Protocol for the Work And Vocational advicE (WAVE) randomised controlled trial testing the addition of vocational advice to usual primary care (Clinical Trials: NCT04543097)

Limitations of current occupational healthcare

In the United Kingdom (UK), vocational advice is variable and is often accessible only to those working for larger organisations (1). Therefore, the first port of call for most people who are struggling with their health and work is primary care, where a patient can request a ‘fit note’ to submit to their employer to allow absence from work. The fit note was introduced in 2010 (replacing the ‘sick note’) and encourages clinicians to discuss with a patient whether they are ‘not fit for work’ or ‘may be fit for some work’ with some adaptations (such as phased return to work; altered hours; amended duties; workplace adaptations), with free text space for the clinician to add details. Legislation introduced in 2022 allowed a broader range of clinicians to complete fit notes, including nurses, physiotherapists, occupational therapists, and pharmacists who can now certify a period of work absence and provide advice about working despite health conditions (2). In practice, however, most fit notes certify patients as not fit for work with just 6% using the ‘may be fit’ for work option (3). Whilst evidence is that work is generally good for health, healthcare practitioners report that they struggle to have conversations about work due to lack of confidence, lack of clear guidance around what they should be saying, and because provision of vocational advice is not embedded in general practice (4).

Improving the provision of vocational advice and support in primary care could benefit patients’ health, quality of life, and society through more active participation in the workforce (5). This is pertinent when the number of days lost from work is considered; in 2022 an estimated 185.6 million days were lost due to sickness absence in the UK (6), where approximately £72 million is spent on health-related sickness benefits per year (7). The key drivers of these costs are common illnesses such as musculoskeletal (MSK) disorders, and mental health (MH) conditions (6). Data from NHS Digital show that there is regional variation in fit notes issued across England with the northwest reporting the highest rate per 100,000 patients and London the lowest rate (8). Of concern though is that 42.6% of fit notes are issued for 5 weeks or longer (8): sickness absence is considered long-term after 4 weeks (9), and evidence shows that the longer a person is absent from the workplace the harder it is for them to return (4).

The inclusion of early work-directed interventions has been demonstrated to be effective and cost-effective for conditions such as depression and MSK pain (10–13). Where models of integrated health and occupational advice have been tested, they have led to fewer days of work absence, earlier return-to-work (RTW) and reductions in healthcare use (14–16). Several UK studies have tested interventions to manage work absence in those with health conditions, and their results have informed the WAVE trial (10,17–19).

Learning from our previous Study of Work And Pain (SWAP)(10) randomised trial and nested qualitative research with patients, vocational advisors and General Practitioners (GPs) highlighted that those with at least two weeks of work absence appeared to benefit more from the vocational advice intervention than those with shorter periods of work absence, a finding supported by other research (10,20–23). The SWAP trial offered a vocational advice intervention to patients consulting with MSK pain, and it is not known whether and how the intervention might be amended for use with a broader range of primary care patients, particularly those with common MH conditions. Parallels can be drawn between RTW processes in MSK pain and MH, and the following interventions have been shown to be effective for both types of condition: case management (24); provision of work accommodations (25); addressing obstacles to work that are clinical, psychosocial and organisational (akin to the Flags model) (26); stepped care (27); programmes tailored to individual patients (28); and telephone-delivered interventions (29). Our previous SWAP trial intervention for adults with MSK pain included these evidence-based interventions (10).

Offering vocational support early, in primary care, where most patients with health conditions are seeking healthcare, offers a key potential solution to the current lack of universal provision of vocational advice. The WAVE trial will test whether our previously developed vocational advice intervention can be adapted for all adults consulting in primary care whose health condition is impacting their ability to work.

This protocol is reported in-line with the Standard Protocol Items: Recommendations for Interventional trials (SPIRIT) reporting guidelines (30).

Aim

The overall aims of the WAVE trial are to determine, in adults consulting in general practice who receive a fit note, whether the addition of a brief vocational advice intervention to usual care leads to fewer days lost from work, and whether offering this vocational advice intervention is cost-effective.

Ethical approval

Ethical approval was granted by National Research Ethics Service (NRES) Committee West of Scotland Research Ethics Committee (REC) 5, September 2020 (REC reference: 20/WS/0127).

Trial design

The WAVE trial is a multi-centre, pragmatic, two-parallel arm, randomised controlled trial with a health economic evaluation and nested qualitative study.

