcomes
Self-reported sickness absence and return to work using text messages
For those on sick leave at baseline, the time to RTW will be calculated from baseline until (1) the employee returns to work in any increased level, or (2) full return to work (working usual hours during an uninterrupted period of at least 4 weeks). The prevalence of no sick leave, part-time sick leave and full-time sick leave during the 12-month follow-up period following baseline will be presented.
Registered sickness absenteeism
The total number of days of registered sickness absence from the SSIA will be calculated in the same way after 3 years as at the 1-year follow-up.
Mental health and stress-related symptoms
Self-reported exhaustion will be assessed by the Emotional exhaustion sub-scale of the Maslach Burnout Inventory - General Survey (MBI-GS). The MBI-GS is a well-established instrument for assessing burnout [26]. Furthermore, stress-related Exhaustion Disorder is assessed by the s-ED measurement instrument [27]. The s-ED is based on diagnostic criteria for stress-related Exhaustion Disorder according to the Swedish National Board of Health and Welfare. Support has been found for its constructive and predictive validity [27]. In addition, the Hospital Anxiety and Depression Scale will be used to assess anxiety and depression among the employees [28, 29]. Finally, self-reported stress as assessed by a single item included in the text messages will be used. This has been found to be valid for monitoring stress in a work-related context [30].
Health problems
Sleeping problems will be measured by the insomnia subscale from the Karolinska Sleep Questionnaire [31]. Furthermore, the European Quality of Life - 5 Dimensions questionnaire (EQ-5D) will be applied to investigate health-related quality of life [32]. Self-perceived general health will be assessed by a single question (Orwelius et al. In manuscript). Presenteeism (being at work while sick) will be assessed by a single question developed by Aronsson et al. [33].
Work ability and working environment
Self-reported work ability will be assessed by three items from the Work Ability Index [34]. Two of the items relate to perceived capacity to work relative to the physical and mental demands of the work. The third item enquires whether the employee believes that s/he can be working at the same workplace in 2 years’ time. Two additional items are intended to measure work performance impairment due to (1) health problems and (2) problems in the working environment. These items have been developed and modified inspired by the Work Productivity Activity Impairment questionnaire - General Health Questionnaire (WPAI-GH) [35, 36]. The second item has been developed and evaluated in Sweden [35,36,37]. Finally, job satisfaction will be estimated by a single self-report item [38].
Prognostic measures
The demand-control-support model [39] will be considered as a possible prognostic measure in this study. The study will also include items from the General Nordic Questionnaire (QPSNordic) relating to ongoing conflicts with the superior, perceived loss of control over work tasks and conflicts between the employee’s values and how the work is actually carried out [40]. One self-efficacy item about how confident participants are about whether they will be back to their usual working hours in their usual work position 3 months after the baseline measurement will be developed and validated within the project. Finally, the OHS consultants will be asked a similar question about whether the employee will be back to his/her usual working hours in 3 months’ time.
Cost-effectiveness
A full economic evaluation of costs for society and for the employer will be conducted [41]. The expenses that will be collected are direct costs, such as intervention costs paid by the employer and possible rehabilitation provided by other caregivers. Indirect costs such as time used by participants for the intervention and travel time will also be collected. Time for the intervention will be calculated on the basis of the intervention program. Time for travel to attend the meetings will be standardized at 1 h per meeting (30 min/single way). Treatment costs will be estimated using national unit costs and the fee charged by the OHS. Lost time due to treatment or travel will be evaluated using national median wages. Information about absenteeism and reduced performance due to health-related and work environment-related problems, i.e. productivity loss, will be collected using validated questions and used to calculate potential benefits of the intervention. Productivity loss will be estimated using national median wages.
