Occupational advice to help people return to work following lower limb arthroplasty: the OPAL intervention mapping study

The impact of hip and knee osteoarthritis on employment

Decreased physical function associated with hip and knee osteoarthritis reduces the likelihood of employment, reduces household income and increases the number of missed workdays for those who are employed. 1 The magnitude of the impact varies according to the degree of activity limitation and disease severity. 2 A diagnosis of hip or knee osteoarthritis is associated with a reduction in work participation and productivity and an increased risk of work loss. 3,4 In a national study of patients in Finland, Kontio et al. 5 found that the age-adjusted incidence of disability retirement owing to knee osteoarthritis was 60 and 72 per 100,000 person-years for men and women, respectively. The highest rates of disability retirement in men were found in construction workers, electricians and plumbers, and in women they were found in building caretakers, cleaners, nurses and kitchen workers. 5

The cost of work-related musculoskeletal disorders that have an impact on a person’s ability to work is difficult to quantify. Direct (the cost of treatment) and indirect (costs related to the impact of the period of ill health) costs are borne by the individual (impact of ill health on quality of life), employers and society (loss of productivity, need for health care, rehabilitation and compensation). 6,7 The Health and Safety Executive calculates the annual cost of workplace injury and ill health on this basis by estimating both the financial cost and the ‘human’ cost. 7,8 The Health and Safety Executive estimates that the total annual cost of workplace ill health due to musculoskeletal disorders is £9.7B, equivalent to £18,400 per case. 7,8 However, these figures do not take account of non-work-related injuries and ill health and, therefore, are likely to be an underestimate of the total cost. In addition to its financial benefits, working has significant physical, mental and emotional health benefits. 9–11 Loss of employment is associated with a reduction in physical function, increased anxiety and depression and increased risk of mortality. 12,13 Therefore, earlier return to work (RTW) has potential health benefits as well as socioeconomic benefits.

The role of lower limb joint replacement in patients of working age

Lower limb joint replacements are successful and cost-effective treatments that relieve pain, restore physical function and improve health-related quality of life for patients with hip and knee arthritis. 14–17 Currently, > 1 million hip and knee replacements are carried out annually in the USA and > 190,000 are carried out annually in England, Wales and Northern Ireland. 18 Projections from 2005 suggest that, by 2030, the number of primary total hip replacements (THRs) and total knee replacements (TKRs) carried out will increase by 174% and 673%, respectively. 19

Recent changes to the state pension age, combined with an ageing UK workforce, resulted in a steady increase in the numbers of hip and knee replacements being carried out in patients of working age between 2007 and 2017. 18 These changes are also reflected in data from North America, which suggest that over half of all hip or knee replacement procedures will be carried out in patients aged < 65 years by 2030. 19 International estimates suggest that between 15% and 45% of patients undergoing either hip or knee replacements are of working age. 20,21

According to data published by the National Joint Registry (NJR) for England, Wales and Northern Ireland, 106,334 primary knee replacements22 and 96,717 primary hip replacements23 were carried out in 2017. Of the 91,923 hip replacement patients with available patient data, 18,812 (20.5%) were aged < 60 years (9778 females and 9034 males) and a further 26,295 were aged 60–69 years (28.6%; 15,375 females and 10,920 males). Of the 102,347 knee replacement patients with available patient data, 17,765 (17.4%) were aged < 60 years (10,259 females and 7506 males) and a further 33,523 were aged 60–69 years (32.8%; 18,161 females and 15,362 males).

Occupational advice for patients undergoing hip or knee replacement

There is currently a paucity of information and guidance to support patients returning to work after hip or knee replacement. Over the last 2 years [during the course of the Occupation advice for Patients undergoing Arthroplasty of the Lower limb (OPAL) study], the Royal College of Surgeons of England has produced written information resources to guide recovery including RTW after both hip replacement and knee replacement (example information is available at www.rcseng.ac.uk/patients/recovering-from-surgery/total-hip-replacement/returning-to-work; accessed 16 April 2019). We are not aware of any other currently available occupational advice or information resources specifically tailored to this patient group.

The UK Government currently funds the ‘Fit for Work’ service (in Scotland, the service is called ‘Fit for Work Scotland’). 24 This initiative is free for the public to use and is designed to support people in work with health conditions and help with sickness absence. It works alongside existing occupational health services and employer sickness absence policies. Patients can access this service via telephone line support, by visiting the ‘Fit for Work’ websites or by e-mailing the team. However, the patient-facing materials are generic and there is no specific information for hip or knee replacement patients. 24

Population

There is limited evidence about the population of patients undergoing hip or knee replacement who are in work and returning to work after surgery. Further information is required to understand the individualised workplace needs of this group, including an understanding of how job classifications (e.g. manual vs. non-manual), employment status (e.g. employed vs. self-employed), the type of employer (e.g. small and medium-sized enterprises vs. large companies; public- vs. private- or third-sector employer) and how the presence of an occupational health service within the organisational structure influences the potential for early RTW.

The target population for a clinical trial is, therefore, not clearly defined.

Intervention

• Current recommendations guiding RTW are limited and inconsistent. Information is rarely individualised and generic information often fails to provide the patient, employers or health-care teams with the advice required.

• The majority of patients undergoing hip or knee replacement undertake an integrated multidisciplinary team programme of education and rehabilitation spanning the surgical episode. The provision and utility of occupational advice within these ‘usual-care’ pathways is not established and the ability of this service to facilitate RTW has not been explored.

• Studies suggested that the vast majority of ‘fit notes’ are not being used correctly. ‘Fit notes’ offer the patient and employer opportunities for early phased RTW. However, most are advising that patients are ‘not fit’ for work, with few doctors making use of the opportunity to advise on patient function and/or work modifications that might facilitate RTW after surgery. 50,51

• There is limited information about the impact of addressing modifiable barriers that prevent RTW or how modifiable psychosocial factors influence RTW behaviours and the specific needs of the patients regarding perioperative care and advice. 48,52

There is, therefore, no appropriate occupational advice intervention available that could be used as the intervention in a clinical trial.

Comparison

There is no information about how and when occupational advice is delivered and who is delivering it to hip and knee replacement patients. The rapid and inconsistent adoption of enhanced recovery and early-discharge pathways has led to variations in provision of perioperative care and advice.

‘Standard care’ is therefore not currently defined for use as a study comparator.

Outcome

• There is currently no standardised method of recording RTW. Dichotomous recording of work status (yes/no) is blunt and does not address important aspects of workplace behaviour including absenteeism, presenteeism, return to usual activities and interference with activities. In the UK, > 20% of patients do not return to usual activities and have restrictions in their ability to work after hip replacement. 39 Measuring RTW should ideally consider specific elements of the job, the duties and the hours worked.

• Assessment of workplace disability and productivity is poorly reported after hip or knee replacement. Validated tools exist (e.g. the Workplace Activity Limitations Scale and Work Limitations Questionnaire) but little is known about their applicability to the UK workforce and their utility as outcome measures for clinical trials. 53

The appropriateness of individual RTW measures for use as primary outcome measures in a clinical trial is currently unclear.

Summary

There is a need for preliminary research to generate relevant evidence and develop an occupational advice intervention to support a future clinical trial. The OPAL study was commissioned to facilitate this.

Intervention mapping stage 1: needs assessment (see Chapters 3–6)

Intervention mapping stage 1 established the rationale for an occupational advice intervention within the target population by evaluating the discrepancy between current and desired practice. It utilised a variety of approaches including a rapid evidence synthesis (see Chapter 3), a cohort study (see Chapter 4), a national survey of practice (see Chapter 4) and patient (see Chapter 5) and stakeholder (see Chapter 6) interviews. This information was then used to create a logic model of the problem considering how the behaviours of the target population increase the risk, prevalence, incidence and burden of the problem and how interpersonal, organisational, community and societal factors influence RTW directly or through influence on the behaviour of the target population. These behavioural and environmental factors were then mapped to specific theory- and evidence-based factors and determinants to help provide an overview of the problem and a framework to address it.

Intervention mapping stage 2: identify intended outcomes and performance objectives (see Chapter 7)

Stage 2 used the findings from stage 1 to specify who and/or what needs to change for patients to make a successful RTW following hip/knee replacement. A provisional matrix of POs for key stakeholder groups was constructed outlining the personal determinants, external determinants and expected outcomes for each objective.

Intervention mapping stage 3: selecting theory-based methods and practical strategies (see Chapter 7)

In stage 3, a list of possible components matched to each performance objective/determinant was generated. Using theory, evidence, experience and consensus, the most practical ways to implement these interventions were identified. These intervention ‘components’ formed the basis of the statements presented to stakeholders as part of the Delphi consensus process (see the next section on IM stage 4) and helped to develop the first iteration of the developed occupational advice intervention.

Intervention mapping stage 4: development of intervention components (see Chapters 8 and 9)

Stage 4 used the information and associated occupational advice strategies identified in the first three IM stages to develop specific tailored tools and materials. To help refine these components, a multistakeholder intervention development group was created to reach agreement about the design, content, delivery, format and timing of the proposed occupational advice intervention. To facilitate this process, a modified three-round Delphi consensus process was employed. Information from the Delphi consensus process was then used to refine and finalise the occupational intervention.

Intervention mapping stage 5: adoption and implementation plan (see Chapter 10)

In stage 5, strategies for the implementation and adoption of the intervention were developed. This stage ran concurrently with the final stages of intervention development as the content, format and method of delivery became finalised. The implementation plan focused on the delivery of the intervention within the realities of the NHS. Therefore, the intervention and the associated implementation plan had to be adaptable to current practice, infrastructure and staffing at each of the three feasibility sites. This flexibility permitted delivery alongside current ‘standard’ care while stipulating the achievement of specified POs, against which the fidelity of the intervention was assessed.

To facilitate the implementation and adoption of the intervention, education and training materials were developed for each of the staff groups involved in its delivery. Appropriate support and training systems were developed and an implementation plan was constructed to assist adoption at each site, which included a site visit and ongoing support from the OPAL investigators.

Intervention mapping stage 6: evaluation plan and feasibility testing (see Chapter 10)

The final stage of the IM process evaluated the intervention by assessing four complementary aspects of its delivery and performance:

1. assessment of intervention fidelity – quantitative evidence that the intervention was delivered against specific POs for both the hospital orthopaedic team (HOT) (staff objectives) and the patient (patient objectives)

2. assessment of intervention quality – qualitative assessment of the intervention delivery obtained by interviewing patients and staff groups about what worked and what did not, why it did not work or why it went well

3. assessment of feasibility data – preliminary comparison of outcomes using data obtained from IM stages 1 (pre intervention) and 6 (post intervention)

4. assessment of economic data – approximate cost estimates for the intervention using derived health economic data.