Setting

This trial will take place within the primary care setting in England. General practices in the West Midlands, London and Wessex areas will participate in recruiting participants. The intervention will be delivered remotely by trained VSWs using phone or videoconferencing, with the option of face-to-face delivery.

Participant eligibility

Inclusion criteria

• - Adults aged 18 years and over,

• - Currently in paid employment (full or part time),

• - Current absence from work of at least two consecutive weeks but not more than six continuous months,

• - Received a fit note,

• - Access to a mobile phone that can receive and respond to SMS text messages,

• - Able to read and write English,

• - Able to give full informed consent,

• - Willing to participate.

Exclusion criteria

• - Long-term work absence defined as over six continuous months,

• - Pregnancy or maternity leave,

• - Patients presenting with signs or symptoms indicative of serious illness requiring urgent medical attention (‘red’ flags),

• - Severe mental health problems (e.g. severe depression with risk of self-harm, exacerbation of schizophrenia or bipolar disorder, cognitive impairment, or lack of capacity), high vulnerability (e.g. palliative stages of illness, recent bereavement, dementia).

Participant identification

Potential participants will be identified when they consult at one of the participating general practices and are issued a fit note for time off work. General practices will choose one of three methods to identify and invite potential participants. Figure 1 provides a flow chart of how participants progress through recruitment.

• Download figure
• Open in new tab

Figure 1:

Participant recruitment

Participant recruitment

Figure 1 details the participant identification and recruitment processes. Where patients have been sent a study pack, they will be invited to answer:

• whether they are still absent from work (Yes / No),

• whether the duration of their absence is more than 2 weeks (Yes / No),

• whether their absence is less than 6 months (Yes / No),

• whether the fit note is for their employer (Yes / No) (to exclude those issued a fit note for employment support allowance).

Patients who respond ‘No’ to any of these questions will be advised not to complete the remainder of the questionnaire. Participants who respond “Yes” to all the questions will be asked to complete the consent form and baseline questionnaire and return these to Keele Clinical Trials Unit (CTU).

Consent

The invitation letter will introduce the study and explain, or remind them, how they were selected to be invited. The Participant Information Sheet (PIS) will summarise the study and tell the patient what is involved should they wish to participate. The contact details of Keele CTU will be provided should potential participants have any further questions about the study or have any difficulty in completing the consent form or baseline questionnaire.

Those who wish to participate will be asked to complete, sign and date a consent form confirming that: they have read and understood the PIS and are willing to take part in the study; understand that a questionnaire will be sent at 6 weeks; consent to receive a fortnightly SMS text message to collect data on RTW for up to 6 months; that they understand and consent to randomisation; and that they are aware they can withdraw at any time without giving a reason and that if they do withdraw that their clinical care will not be affected.

Participants may also consent to the optional aspects of the trial, again signing and dating the consent form to confirm that they: may be invited to participate in an interview; understand that their GP will be notified about their participation in the study; were asked to consent to a pseudonymised electronic copy of relevant sections of their general practice medical records to allow authorised members of the research team to extract information relevant to the study.

The returned consent form will be checked for completion. Recruitment will be complete when the returned questionnaire confirms the patient is eligible, they have signed and dated their consent form and completed the primary outcome data. Any missing data from the consent form, eligibility questions or primary outcome, will be followed up by post, telephone or email from Keele CTU.

Randomisation and allocation concealment

On confirmation of eligibility and receipt of a consent form participants will be randomised by Keele CTU to either usual care or usual care plus the vocational advice by computer-generated stratified block randomisation (ratio 1:1). Stratification will be by centre (West Midlands, London, Wessex) and main health condition resulting in time off work (MH, MSK pain or ‘other’ health condition). All participants will be mailed a letter informing them whether they have been randomised to either the usual care or to usual care plus the vocational advice intervention. A letter will be sent to the GP in participating practices of all randomised participants informing them of their patient’s participation and allocation.

Participants, their treating clinicians, and VSWs cannot be blinded to allocation due to the nature of the intervention. Keele CTU staff who may need to contact participants for minimum data collection by telephone will be blinded to allocation. The data will be analysed independently by two statisticians, one of whom will be blinded to intervention allocation; the other statistician will be unblinded to allow intervention delivery details and content of case report forms (CRFs) to be reported to the TSC / Data Monitoring Committee (DMC) as required.

Description of intervention and control

Intervention arm

Participants who are randomised to the intervention will be offered the vocational advice intervention in addition to usual care. The intervention will be a modification of that developed and successfully delivered in the SWAP trial (10), comprising a stepped care intervention based on the principles of case management (31). A full description of the content, delivery and training associated with the intervention will be published separately. Figure 2 presents a summary of each step of the intervention.