Data regarding the direct costs and potential savings which emerge from the intervention will be collected in parallel with the implementation of the project and for a 1-year follow-up period. All costs and consequences will be converted to a single year using the consumer price index. Discount of costs or consequences is not needed for the 1 year follow-up because of the short follow-up period. Costs or savings that occur during the long term follow-up will be discounted using a 3 % discount rate [42]. The employer’s economic evaluation will be conducted using a cost benefit analysis; the evaluation of the economic costs for society will be conducted using a cost effectiveness analysis.
Process evaluation
The importance of a process evaluation for studies conducted within the OHS has been highlighted in previous research [43]. The current study will perform a process evaluation of both the OHS and the employee [44]. The purpose of this is to examine: the attitudes of the OHS consultants towards the study; the organizational support for the study (on the part of the OHS-consultants); the content of the control condition and the fidelity of the OHS-consultants to the experimental condition.
The participating OHS consultants´ attitudes towards and knowledge of the research project will be evaluated using a short questionnaire at the start of the study. Examples of items are “To what degree do you feel it is good to be part of the project?” or “To what degree do you understand the aim of the project?”
The OHS consultants in the experimental condition will be asked to rate their overall fidelity to the method for each employee during the study (this is done directly after the joint meeting with the supervisor and the employee). Telephone follow-ups with each participating consultant are conducted (both for the experimental and the control group) during the recruitment (and intervention) process. These follow-ups include questions about recruitment, such as clarifications of the inclusion criteria, as well as logistical or contextual factors related to recruitment. For the consultants delivering the experimental condition, issues concerning the feasibility of the intervention and the fidelity of the OHS consultant are examined. The meetings are summarized in short notes.
After the recruitment period is finished, fidelity to the intervention among the OHS consultants delivering the experimental intervention will be assessed by means of check-lists that will be built around key aspects of the intervention. Examples of such variables are whether contact with the workplace has been maintained and whether employees have been active in the problem-solving process and the planning of future measures. The content and fulfillment of the recommended follow-ups with employees and supervisors will also be addressed. The OHS consultants in the control condition will be given a check list where they are asked to describe the content of the characteristics of the control condition (intervention).
For all OHS consultants in the experimental and the control condition, the check lists will also include questions about the recruitment process (for instance, whether all employees who should be informed about the study actually received this information). Furthermore, organizational-level variables will be addressed, such as how the research project was introduced to the OHS consultants by the management at the OHS and how the study was supported while it was in operation.
Employee satisfaction with the intervention or treatment as usual will be assessed 6 months after the completed intervention. The assessment will consider aspects such as the quality of communication with OHS personnel, the relevance of the intervention, problem-solving skills learned during the intervention, planned adjustments at work, influence on the planned adjustments, agreement with the supervisor, the implementation of planned adjustments, whether follow-up contacts with the OHS were held, and questions about treatment satisfaction.
Data analysis
Statistical methods adapted for cluster randomized designs will be applied. The plan is to apply various forms of regression models and mixed models. Intention-to-treat analyses will be carried out. Where relevant, per-protocol analyses will also be performed. If it appears that potential confounders have been unevenly distributed and can be expected to affect the results when the two interventions are compared, these factors will be adjusted for in the analyses. Possible interaction effects on the outcome for (1) gender x treatment (intervention) and (2) sick leave status (sick leave/no sick leave) x treatment will be checked for. If they are statistically significant, subgroup analyses will be considered.
Statistical power
Sample sizes of 75 in Group 1 and 75 in Group 2, which were obtained by sampling 15 clusters with an average of five subjects each in Group 1 and 15 clusters with an average of five subjects each in Group 2, achieve 84% power to detect a difference between the group means of at least 25.00 days of registered sickness absence 1 year after baseline. The standard deviation of subjects is 50.00. The intra-cluster correlation coefficient is 0.010. The coefficient of variation of cluster sizes is 0.650. A two-sided t-test was used with a significance level of 0.050. This test used degrees of freedom based on the number of subjects. No missing values are expected since the primary outcome is based on register data. The aim of the current study is therefore to include a total of 150 participants.