In addition, the feasibility stage collected information that would help to shape the design and development of a future clinical trial by assessing screening, recruitment, consent and follow-up procedures and rates at each of the study sites. A formal pilot study was not undertaken at this stage as per the commissioning brief.

The OPAL IM approach described above is outlined in Figure 1 and a diagram describing development of the OPAL occupational advice intervention is shown in Figure 2.

FIGURE 2.

Diagram of the stages of development of the OPAL occupational advice intervention.

Overview

A rapid review methodology was used. Given that the commissioner had already identified an evidence gap relating to occupational advice interventions for patients undergoing hip or knee replacement, and the need for primary research and a future trial (if feasible), a full systematic review was not warranted. The purpose of the rapid review was to identify interventions that showed evidence of benefit (or a signal of benefit where the study is underpowered) to explore the content of the interventions and identify aspects that could inform the development of the intervention for people undergoing lower limb joint replacement.

The term ‘rapid review’ covers a range of methods and there is no generally accepted definition, although generally the approach addresses a trade-off between time and methodological rigour and comprehensiveness of the end product. 80 We focused on the systematic review evidence in the first instance, included only English-language articles published in the last 20 years, restricted the range of databases searched and double-checked a proportion of the literature searches, rather than 100% (which is accepted practice for a full systematic review). The protocol for the rapid review is available on PROSPERO (protocol registration number CRD42016045235). 81

Literature searches

There were two sets of searches: one for systematic reviews and one for primary studies reported outside the search dates or remit of the reviews identified.

The Cochrane Database of Systematic Reviews and Database of Reviews of Effectiveness were searched in August 2016 for systematic reviews up to 2015. Additional supplementary searches were undertaken for the period 2015 to July 2016 in MEDLINE and EMBASE. The search combined various terms for ‘occupational advice’ and ‘return to work’ with terms for ‘systematic reviews’. There was no restriction for type of population (e.g. elective surgery) so that the searches were as comprehensive as possible. The following five databases were searched for primary studies in August 2016: Cumulative Index to Nursing and Allied Health Literature (CINAHL), Cochrane Central Register of Controlled Trials, MEDLINE, EMBASE and OTseeker. The strategy combined terms for ‘surgery’ and terms for ‘return to work’ and ‘occupational advice’. The full search strategies are reported in Appendix 2.

An information specialist undertook the searches. Both sets of searches were restricted to English-language studies published in the previous 20 years (since 1996). Records were downloaded, added to EndNote (Clarivate Analytics, Philadelphia, PA, USA) bibliographic software and deduplicated.

In addition, reference and citation checking of included studies was undertaken to identify further potentially relevant records.

Eligibility criteria

The eligibility criteria that were applied are displayed in Table 2. We anticipated the literature outside elective surgery to be vast and dominated by RTW following mental ill health and musculoskeletal problems, such as back and neck pain, where generalisability to hip and knee surgery is less certain; hence, we initially excluded studies in which the participants were not undergoing an elective surgical procedure.

However, following an initial screening of the search results, where only a small number of studies were identified for elective surgical populations, we widened our inclusion criteria for the population. Hence, the review also included systematic reviews that evaluated occupational advice interventions, aiming to support RTW, targeted at participants with chronic musculoskeletal problems as this population was considered most similar to our target population of interest. Owing to resource constraints, it was not feasible to widen the inclusion criteria in a similar way for the supplementary primary study searches.

Study selection

The titles and abstracts of all studies identified by the literature search were screened for inclusion by one reviewer, with 30% screened by a second researcher. The full texts of potentially eligible studies were retrieved and assessed for eligibility by a single reviewer, with 100% also being assessed by a second reviewer, following the development and piloting of a screening tool. Any disagreement between the reviewers regarding this sample was resolved via discussion with a third reviewer.

Data extraction

A standardised data extraction form was developed and piloted to record key information, such as population, study design, intervention details, outcomes, surgical procedure type and results. Items related to the intervention followed the Criteria for Reporting the Development and Evaluation of Complex Interventions in healthcare,82 with outcome data extracted from the primary studies, and summary information provided for the systematic reviews. The data extraction forms can be found in Appendix 2. Data extraction was undertaken by a single reviewer, and checked by a second reviewer.

• For primary quantitative studies, the data extraction form recorded information including population; intervention (e.g. content of the intervention, material and tools used for delivery, who delivered, setting and any theoretical basis, such as behaviour change theory); process measures related to the delivery of interventions, such as barriers and facilitators; stakeholder perspectives (i.e. patients, health-care professionals and employers); study methods (e.g. study design, how outcomes were measured and length of follow-up); outcomes (e.g. what outcome measures are used in studies to assess RTW, return to normal activities and social participation?); and surgical procedure type.

• For primary qualitative studies, data were extracted for the following items: population; study objective; surgical procedure type; method of evaluation and underpinning methodology, views and experiences (related to return to work, normal activities and social participation); and process measures related to delivery of interventions.

• For reviews, the data extraction form also collected information such as objectives of the review; search strategies (e.g. searched databases, date of literature search, languages and inclusion/exclusion criteria); number of studies included in the review, sample sizes and details of data synthesis; types of studies included/setting, population, interventions assessed and outcomes assessed; quality assessment tools used; analysis (e.g. meta-analysis); results of the review; key conclusions; and limitations.

Assessment of risk of bias

Careful consideration was given to the risk-of-bias tools that were selected for use in our evidence synthesis, with a recent systematic review noting there being several limitations of existing tools regarding their scope, guidance for judgements on the risk of bias, and measurement properties. 83 Each of the tools listed below was considered to be appropriate for the different study type to adequately capture biases, with further information provided in the corresponding references for each tool. The quality of the included studies was assessed at the study level by one researcher and checked by a second. Specifically:

• for systematic reviews – the Assessing the Methodological Quality of Systematic Reviews (AMSTAR) tool,84,85 a measurement tool to assess the methodological quality of systematic reviews

• for RCTs – the Cochrane risk-of-bias tool86

• for non-randomised studies (including non-randomised controlled trials, controlled before-and-after and interrupted time series studies) – the ROBINS-I (Risk Of Bias In Non-randomised Studies – of Interventions) tool87

• for qualitative studies – the Critical Appraisal Skills Programme (CASP) qualitative checklist. 88

Data synthesis

Details of studies were tabulated and presented in a narrative synthesis to address the review questions. A meta-analysis was not possible owing to heterogeneity of studies and limited availability of RCTs. Key study characteristics have been tabulated, and the outcome domains investigated in the studies and specific outcome measures used have been mapped.

Many of the included systematic reviews had broad inclusion criteria and included primary studies outside the remit of interest (i.e. occupational advice interventions). Therefore, for the systematic reviews, the relevant primary studies were pulled out for closer examination, with the studies reported according to whether they featured a (1) surgical population or (2) wider musculoskeletal population. Mapping of the content of the interventions was also undertaken to allow exploration of all intervention components, materials and tools, any underlying theoretical basis, and any issues related to delivery and implementation. Data were explored and described by individual review question. No subgroup analysis was planned as part of this review.

Systematic reviews

Study selection

There were 859 records that were screened for relevance following deduplication of the results of the searches for systematic reviews (Figure 3). On reviewing titles and abstracts, 812 records were excluded, with 50 obtained in full-text form to assess eligibility for inclusion. A total of 17 systematic reviews were included, as listed in Appendix 2. The 33 excluded reviews and their associated reasons for exclusion are available in Appendix 2.

FIGURE 3.

Study selection for the review of systematic reviews.

What components of the interventions are likely to be generic across conditions and surgical procedures and, therefore, generalisable to an occupational advice intervention prior to planned surgery for hip or knee replacement?

The effective interventions tended to involve rehabilitation programmes, which took a multidisciplinary approach in general. In the majority of cases, it was not possible to disentangle the separate elements to determine whether certain components were playing more of a role in the effectiveness than others. The key components of the interventions that kept appearing irrespective of the condition and/or surgical procedure under consideration are summarised in Figure 4.

FIGURE 4.

Summary of key components across effective interventions.

Primary studies (surgical)

Study selection

The literature search of electronic databases identified 1179 potentially relevant records for the primary studies (Figure 5). After removal of duplicates, 989 primary studies were screened for relevance. A total of 856 primary studies were excluded on the basis of title and abstract and 140 full papers were retrieved for more detailed evaluation, which included seven obtained via reference and citation checking. A total of 136 papers were excluded and four studies met the inclusion criteria, with the included primary studies listed in Appendix 2. One of these studies had already been identified in the review of reviews and was also included here for the sake of completion so that it was quality assessed and discussed in conjunction with the only other identified study of a surgical population. 124 Details of excluded studies are also provided in Appendix 2.

FIGURE 5.

Study selection for review of primary studies.

Barriers to and facilitators of intervention delivery and stakeholder perspectives

Truncated data extraction tables from the two qualitative studies on stakeholder perspectives are provided in Appendix 2.

One UK study of 10 employed patients who had undergone TKR identified several facilitators and barriers from the patient perspective. 132 Three key themes were identified that have relevance for delivery of an occupational advice intervention:

1. Delays in surgical intervention and impact on work participation preoperatively. Patients felt that their employment status and need to remain in employment were not fully taken into consideration in the decision-making process about whether surgery should take place or be delayed until they were older. Perceived delays in surgery due to their age had a negative impact on their work before surgery and had the potential to have a negative impact on future employability.

2. Limited and inconsistent advice from health-care providers to optimise RTW. Patients reported that the advice they received focused mainly on the needs of an older retired population and covered the inpatient stay and immediate postoperative period but not RTW. Some patients thought that they should not return to work until they were advised to do so. Some reported that they could have returned to work earlier. Advice appeared to be generic rather than tailored.

3. Rehabilitation to optimise recovery and RTW. Patients reported that the postoperative rehabilitation they received was variable, that their need to return to work was not routinely considered and that they would have benefited from a more tailored approach. However, rehabilitation staff played an important role in giving them confidence to progress in their recovery.