• Download figure
• Open in new tab

Figure 2:

Overview of the intervention

Sample size

The trial is powered to detect a 25% reduction in days off work over 6 months between the intervention and control arms, equating to an Incidence Rate Ratio (IRR) of 0.75 (e.g., mean days off work reduced from 30 days in the control arm to 22.5 days in the intervention arm). A sample size of 720 gives 80-90% power to detect an IRR of 0.75 based on a 5% two-tailed significance test and assumed dispersion parameter of 1.4 (derived from the previous SWAP trial) allowing for 20% loss to follow-up (10).

Primary outcome measure

The primary outcome measure is number of days off work over 6 months and will be assessed by asking participants to report how much time off work they have had due to their health condition. See Table 1.

Secondary outcome measures

Participants will be asked every 2 weeks whether they have returned to work, for a period of 6 months or until a sustained RTW is achieved (defined as return to any work for at least 4 weeks). RTW will be measured via SMS text message using the following questions;

• lr Have you returned to work? Yes/No,

• lr If yes, on which date did you return to work e.g. 13SEP2021.

Work interference will be measured using the Work Productivity Activity Impairment (WPAI) questionnaire (32). The WPAI measures impairments to work and activities in the past 7 days and has been validated in many health conditions including MH, MSK pain, respiratory, digestive cardiovascular and other conditions. A single item work productivity question from the WPAI, asking participants to what extent their health condition has impacted their performance at work will also be used as a secondary outcome measure.

Where mediators are collected at 6 months they will also be treated as secondary outcome measures.

Moderators and mediators

The questionnaires will include anticipated mediators that have been included within the initial logic model underpinning the intervention. These mediators have been selected based on published evidence indicating that they are important in the relationship between health and work, and they are modifiable through interventions like vocational advice. The measures include personal (physical and mental health, and behaviours) and occupational measures and are summarised in Table 1.

Electronic Medical Record (EMR) review

In addition to the data collection from the self-report questionnaires and SMS text messages, participants will have the option to consent for the research team to access their general practice EMR data for the study duration to allow examination of fit notes issued. Medical records for consenters will be extracted electronically and will be pseudonymised at source by the EMR system. Records will be exported by general practice staff with support from Clinical Research Network staff, as required.

Internal pilot

The internal pilot trial will assess recruitment and intervention fidelity over the first four months and be focused on the proportion with follow-up data at six weeks. The anticipated recruitment rate will be approximately four patients/practice/month in the internal pilot and aim to achieve 80% follow-up for the primary outcome at six months. We will scrutinise (i) recruitment (target versus actual, % eligible, % consenting and randomised), (ii) intervention fidelity, and (iii) follow-up rate at six weeks. A ‘stop (Red)/ amend (Amber)/ go (Green)’ set of progression criteria including: (i) recruitment uptake ≤70% of those eligible and consent to participate (Red), 71%-99% (Amber), 100% (Green); (ii) engagement with the intervention, % of intervention arm participants who have at least one contact with a VSW <40% (Red), 40-65% (Amber), >65% (Green); (iii) primary outcome data at six weeks follow-up rate <60% (Red), 60-80% (Amber), >80% (Green). A decision to continue to the main trial will be made if all progression criteria are met at the ‘Green’ level, and to continue but with some adjustments if any criteria are at least ‘Amber’; the trial may be stopped if any of the criteria are ‘Red’ and the Trial Management Group (TMG), Trial Steering Committee (TSC) and Data Monitoring Committee (DMC) agree they cannot be addressed.

Main trial

Baseline participant characteristics will be summarised according to the nature of the data (mean/standard deviation for normally distributed variables; median/inter-quartile range for skewed numerical data; frequency/percent for categorical variables) – overall, and by treatment arm (no formal statistical testing will be carried out). An intention-to-treat analysis will be carried out as the main approach: analysing participants as per randomised allocation. This is in line with the pragmatic nature of the trial – allowing for infrequent referral to occupational health within both arms and a lack of contact with the vocational support intervention in a small proportion of participants in the intervention arm.