One US study obtained the views of 31 mastectomy patients, 18 physical or occupational therapists and five employers. 133 Information provided about patients’ views on RTW was very limited. It is noteworthy that, although ‘many women’ described physical impairments that interfered with their ability to work, only one woman reported being asked by a health-care professional about the physical requirements of her job. However, 81% of therapists reported that job requirements were addressed in their treatment goals. Employers reported that they had written guidelines in place appropriate for people returning to work following surgery but that they would find it useful to have more tailored information about their employee’s physical restrictions, better patient education about expectations for recovery, more counselling services and better timing of clinic appointments to reduce disruption to work schedules. The authors commented that a common theme from all three stakeholder groups was the perceived dependence on doctors to guide the recovery process. It was suggested that some of this responsibility could be delegated to other health-care professionals.

Cohort study

Health economic analysis

We had originally intended to utilise information from the survey of practice to inform the mapping of the ‘standard care’ pathway. However, the findings from the survey highlighted considerable variation in what constitutes standard care at different institutions and according to different individuals (e.g. surgeons and physiotherapists), making it difficult to define and quantify standard care. Therefore, it was not possible to incorporate the survey data in the mapping of standard care for the economic analysis and this analysis was therefore based solely on the data collected from the cohort study.

Survey of practice

A web-based electronic software (SurveyMonkey^®, Palo Alto, CA, USA) was used to create the survey. To achieve national dissemination, a three-armed sampling strategy was used. First, the survey was e-mailed to the 149 individual trust leads for the NJR for England, Wales and Northern Ireland. Each lead was asked to disseminate the survey to relevant members of their clinical teams. Second, a link to the survey was embedded in an article about the OPAL study in the July edition of the NJR eBulletin (https://webactivate.hqip.org.uk/index.php?action=social%26c=284%26m=367; accessed 27 September 2017), which has an e-mail readership of > 3800 stakeholders. Third, to capture clinical teams in Scotland, the survey was distributed via the chairperson of the Scottish Committee for Orthopaedics and Trauma to members for dissemination within local organisations. The survey was available for completion for 6 weeks and collected information from all of the HOT involved in the treatment of hip and knee replacement patients.

The survey collected information specific to each member of the HOT. The survey explored (1) when each group interacted with patients as part of their preoperative pathway, (2) whether or not RTW advice was routinely given during this interaction, (3) the methods used to deliver RTW advice, (4) confidence in delivering advice and (5) the need for an occupational RTW advice intervention. The survey also offered the participants the opportunity to provide free-text comment. The survey was released on 1 July 2017 and responses were collated 6 weeks later, with the last response received on 11 August 2017. The OPAL survey of practice document is available at www.journalslibrary.nihr.ac.uk/programmes/hta/152802/#/ (accessed 2 May 2020).

Each question was summarised using simple descriptive statistics and, where appropriate, by job role. Direct comments from the ‘free-text’ question were grouped based on positive or negative experiences of delivering RTW advice.

Cohort

Screening data

Overall, 765 people were screened, of whom 202 (26.4%) were in work in the 6 months prior to surgery and were eligible for inclusion. All 202 patients in work met the other eligibility criteria. Figure 6 details the flow of participants through the study and details reasons for exclusion and non-participation. In total, 162 patients (80.2% of eligible patients) consented, of whom 154 (95.1%) provided baseline data (77 hip replacement patients and 77 knee replacement patients) and were followed up. Participants were recruited from all four sites over a 5-month period (1 November 2016 to 30 March 2017): Nottingham (n = 42), Norfolk and Norwich (n = 12), Middlesbrough (n = 62) and Northumbria (n = 38).

FIGURE 6.

Screening log data describing work status prior to surgery and intention to return to work after surgery.

The OPAL study intended to capture data from two defined groups of patients (see Figure 6). Group A comprised patients in work prior to surgery and planning to return to work after surgery, and group B comprised patients in work prior to surgery but planning to retire after their operation. However, only six screened patients (three consenting) were in group B, limiting the ability to analyse data from this group.

Follow-up data

Follow-up rates

All participants who had not withdrawn from the study were followed up at weeks 8 and 16; however, a subsample of participants was followed up 24 weeks post surgery. In total, 148 participants were provided with week eight questionnaires (73 hip and 75 knee participants), 139 were provided with week 16 questionnaires (70 hip and 69 knee participants) and 87 were provided with week 24 questionnaires (51 from Middlesbrough and 36 from Nottingham; 41 knee and 46 hip participants) (Figure 7 and Table 4). The Middlesbrough and Nottingham sites were the first to open for the OPAL study, so the participants reached week 24 first and became the subsample at this time point. This differs from the plan of including only 45 participants, as stated in the protocol. It can be seen that 83.8% of the participants replied to at least one of the follow-up questionnaires; the average response rate to the follow-up questionnaires was 61.6%.

FIGURE 7.

Flow of patients through the cohort study.

TABLE 4 - Cohort study returned questionnaires

Percentage given out of those who completed baseline (n = 154).

Time point	Group				Total
	Hip replacement		Knee replacement
	Number sent	Number returned (%)	Number sent	Number returned (%)	Number sent	Number returned (%)
Baseline	80	77 (96.3)	82	77 (93.9)	162	154 (95.1)
Week 8	73	50 (68.5)	75	43 (57.3)	148	93 (62.8)
Week 16	70	53 (75.7)	69	51 (73.9)	139	104 (74.8)
Week 24	41	23 (56.1)	46	18 (39.1)	87	41 (47.1)
Completed at least one follow-up questionnaire, n (%)a	65/77 (84.4)		64/77 (83.1)		129/154 (83.8)

Oxford Hip and Knee Scores

The OHS/OKS raw scores and a categorised representation are provided in Appendix 3.

At week 8, 93 participants completed the questionnaire (62.8% of those who were sent the questionnaire) and the average score was 33.6 for hip and 28.3 for knee participants. This increased to 38.6 and 54.1 at week 16, and remained similar at 24 weeks (see Table 4). The proportions of participants who were classified as ‘satisfactory’ increased from 1.3% at baseline to 19.4% at week 8 and to 45.2% at week 16 (see Appendix 3). Similarly, the proportion of those who classified their joint symptoms as ‘severe’ decreased from 49.4% at baseline to around 6% at week 16.

Return to work

Only 78 participants (50.6%, 37 hip and 41 knee participants) indicated that they returned to work within their period of follow-up. Of these, 74 participants (94.9% of returnees, 48.1% of entire cohort) provided a return date, allowing time between surgery and RTW to be calculated. On average, those who did return did so 10 weeks after surgery, with a range of 1 to 27 weeks. Return times are presented in Figure 8, and detailed in Table 5, for those who returned to work and those classified as early returners.

FIGURE 8.

Time to RTW after surgery. Note that two participants returned their 24-week questionnaires late.

TABLE 5 - Length of time (weeks) after surgery before participants returned to work

Q, quartile.

Early return was defined as returning in ≤ 6 weeks.

Length of time (weeks)	Group		Total (N = 154)
	Hip replacement (N = 77)	Knee replacement (N = 77)
Participants who returned to work	n = 36	n = 38	n = 74
Mean (SD)	9.7 (5.5)	10.3 (5.4)	10.0 (5.4)
Median (Q1, Q3)	8.5 (6.2, 13.1)	10 (6.3, 13.1)	9.4 (6.3, 13.1)
Minimum, maximum	1, 26.9	1.9, 27	1, 27
Participants who returned earlya to work	n = 8	n = 9	n = 17
Mean (SD)	3.6 (1.8)	3.9 (1.3)	3.8 (1.5)
Median (Q1, Q3)	4.5 (1.9, 5)	4.4 (3, 4.7)	4.4 (2.7, 5)
Minimum, maximum	1, 5.4	1.9, 5.7	1, 5.7

At 8 weeks’ follow-up, 27 of the 93 (29%) respondents had returned to work (12 hip and 15 knee replacement participants). Fifty-six (60.2%) respondents had not yet returned to work but intended to and nine (9.7%) stated that they no longer intended to return to work. At 16 weeks’ follow-up, 47 of the 103 (45.6%) participants stated that they had returned to work in the previous 8 weeks (23 hip and 24 knee participants), 17 (16.5%) had not yet returned to work but intended to and nine (8.7%) stated that they no longer intended to return to work.

Returning to work, workplace productivity and use of fit notes

When returning to work, 48.7% of the 78 participants who returned stated that they were working their usual hours and carrying out their usual duties in their first week, and a similar percentage (47.4%) returned with amended duties or hours. For those who had returned with reduced hours, the average amount of time worked in their first week back was 16.3 hours (range 3–40 hours). This is around half of the average amount of time worked before their operation (34 hours, range 6–65 hours). When asked about adaptations that had been made to their workplace and alterations to their pattern of work, only 16.7% (13/78) and 28.2% (22/78), respectively, said that any changes had been made.

Based on the Work Limitations Questionnaire, the average productivity loss in the 2 weeks prior to surgery was 30.4% for THR (SD 34.1%, range 0–100%) and 24.2% for TKR (SD 31.7%, range 0–100%) participants. For the patients who had returned to work after surgery, this had reduced to 19.7% and 5.1% for THR and 11.1% and 5.6% for TKR participants at 8 weeks and 16 weeks post surgery, respectively (Table 6).

TABLE 6 - Workplace Limitations Questionnaire data for the percentage of time lost at work for the cohort participants at each time point

Q, quartile.

Only 87 participants were invited to fill in a week 24 questionnaire.

Percentage of time lost	Group		Total (N = 154)
	Hip replacement (N = 77)	Knee replacement (N = 77)
Baseline	n = 65	n = 62	n = 127
Mean (SD)	30.4 (34.1)	24.2 (31.7)	27.4 (32.9)
Median (Q1, Q3)	25 (0, 50)	0 (0, 50)	0 (0, 50)
Minimum, maximum	0, 100	0, 100	0, 100
Week 8	n = 19	n = 18	n = 37
Mean (SD)	19.7 (30.7)	11.1 (26.0)	15.5 (28.5)
Median (Q1, Q3)	0 (0, 25)	(0, 0)	0 (0, 25)
Minimum, maximum	0, 100	0, 100	0, 100
Week 16	n = 39	n = 36	n = 75
Mean (SD)	5.1 (13.1)	5.6 (12.1)	5.3 (12.1)
Median (Q1, Q3)	0 (0, 0)	0 (0, 0)	0 (0, 0)
Minimum, maximum	0, 50	0, 50	0, 50
Week 24a	n = 16	n = 15	n = 31
Mean (SD)	10.9 (22.3)	11.7 (26.5)	11.3 (24.0)
Median (Q1, Q3)	0 (0, 12.5)	0 (0, 25)	0 (0, 25)
Minimum, maximum	0, 75	0, 100	0, 100

Around half (50.5%) of those responding at week 8 stated that they had been given a fit note after their operation. The majority of these fit notes stated that the participant was not fit for work (87.2%) or may be fit to work taking into account a phased return (8.5%). Very few of these fit notes (n = 1 at 16 weeks) stated that amended duties may be needed. The mean length of time that the first fit note supplied to patients post surgery applied to was 5.6 weeks (range 2–10 weeks), and this was similar for hip (5.7 weeks) and knee (5.4 weeks) patients.