For the primary outcome measure of days off work, we will present descriptive statistics on both mean (and median) number of days off work with standard deviation (and interquartile range) – for the time intervals baseline to six weeks and six months follow-up, by trial arm. The inferential analysis will be carried out by negative binomial (or Poisson) regression models adjusting for age, sex, centre, main health condition for which the fit note was issued (MH, MSK, other) and time off work due to health condition in the months prior to trial participation (fixed effects). If there is over-dispersion (skewness) in the outcome data, then the negative binomial model will likely be preferred to the Poisson model; the goodness of fit of each model will be assessed to determine which model is most appropriate through scrutiny of the likelihood-ratio test and Akaike/Bayesian Information Criterion (AIC/BIC). The estimated effect will be presented as an incidence rate ratio; the results will be given as point estimates with 95% precision interval and associated p-value. The primary endpoint evaluation will be number of days off work over the six months follow-up, with days off work over the initial six weeks follow-up as a secondary endpoint.

Similarly, descriptive summaries of data (mean (SD) / median (IQR) / frequency count (percent)) by trial arm will be presented for secondary outcomes. Proportions of participants accessing each step of the intervention (1, 2 and 3) and the content of the intervention as detailed on CRFs will be reported for the intervention arm. A mixed-model approach will be carried out for between-arm estimation of mean differences (for numerical outcomes) or odds ratios (for categorical outcomes) through linear or logistic link functions, respectively. The regression models will include the same covariates as outlined above for the primary outcome analysis and additionally the corresponding baseline value (as appropriate) e.g. baseline RTW self-efficacy score for evaluation of between-arm difference in mean RTW self-efficacy at follow-up. Time to sustained RTW data, collected by fortnightly SMS text messages, will be evaluated through survival analysis methods: life table and Kaplan-Meier descriptive summaries and Cox regression modelling with covariates as detailed above.

A complier average causal effect (CACE) evaluation will be undertaken to obtain unbiased estimation for the comparison of the primary outcome for those participants in the intervention arm who had at least one contact with the VSW versus similar participants in the control arm. A small number of pre-specified subgroup analyses will be carried out evaluating whether between-arm differences in the primary outcome measure (number of days off work over six months) contrast across potential baseline moderators: baseline main health condition resulting in the fit note (MSK, MH, other) and duration of work absence in previous six months. Statistical estimates will be obtained through including interaction terms for trial arm × baseline subgroup within the statistical model of treatment effect. All statistical analyses are focused on superiority testing based on 5% two-tailed significance level.

Analyses will be carried out blind to intervention allocation (with the exception of the CACE analysis and CRF data analysis describing the content of the vocational advice intervention) and double-analysed by two statisticians. Data collection, checking and verification will be performed according to Keele CTU SOPs to ensure rigour. A detailed Statistical Analysis Plan (SAP) will be developed with guidance from the DMC and TSC to ensure transparency in the statistical analysis of the trial.

Mediation analysis

Mediation analysis is a statistical approach for testing hypothesised causal pathways between variables thought to be important in explaining treatment outcome. Testing such causal mechanisms is an important aspect of process evaluation (35). Variables hypothesised to be key mediators of the intervention (reported in Table 1) will be finalised through the intervention adaptation in the feasibility phase. In the trial, data on these variables and on the primary outcome will be collected at baseline, 6 weeks and 6 months. Change in each mediator between baseline and each follow-up point will be described. Multilevel causal modelling techniques will be used to identify the proportion of the intervention effect on days absent from work explained by change in the potential mediators (indirect effects). Each mediator will be first analysed separately, then combined into a single multiple mediation model to assess the combined effect of the potential mediator variables on the outcome of days absent from work over six months. Latent growth curve models will be used for this analysis as they allow all three time points to be used within the analysis. These indirect effects on outcome will be expressed as a regression coefficient with bootstrapped 95% CI.

Analysis

All interviews will be audio-recorded, fully transcribed and anonymised. An inductive, exploratory framework will be adopted using thematic analysis, and influenced by grounded theory(46). First, a sample of early transcripts will be independently coded by two researchers with experience in qualitative analysis, and a coding framework agreed upon, which will be applied in subsequent coding. Coded data will be analysed by the qualitative social science researcher and a second research team member independently to develop categories and themes for discussion. The constant comparison method (47) will be used in the analysis, looking for connections within and across interviews, and across codes, highlighting data consistencies and variation. Analysis will be an iterative process, with emergent findings used to further refine topic guides for subsequent interviews. Comparisons will then be made using a framework approach (48) between the experiences of participants with different health conditions looking for similarities and differences in the separate accounts particularly related to how the methods of identification, recruitment and intervention delivery can be improved. We will also use the acceptability framework of Sekhon et al (49) to sensitise the analysis.