Further details about the mode of RTW and fit notes provided are in Appendix 3.

Returning to driving and normal activities

At week 16, 58 of the 79 (73.4%) participants had returned to driving when expected; at baseline, this was estimated to be 6 weeks. Similarly, 48 out of 85 (56.6%) participants said that they had returned to normal activities when they expected to (around 9.3 weeks as stated at baseline).

Access to occupational advice

Overall, 44 (28.6%) participants reported having access to an occupational health service through their employer at baseline. However, when asked at week 8, only 36 (23.4%) participants stated that they had received advice about returning to work post surgery. For those participants who received advice, the advice came from a variety of sources including surgeons, GPs, occupational health teams, physiotherapists, occupational therapists and employers (see Appendix 3).

Predictors of return to work

Because only 78 participants returned to work within our follow-up time frame, 74 of whom provided a return date, the number of variables to be included in the model was limited. Factors including age, sex, BMI, ethnicity, type of employer, number of hours worked and standardised outcome measures showed little or no evidence of predicting RTW. Size of employer, especially working for a micro employer, showed a sign of prediction when used solely in a model; however, when other factors were also included, this became non-statistically significant (see Appendix 3).

Although other studies found factors that were predictive of RTW within this population,42,44,48,146 our lack of predictive factors may be due to the relatively small sample size. Given the small numbers, no further statistical analyses were undertaken.

Health economics

Survey of practice

Responses were received from a total of 152 participants from 59 different public and private health providers across England (n = 47), Wales (n = 1), Scotland (n = 10) and Northern Ireland (n = 1). These included 78 surgeons, 20 physiotherapists, 25 occupational therapists and 25 nurses/specialist nurses/extended-scope practitioners. A further four participants who labelled their role as ‘other’ were excluded from the analysis as their role within the HOT and input into the orthopaedic surgical pathway was unclear.

Sampling

From the cohort, a purposive sample of 45 patients who intended to return to work following surgery were interviewed at approximately 16 weeks post surgery. Patients were sampled to provide an equal proportion of participants having had hip or knee surgery, representing a range of work roles and employing organisations. Interviews were conducted by telephone. We had originally planned to interview a subgroup of patients not intending to return to work; however, these did not take place as only three participants met the criteria (the cohort study screening information is available at www.journalslibrary.nihr.ac.uk/programmes/hta/152802/#/; accessed 2 May 2020).

A framework approach was used to design the semistructured interviews and analyse data. 147,148 This method is widely used in health research and is particularly recommended for use in multidisciplinary health research teams. As a range of stakeholder groups and patients were to be interviewed, this was an appropriate design. The theoretical framework reflected an essentialist/realist perspective, reporting on the experiences, meanings and reality of the participants, rather than examining the ways in which the broader social context impinges on those meanings. Interview schedules informed by initial piloting with service users were used (see Appendix 4).

Data analysis

The analysis procedure followed the seven stages proposed by Gale et al. :148 transcription, familiarisation with the interview, coding, developing a working analytical framework, applying the analytical framework, charting the data in the framework matrix and interpreting the data.

The patient interviews were conducted by Carol Coole and Fiona Nouri (both occupational therapists by background). Both interviewers were experienced in conducting qualitative research, and in relation to the topic of work and health. Interviews were conducted individually and by telephone. This was a pragmatic decision made in order to recruit the intended sample and conduct the interviews using the resources available.

All interviews were digitally recorded and transcribed verbatim, and NVivo 10 (QSR International, Warrington, UK) was used to manage the data, which were analysed thematically. Following familiarisation, the first few transcripts were independently coded by the researchers who conducted the interviews, who then compared, revised and agreed a set of codes and/or categories to form a working analytical framework. This framework was used to code the remaining transcripts. Summarised data were charted into a matrix to facilitate comparison of data across cases as well as codes and categories. Potential themes were identified independently by the interviewers, who discussed, revised and agreed the final themes.

Themes identified from the interview analysis

The following themes were identified relating to RTW after hip or knee replacement:

• preoperative context

• postoperative context

• advice received

• GP role and fit note

• barriers to and facilitators of RTW

• perceptions of an occupational advice intervention.

Patient interview analysis

Summaries of the analysis for each identified theme (see previous section) are described in the following sections. Direct quotations supporting these themes are provided in Appendix 4.

Sampling

A sample of eight employers (the term ‘employers’ is used in the broadest sense, encompassing a range of individuals within the workplace, including managers, human resources staff, occupational health staff and colleagues) around each study site were recruited from organisations of differing sizes and sectors via local employer organisations and contacts. Eligible participants had experience of employees undergoing THR or TKR in the previous 12 months. Data were also collected from HOTs and local GPs. A sample of 12 orthopaedic surgeons, 12 AHPs and nurses and 12 GPs were interviewed across the sites to provide sufficient diversity of views and experiences. Interviews with these stakeholders were conducted by telephone, face to face or in small focus groups. Interview schedules informed by initial piloting with stakeholders were used (see Appendix 5).

A methodology and framework approach similar to that described in Chapter 5 was used.

Data analysis

The stakeholder interviews were conducted by Carol Coole, Fiona Nouri (occupational therapists by background) and Melanie Narayanasamy (social scientist by background). All interviewers were experienced in conducting qualitative research, and in relation to the topic of work and health. Interviewees were offered face-to-face or telephone interviews, either as a group or individually, according to preference. This was a pragmatic decision made in order to recruit the intended sample and conduct the interviews using the resources available.

Interview conduct

Twenty-five workplace representatives were interviewed, 15 by telephone and 10 face to face, between September 2016 and June 2017 (mean duration 36 minutes). Recruitment was extended outside the geographical catchment of the study sites. The characteristics of the participants are listed in Appendix 5.

Twelve interviews were conducted with AHPs and nurses: six by telephone and six face to face. The mean interview duration was 52 minutes. Characteristics of the AHP/nurse participants are listed in Appendix 5.

Twelve interviews were conducted with consultant orthopaedic surgeons (mean duration 51 minutes). One was interviewed by telephone and 11 were interviewed face to face, either in a group or individually. Characteristics of the surgeon participants are listed in Appendix 5.

Sixteen interviews, 10 by telephone and 6 face to face, were conducted with GPs (mean duration 36 minutes). Characteristics of the GP participants are listed in Appendix 5.

Themes identified from the interview analysis

The following themes were identified relating to RTW after hip or knee replacement.

Workplace representatives (n = 25):

• experiences of accommodating patients undergoing THR and TKR in the workplace

• barriers to and facilitators of RTW

• perceptions regarding an occupational advice intervention.

Clinicians [AHPs/nurses (n = 12), surgeons (n = 12) and GPs (n = 16)]:

• decision to have surgery and expectations of recovery

• advising patients about work and other activities

• barriers to and facilitators of RTW

• perceptions regarding an occupational advice intervention.

Workplace representative interview analysis

Summaries of the qualitative analysis are provided below. Direct quotations supporting the themes are provided in Appendix 5.

Clinician interview analysis

Population: the return-to-work population

Invervention: an occupational advice intervention

Comparator: advice currently provided to return-to-work patients

Outcome: measurement of return to work

Stakeholder recruitment

Five stakeholder groups were identified for inclusion in the modified Delphi process, mirroring the groups involved in IM stage 1. The sampling strategy for each stakeholder group is outlined in Table 1, with participants chosen to maximise patient, public and professional engagement. Participants were nationally sampled from across England and Scotland (see Appendix 7, Table 67). To ensure the validity of the consensus process, a minimum of five individuals from each stakeholder group were recruited. A maximum limit of 15 individuals from any given stakeholder group was chosen to ensure that one group’s opinions did not overwhelm the opinions of others in the consensus process. In total, 66 participants were invited to participate in round 1 of the Delphi process (see Appendix 7).

Statement development

Statements relating to the proposed content, format, delivery, timing and measurement of the occupational advice intervention were developed within the OPAL investigator group. Owing to the breadth of statements and their inter-related nature, a step-wise approach to the presentation of individual statements to the Delphi group was adopted.

Round 1 focused on defining the content of the intervention in two sections. Section 1 focused on the content (‘written’ advice and information) and section 2 focused on activities to deliver content (actions or processes for patients, employers and health-care members to undertake). These statements were piloted by two surgeons, two GPs and three patients prior to distribution to the Delphi panel.

The information from round 1 was then used to refine the statements for round 2, which focused on defining the format, delivery, timing and measurement of the content examined in round 1. The statements for round 2 were grouped under headings in order to enable the Delphi panel to explore different approaches to these specific areas. Round 3 was then used to clarify any areas of residual uncertainty from rounds 1 and 2 and to present the draft occupational advice intervention back to the Delphi participants for comment.

Definition of agreement

Participants were asked to rate individual statements in the Delphi questionnaire, with possible options being strongly agree, agree, disagree, strongly disagree and don’t know. Participants were asked to rate the importance of the content or action given in the statement. For a subset of statements in round 1, participants were also asked to rate the deliverability of the content or action alongside current health-care provision. This was done to assess whether or not the stakeholders felt that certain actions, despite being important, were achievable owing to their experience of current service delivery, funding and logistics. Therefore, for this subset of statements, the participants were asked to provide two ratings: one for ‘importance’ and one for ‘deliverability’.

At the end of each section, participants were also able to insert comments in a free-text box or provide additional suggestions relating to the intervention that could be evaluated in subsequent rounds. In rounds in which statements from a previous Delphi round were being re-presented, these were presented alongside controlled feedback, with modal round 1 rating for these statements; the proportion of each response option selected by the other participants; and a reminder of the participant’s own previous ratings.

Delivery of the Delphi survey

The Delphi survey was delivered via e-mail using an online web-based survey platform (SurveyMonkey). Round 1 was delivered between 25 September 2017 and 13 October 2017, round 2 between 22 November 2017 and 13 December 2017 and round 3 between 1 June 2018 and 22 June 2018. The OPAL Delphi questionnaires are available at www.journalslibrary.nihr.ac.uk/programmes/hta/152802/#/ (accessed 2 May 2020). The survey included a covering e-mail to the participants and an electronic link to the questionnaires. This e-mail informed the participants of the details of each round and provided instructions on completing the survey. Round 1 and 2 questionnaires required respondents to provide their initials and occupation to assist the investigators in identifying respondents. Round 3 e-mails included four documents from the developed occupational advice intervention (a summary of the intervention, an occupational checklist, a patient ‘return-to-work’ workbook and an employer booklet) for participants to review and comment on. A minimum of two reminders were sent to non-responders during the final week of the surveys.

Analysis of data

An a priori consensus level of 70% (strongly agree/agree or strongly disagree/disagree) across all stakeholder groups combined was set. 151 For statements that failed to reach consensus across the overall group, further analysis was undertaken based on responses for each of the five stakeholder subgroups. The following rules were then employed to determine which statements were discarded and which were re-presented in the next round:

• if no or only one stakeholder group reached concordant consensus (≥ 70% agreement or disagreement), then the statement was withdrawn

• if two or more stakeholder groups reached concordant consensus (≥ 70% agreement or disagreement), then the statement was re-presented in the next round

• in the situation in which one or more stakeholder group reached ‘agreement’ and another group reached ‘disagreement’, the statement was discussed among the investigators and a decision on inclusion/exclusion of the statement was made.

As stated previously, a subgroup of round 1 statements were rated for their deliverability in addition to their perceived importance. For these statements, consensus was reached if the 70% threshold was achieved for both the importance rating and the deliverability rating. Statements that reached consensus for one of the domains were analysed by stakeholder grouping as described above.

Round 1 statements relating to the content of the intervention were ranked according to the level of agreement to enable the investigators to determine which items of content were the most important to include in the intervention.

In round 3, the intervention documents created based on the results of the first two rounds were presented and descriptive open feedback from participants was recorded.

Round 1

Responses were received from 43 of the 66 invited participants (65%), comprising 14 patients, 8 surgeons, 6 GPs, 11 AHPs and nurses, and 4 employers. In section 1 (‘written’ advice and information), consensus was reached for 26 of 32 statements (81%). Of the remaining six statements, five reached consensus for two or more stakeholder groups and were therefore taken forward to round 2, and one statement was discarded. A full summary of all round 1, section 1, responses and analysis is reported in Appendix 7. The top 10 ‘section 1’ statements reaching consensus, ranked based on the level of agreement (percentage who responded strongly agree or agree), are listed below.

Question: is it important that an occupational advice intervention commenced prior to hip or knee replacement includes the following (percentage strongly agree or agree)?

1. information about exercises and rehabilitation following surgery (100%)

2. information about returning to driving (100%)

3. a broad overview written for all stakeholders, of what to expect following surgery (rates and timing of expected recovery) (98%)

4. information about managing pain, types of analgesia and side effects (98%)

5. information about postoperative precautions, restrictions and activities to avoid following surgery (95%)

6. information about symptom management in relation to RTW and specific occupations (e.g. expected levels of fatigue, pain and swelling) (95%)

7. tips and tricks to help the patient manage around their home with day-to-day activities immediately following surgery (95%)

8. information regarding postoperative complications and their management (95%)

9. signposting to DVLA guidance (95%)

10. information for the patient about who to ask if they are having a problem with returning to work (93%).

In section 2 (actions or processes for patients, employers and health-care staff to undertake), participants were asked to rate both the importance and the deliverability of each statement. Of the 32 components presented, 10 (31%) reached consensus for both importance and deliverability. Of the remaining 22 statements, 14 reached consensus for importance but not deliverability, two reached consensus for deliverability but not importance and six did not reach consensus for either. Of these statements, seven reached consensus for both importance and deliverability for two or more stakeholder groups and were therefore taken forward to round 2, and 15 statements were discarded. A full summary of all round 1, section 2, responses and analysis is reported in Appendix 7. The top 10 ‘section 2’ statements reaching consensus, ranked based on the level of agreement (percentage who responded very important or important), are listed below.

Question: how important/deliverable do you believe the following components are if an occupational advice intervention commencing prior to hip or knee replacement was to be developed (percentage who strongly agree or agree with the importance and deliverability of the statement)?

1. a postoperative mechanism for the identification of patients who are not progressing towards RTW as planned (important 95%, deliverable 71%)

2. guidance for health services defining ‘best practice’ for patients returning to work after hip or knee replacement surgery (important 93%, deliverable 82%)

3. training for members of the hospital orthopaedic care team to increase awareness about RTW issues (important 88%, deliverable 82%)

4. interaction between the health-care team and the patient by telephone, e-mail or ‘online’ so that members of the care team can monitor progress and help the patient use the advice and information provided (important 88%, deliverable 70%)

5. guidance on when in the RTW process patients can safely be discharged back to primary care for continued management of their RTW (important 86%, deliverable 80%)

6. a mechanism for preoperative identification of patients at ‘high risk’ of prolonged sickness absence following surgery (important 86%, deliverable 74%)

7. routine preoperative therapy assessment during which a RTW plan is developed between the patients and the hospital orthopaedic care team (important 84%, deliverable 80%)

8. a separate intervention for hip and knee replacement patients who are not progressing towards RTW as planned (important 84%, deliverable 79%)

9. a process by which work status can be included in referral information for all patients referred from primary care into secondary care for consideration of hip or knee replacement (important 79%, deliverable 79%)

10. information from patients who have experienced the process of returning to work after hip or knee replacement within the preoperative education process (important 76%, deliverable 73%).

Round 2

Responses were received from 26 of the 66 participants (39%), comprising eight patients, seven surgeons, three GPs, six AHPs and nurses, and two employers.

The 12 questions (five from section 1 and seven from section 2) carried forward from round 1 plus one additional question generated from the free-text comments (see Appendix 7) were presented to participants. In round 2, participants were asked to rate only the importance of these statements, having been made aware of the responses from round 1 and the overall group’s position regarding the deliverability of the component if applicable. Of the 13 statements, 10 reached the threshold for consensus (see Appendix 7).

Based on the responses to round 1, a further 81 statements grouped into 13 categories were generated. These statements related to the format, delivery, timing and measurement of the agreed content components from round 1. Each category included three to nine statements relating to a common category theme (see OPAL Survey questionnaires at www.journalslibrary.nihr.ac.uk/programmes/hta/152802/#/; accessed 2 May 2020). This allowed the participants to compare different options presented within the category in the context of the other available options and reach a position on each statement accordingly. This allowed the investigators to explore different approaches to a given problem. For example, the first category asked participants to rate a set of five statements relating to which health-care team member should have responsibility for delivery and co-ordination of the occupational advice intervention. If at least one statement in a given category reached consensus, this was taken as representative of the Delphi group’s position relating to the given category and the remaining statements were discarded.

Overall, 49 statements (60%) reached consensus (44 agreement and 5 disagreement). At least one statement in every category reached consensus (see Appendix 7).

The 13 categories presented to participants and a summary of the responses are listed below:

1. Responsibility for delivery and co-ordination of the RTW intervention. Participants agreed that the orthopaedic surgeons should not be responsible for delivering and co-ordinating the RTW intervention (88% agreement). Surgeons and patients agreed (both > 70% agreement) that the intervention should be co-ordinated by primary care teams. However, GPs felt that the intervention should be co-ordinated in secondary care by a nominated member of the orthopaedic team (100% agreement).

2. Preoperative identification of patients at ‘higher risk’ of prolonged sickness absence following surgery who may require additional individualised help and support. There was agreement (80%) for a face-to-face assessment with a staff member trained in the RTW intervention to help identify patients at ‘higher risk’ of prolonged sickness absence following surgery. No agreement was reached on other forms of assessment, including written, telephone, patient’s self-referral and surgeon assessment.

3. Preoperative assessment. Participants felt that a greater level of involvement from the therapy team would be beneficial irrespective of whether or not the patient was deemed to be at ‘high risk’ of an extended period of sickness absence after surgery (80–88% agreement). They also felt that development of a RTW plan prior to surgery (80% agreement), discussion with employer (100% agreement) and routine inclusion of RTW information in pre-assessment/education classes (88% agreement) were important.

4. Postoperative identification of patients at risk of an extended period off work after surgery. There was overall agreement for additional monitoring for RTW patients after surgery, including routine physiotherapy (84% agreement), closer contact with the therapy team (92% agreement) and access to telephone support (92% agreement). Participants also felt that patients should meet with their employer after surgery to discuss their recovery and plan for RTW (72% agreement).

5. Additional care for patients identified as being at ‘higher risk’ of an extended period off work after surgery. Participants agreed that ‘high-risk’ patients should receive additional physiotherapy (92% agreement) and occupational therapy (76% agreement) input and that this should continue until RTW had been achieved (76% agreement). This could be supplemented by additional information (92% agreement) and referral to the Fit4Work (Leeds, UK) service (84% agreement).

6. Scope of training for staff. The group felt that all members of the HOT involved in the treatment of hip and knee replacement patients should receive training regarding the intervention (76% agreement) and that training should also be offered to local GP groups (72% agreement).

7. Communicating occupational status and progress between stakeholders. There was widespread agreement for statements relating to improved communication between stakeholders through greater information in referral, clinic, discharge and therapy service letters (84–96% agreement). Participants agreed that greater information sharing between primary and secondary care and employers would be beneficial (84–96% agreement). However, there was also a feeling that it was the patient’s and not the health-care team’s responsibility to communicate with their employer about their RTW (76% agreement).

8. Fit notes. Every patient should be offered a fit note (96% agreement) and it should be completed in accordance with Department for Work and Pensions Fit Note Guidance152 (96% agreement). Participants felt that providing short-duration fit notes (2 weeks) to discourage extended periods off work was wrong (72% agreement) but that GPs should be responsible for administering subsequent fit notes after the initial fit note given on discharge (76% agreement).

9. Format and delivery of patient information. Written materials were favoured (76% agreement), although participants felt that only the most ‘important’ information (based on round 1 ranking) needed to be included, with additional information made available through other sources (e.g. website) (80% agreement). It was felt that an interactive booklet in which information and progress could be recorded and shared (80% agreement) and interactions with their employer documented (92% agreement) would be beneficial. Online materials should be accessible by GPs prior to referral (88% agreement).

10. When should the intervention commence? There was uncertainty about the timing of the intervention. AHPs and GPs felt that it should start during the pre-assessment process. Overall, there was agreement that it should not start after surgery (88% agreement).

11. Defining RTW. Participants struggled to agree on a definition of RTW. However, they agreed (72%) that RTW should not be defined as the patient returning to the same job (usual hours and duties) and activities outside of work they were doing prior to surgery.

12. The aim of the intervention. There was agreement that the RTW intervention’s primary aim should be to return the patient to their preoperative work role and level of occupational performance (76% agreement).

13. Measuring RTW. There were a number of ideas about how RTW could be measured. The group felt that RTW was not a binary outcome (72% agreement). There was agreement on more detailed assessment, including the use of graded assessment based on specific work milestones, for example return to place of work, return to normal hours and return to normal workplace activities (92% agreement) or the resumption of specific work-related activities (e.g. climbing, lifting and manual tasks) based on a list of preoperative workplace activities (72% agreement). Rate of further sick leave (80% agreement), requirement and duration of occupational adjustments (84% agreement) and patients’ reports of their experiences and expectations were also felt to be important.

Following round 2, the investigator group convened to further refine the intervention and develop supporting materials and resources (discussed further in Chapter 9). During this process, a number of materials were drafted, which were subsequently shared with the Delphi participants in round 3.

Round 3

In round 3, the finalised occupational advice intervention along with selected patient and staff materials were circulated to 65 of the 66 Delphi participants for comment (one participant had withdrawn). Responses were received from 11 participants (four surgeons, one physiotherapist, one GP, three employers and two patients), comprising a constructive appraisal of the intervention from nine respondents (two employers responded but did not comment on the intervention) as well as highlighting typographical and formatting issues. The feedback was positive in all cases and all of the comments received are reported in Appendix 7.

Content, format and delivery

The majority of the written information presented to the Delphi group reached consensus either in round 1 (26 of 32 statements) or when re-presented in round 2 (three of five statements). During the interviews (see Chapters 5 and 6), both patients and employers had stated a preference for written materials, although there was a concern about patients becoming disengaged if the amount of information was overwhelmingly large. The Delphi process suggested that only the most ‘important’ information (based on Delphi ranking) needed to be included in any written materials and that information could be made available in other formats (e.g. a website).

There was support both from the interviews and through the Delphi process for an interactive booklet in which information and progress could be recorded and shared with other stakeholders invested in the RTW process. The OPAL investigators thought that this approach had the potential to allow patients to record individualised information about their workplace, the impact of their health condition, plans for returning to work and progress after surgery, providing an individualised intervention. This individualised approach also had the potential to avoid having a separate intervention for hip and knee replacement patients as long as patients who did not make the anticipated progress received individualised ‘joint-specific’ support after surgery. On this basis, the OPAL investigators created an interactive patient workbook that enabled the patient to develop a RTW plan, tailored to their own circumstances, that could be shared and agreed with their employer. The decision was made to provide key information in the booklet, with additional information available via a website that was signposted within the booklet.

There was greater uncertainty about the ‘active’ elements of the intervention. Two key areas that stakeholders felt were important were (1) the provision of additional preoperative and postoperative therapy (over and above standard care) in which RTW issues could be addressed and (2) the identification of ‘high-risk’ patients, with additional support made available for this group. The initial draft intervention (see Chapter 7) mirrored this position and aimed to provide three preoperative and three postoperative interactions between the patient and hospital team to discuss RTW, with additional input for ‘high-risk’ patients. However, the investigators identified difficulties with this approach.

First, our cohort data (see Chapter 4) failed to identify a ‘high-risk’ population and the current literature describing predictors of RTW after hip or knee replacement (see Chapter 1) was limited. 42,44,48,146 This meant that we were not able to confidently identify a ‘high-risk’ group in need of a more intensive intervention. The OPAL investigators therefore decided not to pursue a tiered high-/low-risk intervention and instead focused on developing an intervention that could be tailored to the patients’ needs with the ability to offer more or less support as required.

Second, there was concern about the cost, time and logistics associated with the implementation of a resource-intensive intervention requiring an additional six patient interactions. The survey of practice (see Chapter 4) and interviews (see Chapters 5 and 6) demonstrated that services varied significantly in their structure and the resources available, and concerns were expressed about the implementation and sustainability of an intervention requiring significant additional resources. Therefore, a flexible model that allows the intervention to be delivered at different times by different people in different trusts dependent on the timing and delivery of current standard care in individual trusts was considered desirable. Despite the need to limit resource use, the investigators felt that it was important to have a hospital team member actively engaging with RTW patients. This reflected a key aspect of RTW interventions reported in the rapid evidence review (see Chapter 3) in which there was an element of counselling and guidance and the ability to co-ordinate the wider multidisciplinary team. We therefore developed a return-to-work co-ordinator (RTWC) role that had a range of responsibilities including co-ordination of the RTW process, encouragement and supporting completion of the interactive patient workbook, being a point of access for problems and signposting and assisting with referrals to other existing services should this be required (see Staff roles).

The process to identify and support patients having problems and not making progress as expected was then considered. The Delphi group agreed that a postoperative mechanism for the identification of patients who were not progressing towards RTW as planned was required. They felt that the intervention should include the ability for the health-care team to interact with the patient by telephone, by e-mail or ‘online’ to help monitor progress. These actions were linked to the RTWC role through the requirement for the co-ordinator to contact patients prior to surgery to support and monitor their RTW process and be available after surgery via a dedicated RTW contact line (telephone or e-mail). This provided patients with access to a designated point of contact who could offer additional help and support if needed. This, in effect, was an extended version of ‘usual care’, in which problems after surgery are assessed by the patient’s clinical team and treated accordingly (e.g. group physiotherapy, one-to-one physiotherapy, referral to occupational therapy and referral to occupational health services). However, for the purpose of the intervention, this task was centralised and administered by the RTWC.

Issues relating to fit notes and communication between stakeholders were more easily accommodated within the intervention through the development of specific guidance and examples of correct completion (see Materials and resources). Although there was agreement that recording work status in referral information from primary care would be beneficial, the OPAL investigators felt that this was not logistically possible within the study time frame owing to the large number of general practices referring into secondary care teams.

Finally, throughout the needs assessment and Delphi process, there was a lack of clarity about who should be responsible for administering the intervention. The prevailing opinion was ‘anyone but me’. Therefore, as work-focused health-care is a relatively new concept and delivery of occupational advice is not perceived to be the role of any particular health professional, the OPAL investigators agreed that all members of staff should be encouraged to take an active part in the intervention and be aware of each other’s role in delivery. This would help to embed the concept within the pathway and supported ‘organisational’ change. Training for all members of the HOT was therefore provided to increase awareness about RTW issues.

Timing: when should the intervention be initiated?

Given the residual uncertainty following the Delphi process about when the intervention should start, the OPAL investigators reflected on the information from the IM stage 1 needs assessment. It was subsequently decided that the optimal time to initiate an occupational advice intervention was the outpatient consultation during which patients were listed for surgery. This was based on the following factors:

• Patients and HOT members (AHPs and nurses) take the lead from their surgeon. Surgeons are integral to the delivery of information to their patients, who, in many cases, will not contemplate or consider returning to work without their permission. AHPs and GPs involved in the care of these patients often defer decisions relating to RTW to the surgeon (from interviews).

• In over two-thirds of cases, the only time the patient sees their surgeon prior to surgery is in the initial outpatient consultation. Surgeons then do not see patients again until the morning of their operation, limiting the opportunity for interaction between the surgeon and their patients (from survey).

• Over 90% of surgeons do not offer routine advice to patients returning to work after surgery and, when it is delivered, it is ad hoc verbal advice based on anecdote and personal experience. Patients, GPs and the HOT look to surgeons to lead the RTW process; however, they do not routinely provide advice and when they do it is not based on specific guidance or best practice (from survey/interviews).

• The preoperative assessment process is extremely varied between trusts. The composition, timing (sometimes only 1 week before surgery) and staffing of these services would make it difficult to embed an intervention that fulfils the individualised needs of this patient group (from survey).

• The outpatient consultation is a consistent step within the preoperative pathway in all institutions. Early discussion of RTW issues allows adequate time for patients and employers to develop, communicate and instigate a suitable plan to enable early and sustained RTW (from survey/interviews).

A pragmatic decision was made to allow access to the RTWC up to and including 16 weeks after surgery. Cohort data (see Chapter 4) suggested that the majority of patients had returned to work by 16 weeks and this, therefore, seemed an appropriate time point for the end of the intervention. This should allow some additional time to access the support following the standard postoperative review by the surgical team at 8–12 weeks after surgery when the patient was routinely either discharged back to primary care or offered further appointments (usually at 12 months post operation).

Timing

• The intervention supported patients throughout their surgical pathway. It started in their outpatient appointment during which they are listed for surgery and continued until 16 weeks after their surgery.

Patient identification

• All patients in work and intending to return to work after surgery were identified as RTW patients at their initial outpatient clinic appointment. The definition of ‘work’ included being in full-time or part-time employment or self-employment. It also included patients who are full- or part-time carers or who work as volunteers. The identification process was facilitated by the use of an occupational checklist completed by patients prior to their clinic appointment. The checklist was administered by a member of the outpatient clinic team when the patient arrived for their appointment.

• The surgical team used the information on the occupational checklist to aid surgical decision-making with respect to surgery and allow an individualised preliminary discussion of RTW with the patient.

• Patients who were subsequently listed for hip or knee replacement surgery (and consented to participate in the OPAL study) were signposted to the OPAL intervention resources (OPAL patient RTW workbook, employer information resource, website and local RTWC) by their surgical team.

Delivery of information

• All patients in work and intending to return to work after surgery were provided with the following resources at the point at which they were recruited into the OPAL programme (in clinic after they are listed for surgery):

  • The patient RTW workbook. This was designed as an eight-step interactive workbook. Completion of the workbook helped patients to list and understand their current job demands, set a provisional RTW date, identify potential barriers and solutions to safe and appropriate RTW and develop a provisional RTW plan that could be shared with their employer/work colleagues. The completion of the workbook was the responsibility of the patient but was overseen by a designated RTWC who was a member of the orthopaedic team.

  • The employer RTW information resource. This mirrored the information in the patient workbook. It explained the OPAL project and the steps the patient will follow when completing their patient RTW workbook and provided useful information for employers and work colleagues with respect to returning to work after hip or knee replacement. The patient was provided with a copy to give to their employer, manager, occupational health link or other relevant work colleagues.

  • Signposting to the OPAL website: www.opalreturntowork.org.uk (accessed 1 April 2019). This contained additional information and advice for patients, employers, HOTs and GPs to access.

Assessment by a designated member of the orthopaedic team

• All patients were contacted by a RTWC prior to surgery. The RTWC was a designated member of the orthopaedic team identified at each site who was involved in the assessment, management or education of hip and knee replacement patients. The RTWC offered support to patients, encouraged them to complete the patient RTW workbook and discussed the plans they have developed. This contact took place a minimum of 4 weeks prior to surgery.

• The RTWC also encouraged patients to share their plans with their employer if they had not done so already.

Support, review and escalation

• During the course of their assessment, the RTWC offered additional support to patients based on need. This decision was made on an individual patient basis having discussed and reviewed the information in the patient RTW workbook. Additional support could involve review and input from local therapy teams (in hospital or the community) and could be arranged either preoperatively or post discharge.

• The RTWC facilitated a mechanism that allowed patients to contact them following their surgery (e.g. answerphone or e-mail). If indicated, this could prompt further review and referral back into local therapy services.

Communication

• The intervention included mechanisms and guidance to support communication within the hospital team, between the hospital team and primary care and between the patient and their employer:

  • signposting ‘in-hospital’ teams (e.g. pre-assessment, ward nurses and doctors, inpatient and outpatient therapy services) to patients in the OPAL programme by the RTWC

  • guidelines for clinic letters, fit notes and discharge communication to support communication between secondary and primary care

  • the employer RTW information resource and specific instruction and advice in the patient RTW workbook to assist communication between the patient and their employer

  • communication between the patient and the RTWC via the telephone/e-mail service

  • a comprehensive training platform for staff to embed the OPAL programme within the participating unit’s/surgical team’s practices.

Training

• The OPAL intervention provided training for members of the hospital orthopaedic care team who interact with RTW patients to increase awareness of RTW issues across the orthopaedic department.

Overview

An implementation strategy was developed to support the adoption and implementation of the OPAL intervention at three NHS hospital trusts in England. At each hospital, a principal investigator (PI) (a consultant orthopaedic surgeon) was identified to assist the OPAL investigators as they had knowledge of local service structure and personnel. The PIs identified staff members for the OPAL intervention roles (e.g. RTWC) based on local service structure and personnel. The person used in the RTWC role was chosen by the PI after discussion with the OPAL research team based on seniority, experience and time available for the role. The PI then planned how the intervention would work alongside ‘standard’ care at their institution.

Support and clarification were provided by the OPAL investigators when needed; however, the local delivery of the intervention was largely determined by the local teams through negotiation with the PI and nominated RTWC. In this way, the intervention could be delivered pragmatically alongside current care while also stipulating the achievement of specified POs against which the fidelity of the intervention could be assessed. The investigators were then able to support local implementation through training sessions and specific training resources tailored to the roles and clinical areas.

At each site, potential users of the intervention were identified using eligibility criteria similar to those used in phase 1: (1) placed on the waiting list for hip or knee replacement during their outpatient appointment with the surgical team, (2) in work in the 3 months prior to being added to the waiting list for joint replacement and (3) intending to return to work following surgery.

Training

The OPAL training for hospital staff was undertaken between May and July 2018. Different approaches were required at each site; training logs are provided in Appendix 9. At the Nottingham and Middlesbrough sites, visits were conducted on a number of occasions with staff groups from a variety of clinical areas, to ensure coverage within the department. In Middlesbrough, two or three large-group training sessions were held. In Nottingham, a more labour-intensive one-to-one approach was needed after an initial group session failed to include all of the required personnel. In contrast, the Northumbria site invited ‘key’ staff to the training (including the PI, nominated RTWC and members of the local research team). The RTWC then facilitated the implementation of the OPAL intervention at this site through a systematic series of training with the local team, collating the information and materials needed by each team member and cascading this internally as required.

To support training, the patient and staff materials produced to support the OPAL intervention were made available to each site as ‘site files’ that could be stored and used for reference in clinical areas delivering the intervention (outpatient departments, wards and the RTWC office). To supplement these materials, specific training resources were created, including worked examples of completed study paperwork and fit notes, and training PowerPoint slides. A full list of training resources can be found in Appendix 9. For examples, see OPAL examples of developed materials at www.journalslibrary.nihr.ac.uk/programmes/hta/152802/#/ (accessed 2 May 2020).

Performance objectives to support implementation

As part of the implementation strategy, the OPAL investigators identified a number of POs required to facilitate adoption at each site.

The POs were to:

• recruit the participating sites and PIs

• recruit the RTWC

• train staff in delivering the OPAL intervention

• support the staff in delivering the OPAL intervention.

The POs for the local site PI and RTWCs were to:

• familiarise themselves with the OPAL intervention and associated materials

• inform and enthuse their HOT about the study, and promote engagement

• identify potential RTWCs and deputies (PI role)

• identify surgeons willing to support their patients’ involvement in the study

• arrange training events/meetings between the HOT and the OPAL investigators

• set up a helpline for patients to access a RTWC

• support the team in identifying existing staff members to act as OPAL champions for their subteam (e.g. outpatient clinic, ward and therapy team).

As this was a feasibility study, the implementation strategy focused around adoption for a 6- to 8-week period during which the HOT was to deliver the intervention. Therefore, it was not necessary to identify future adopters and maintainers of the programme.

Methods

The methods used for the ‘feasibility’ stage were similar to the methods used in IM stage 1. 54 Questionnaire and interview data were collected from patients returning to work after hip or knee replacement. Assessment of the intervention considered four inter-related themes:

1. Assessment of intervention fidelity (were the stated POs delivered?). Data collected from participants were mapped against the final staff and patient POs. Evidence was collected from a variety of sources including the baseline and follow-up questionnaires, the patient RTW workbooks (evidence of patient activity) and the RTWC checklists (evidence of RTWC activity) for each patient. All POs were assessed except for PO10 (patient communicates with employer regarding surgical outcome and progress/recovery, by telephone, e-mail or face to face) and PO13 (patient adheres to postoperative rehabilitation plan and advice) owing to an omission on the postoperative questionnaires. Examples of the evidence sources used to determine intervention fidelity for the patient and staff POs are provided in Table 19. A complete description of the evidence sources for all POs is provided in Appendix 9.

2. Assessment of intervention quality (what did patients, staff and employers feel about the intervention and how it was delivered?). Structured interviews explored patient and stakeholder (HOT members and employers) perceptions of the intervention. Interviews explored the understanding, opinions and experiences of the intervention and the study processes associated with its delivery in the context of a research study.

3. Assessment of feasibility data (did the intervention facilitate early supported RTW?). Data collected from the feasibility study (rates and timing of RTW, functional outcomes scores, health utility measures and work-related scores) were compared with similar data collected in IM stage 1 to generate a preliminary comparison of patients who did (IM stage 6) and did not (IM stage 1) receive the OPAL intervention.

4. Assessment of economic data (how much does the intervention cost to deliver and what is the associated health utilisation?). Health-care utilisation data were collected using questionnaires, allowing costs to be assigned to activities. In addition, costs were attached to data collected in the RTWC checklist that documented the time spent by the RTWC supporting the delivery of the intervention. The timing of RTW after surgery in patients receiving the intervention was also explored.

Intervention mapping stage 6 also supported collection of other key information, such as (1) patients’ and surgeons’ views on their willingness to participate in a future trial, (2) potential rates of recruitment and the proportion of eligible patients consenting and (3) information about the behaviour and distributional characteristics of RTW outcomes that would inform the power calculation for any subsequent trial. It therefore captured data that allowed recommendations about the conduct of a future trial to be made.

Feasibility study

Participants were recruited from three sites (Middlesbrough, Nottingham and Northumbria). Participants were asked to complete questionnaires at baseline (preoperatively, when listed for surgery) and at 8 and 16 weeks after surgery as in IM stage 1, to allow comparison with data collected during this stage. The aim was to recruit 30 patients.

All patients attending hip and knee replacement clinics at the study sites were screened, prior to their surgical appointment, using the developed occupational checklist (see OPAL examples of developed materials at www.journalslibrary.nihr.ac.uk/programmes/hta/152802/#/; accessed 2 May 2020). This checklist was taken into the appointment, providing the surgeon with information about their work and work-related activities that could be considered as part of the decision to offer surgery. Where patients were offered surgery, the surgeon confirmed eligibility and facilitated referral to the local research team for further information about the OPAL study, for consent and enrolment. Patients who consented were then provided with the OPAL resources (patient and employer workbooks, and website access) and their contact information was passed on to the RTWC. The design of the feasibility study is shown in Figure 16.

FIGURE 16.

Feasibility study design.

Results

Feasibility recruitment commenced on 1 June 2018 and ended on 14 August 2018. A total of 147 patients were screened, of whom 35 (24%) were eligible for recruitment. In total, 26 of a planned 30 patients were recruited (11 in Northumbria, 8 in Middlesbrough and 7 in Nottingham) (Figure 17). Consent forms, contact details forms and baseline questionnaires were received from all 26 participants. Four participants were withdrawn from the study for the following reasons: participant’s care transferred to a neighbouring trust (n = 1), surgery deferred until later in the year at the participant’s request (n = 1) and participant requested withdrawal from the study (n = 2). A further participant had their surgery delayed for medical reasons and was still awaiting surgery when follow-up for the feasibility was closed (surgery date 22 March 2019). A total of 21 participants were included in the analysis. Follow-up data were received from 18 participants at 8 weeks and from 14 participants at 16 weeks. Copies of the patient workbooks and RTWC checklists were received for 10 and 19 of the 21 remaining participants, respectively. The results are presented under the headings of the four inter-related themes described in Methods.

FIGURE 17.

Flow of patients through the feasibility study.

Assessment of feasibility data

The participants in the feasibility phase were similar to those in the cohort phase; participant characteristics, job titles and details of the activities relating to their jobs can be found in Appendix 9. The flow of participants through the feasibility study is shown in Figure 17. The average questionnaire return rate was 69.6% (see Appendix 9).

Interaction with the intervention

At baseline, most participants completed the occupational checklist prior to their appointment with the surgical team and stated that it was referred to during their consolation (76.9% for both); on average, it took 10.7 minutes to complete. Twenty-two (84.6%) participants talked about their job when discussing the options for treatment, 23 (88.5%) stated that a surgical team member mentioned the OPAL programme, but only 18 (69.2%) discussed how and when they might return to work. All but one of the participants had the OPAL programme explained to them. These results suggest that the initial introduction to the OPAL programme was implemented.

By contrast, only two participants stated that they had contacted the RTWC by telephone following their operation (4.8% of those followed up), and two had used the website. Seventeen of the eighteen (94%) responders at 8 weeks had completed the workbook. Patients took an average of 38.6 minutes to complete the workbook (range 5–90 minutes). Twelve of the 18 (66.7%) participants found the workbook helpful, and nine (50.0%) said that it helped them to develop a RTW plan. Full details can be found in Appendix 9.

Returning to work

At 8 weeks, 7 of the 17 responders had returned to work. At 16 weeks, a further three participants had returned to work. Overall, 38.5% (10/26) of the participants in the feasibility phase stated that they had returned to work; however, 26.9% of the participants (including those who were not followed up) provided no data for this question. Of the returnees, there were five hip and five knee replacement patients.

The average return time was 7.4 weeks (ranging from 0.6 to 17.7 weeks) (Figure 18). This was approximately 2.6 weeks on average earlier than in the cohort phase; however, it should be noted that the sample size here was significantly smaller than in the phase 1 cohort study. There was a difference in return times for the two types of operation, with hip patients returning on average 5.2 weeks after surgery, compared with 9.7 weeks for knee patients. Of those who returned to work, seven (70.0%) did so with reduced hours, on average reducing to 13.7 hours in their first week back at work (ranging from 6 to 20 hours). Full details can be found in Appendix 9.

FIGURE 18.

Bar chart of time to RTW after surgery for the participants in the feasibility study.

Fit notes

On average, participants requested 1.6 (range 1–3) additional fit notes after discharge. The average length of fit notes reported at 8 weeks was 6.1 weeks (ranging from 0 to 12 weeks) (see Appendix 9).

Oxford Hip and Knee Scores

At baseline, the average OHS and OKS were 17.4 and 17.3, respectively; this was comparable to the phase 1 cohort study. This increased to 35.6 for hip participants and 29.2 for knee participants 8 weeks post surgery, and to 43.0 for hip and 29.4 for knee participants at week 16. These follow-up results were similar to those in the phase 1 cohort (see Appendix 9).

Work Limitations Questionnaire

As in the initial cohort study, participants completed the Work Limitations Questionnaire135,153–155 at each time point, to indicate time lost at work. At baseline, there was an average of 41.4% time loss due to their joint problems; this decreased to 23.1% at week 8 and 17.6% at week 16. These results are similar to those seen in the phase 1 cohort (39.4%, 16.8% and 16.8% across the corresponding time points).

Self-efficacy

The General Self-Efficacy Scale156 was included in the feasibility phase baseline questionnaire. This was included on the advice of Trial Steering Committee members as it was a variable felt to be important that we failed to collect in the phase 1 cohort study. This scale ranges between 10 and 40, with higher scores indicating higher self-efficacy. The participants had an average score of 32.6 (range 22–40), implying good self-efficacy; the full results can be found in Appendix 9.

Readiness to return to work

The Readiness to Return to Work Scale157 was included in the follow-up questionnaires. This scale has two sections: one for those already back at work (13 questions) and one for those who are not yet there (nine questions). The scale asks participants to indicate how well they agree with a selection of statements, from ‘strongly disagree’ to ‘strongly agree’, and they are scored on a five-point Likert scale. Full results can be found in Appendix 9. In the case of those not back at work, the results indicated that participants wanted to get back to work, thought that it was possible and were working towards achieving it. At week 8, 50% of the respondents stated that they did not think that they were ready to go back to work, and 40% had not yet set a date for their return. Of those who had returned to work, the responses indicated that they were working towards staying at work, had found strategies to make it possible to be at work and were not concerned about having to take more time off.

Is a future clinical trial feasible?

The feasibility study demonstrated that it was possible to collect relevant data to answer the research question and that an economic analysis could be conducted alongside a future definitive trial. The OPAL study has defined and clarified the following key components for a future trial.

Population

Information collected about patients returning to work after hip or knee replacement defined the target population for a future randomised control trial. The cohort study, structured interviews, evidence synthesis, Delphi consensus process and feasibility study provided information that allowed this population to be characterised and their needs assessed, and barriers to and facilitators of RTW after surgery to be defined. The findings confirmed that this population had varied and complex needs, which supported the use of an individualised approach to managing their RTW. Information collected from key stakeholders (employers, surgeons, AHPs, nurses and GPs) generated a more complete picture of this patient group. These interviews demonstrated that health-care teams and workplaces may not be prepared to, or understand how to, facilitate their patients’/employees’ RTW after surgery. The information generated has supplemented the previous literature20–49 to further define the target population that would benefit from an occupational advice intervention.

Intervention

An intervention was developed that addressed the key aspects of the commissioning brief; that is, it was based in hospital, started prior to surgery, was individualised, provided target support, was proactive and was designed in a pragmatic fashion to support delivery alongside routine care in an NHS setting. The intervention was designed using an iterative process using the IM approach,55–59 which allowed the OPAL investigators to revisit, revise and adapt the intervention as new information became available.

The intervention had a strong theoretical background and was underpinned by biopsychosocial models that supported behaviour change in the target groups (patients and stakeholders in the RTW process). It was manualised as a set of patient and staff POs that defined its content, format, delivery and timing while maintaining pragmatism in the ability for participating sites to administer the intervention alongside standard care. Central to the intervention was the development of an interactive patient workbook that supported the self-directed development of a RTW plan, similar to other recently developed RTW interventions. 166 Implementation during the feasibility stage highlighted specific issues relating to the introduction of a complex intervention. The OPAL intervention was generally well received by patients and stakeholders: positive feedback was received throughout and the developed study materials were reported to be informative and helpful. There was good evidence from the completed patient workbooks that the intervention supported individualised care and, through the development of a RTW plan, acted as a decision aid167–170 enabling shared decision-making in line with best practice. 171,172 The intervention also shared many of the characteristics of the occupational advice interventions identified in the rapid evidence synthesis, including advice about job accommodation; mechanisms to support workplace visits and contact with the employer; education and advice; counselling and guidance through the RTWC; and involvement of the multidisciplinary team.

Furthermore, the OPAL intervention has similarities to another recently tested occupational advice intervention. 166 Grunfeld et al. 166 recently reported on the feasibility and acceptability of a theoretically led workbook intervention designed to support patients with cancer returning to work and confirmed the feasibility of undertaking a definitive trial in this setting.

There is some overlap between the content of the written materials developed in OPAL and those available through the Royal College of Surgeons of England website (example at www.rcseng.ac.uk/patients/recovering-from-surgery/total-hip-replacement/returning-to-work; accessed 26 May 2020). The Royal College of Surgeons of England provides generic written resources covering recovery after both hip and knee replacement. Within these are sections providing information about RTW and timelined guides for recovery after surgery. Because they are designed for all patients, they do not provide the level of detail available in OPAL and cannot provide the individualised support that our phase 1 interviews suggest is needed. The structure, format and delivery of the OPAL intervention has been specifically designed to empower patients to take responsibility for their RTW and provides tools for them to develop an individualised RTW plan. It also encourages active engagement with employers and health-care teams via the OPAL booklets and RTWC role.

Comparison

Usual care that would be the comparator in a future definitive trial was evaluated in the cohort study, patient and stakeholder interviews and survey of practice. This demonstrated a haphazard approach to the delivery of RTW information and significant variation in the way preoperative services were configured. Most patients received little or no information about RTW from their HOT or GP, and only one-third of patients had access to occupational health support at work. These findings were counter to best practice as defined by NICE guideline 138171 and NICE quality standard 15172 (patient experience), which describe the need to ensure that ‘patients experience care that is tailored to their needs and personal preferences, taking into account their circumstances, their ability to access services and their coexisting conditions’ (© NICE 2012. Patient Experience in Adult NHS Services. 172 Available from www.nice.org.uk/guidance/qs15. All rights reserved. Subject to Notice of rights. NICE guidance is prepared for the National Health Service in England. All NICE guidance is subject to regular review and may be updated or withdrawn. NICE accepts no responsibility for the use of its content in this product/publication). Many patients did not have workplace contact until they returned to work. Despite this, a significant proportion of patients felt that current care was sufficient. This may reflect the fact that they returned to work (if they returned to work, this was seen as success) rather than any indication of the quality or timing of their RTW.

Outcome

The measurement of RTW is complex and the evidence synthesis demonstrated that there is currently no standardised method of recording it. Different approaches to measuring RTW were explored during the Delphi process and suggested that a number of complementary measures are needed. Dichotomous recording of work status is blunt and does not address important aspects of behaviour, such as how patients return, use of phased and adapted returns, timing of RTW and secondary sickness absence. Measures such as the Work Limitations Questionnaire135,153–155 and Readiness to Return to Work Scale157 were assessed during the feasibility stage. They provided useful information about time lost at work and information about where patients felt they were in their RTW process.

Study delivery and design

Approximately one-quarter of patients approached for the cohort and feasibility studies were eligible for inclusion, and consent rates for eligible patients were > 80%. This suggests that there are substantial numbers of patients willing to participate in research examining RTW after hip or knee replacement. The response rates for the questionnaires, which reduced with extended follow-up as participants returned to work, highlight a need to put additional efforts into improving the proportion of participants who return questionnaires.

The utility and resource use measures that were included in the questionnaires appear to have been appropriate for the purpose of collecting the health-related quality-of-life and cost data that fed into the economic analysis. The responses to such resource use and EQ-5D questions can be used to inform and improve the design of questionnaires in future research. For instance, participants reported that they had minimal or zero resource use for some items, which could be considered for removal from future questionnaires in order to reduce questionnaire completion burden for participants. Mean utility scores and EQ VAS scores increased over time, in line with what would be expected, which supports the face validity of using the EQ-5D measure.

Health-care resource use was broadly similar for the cohort and feasibility participants, with common key cost drivers; the most notable cost difference was for knee replacement participants over the period of baseline to 8 weeks, when the costs associated with inpatient stay and day-case visits were higher for cohort participants than for those in the feasibility study (who received the intervention). The cost associated with productivity loss following surgery was lower for the feasibility participants (£5929 per feasibility participant vs. £7983 per cohort participant). This cost saving has the potential to offset the cost associated with the intervention and the difference in health-care resource utilisation. It should be noted that, because there were only a small number of participants in the feasibility element, and the feasibility and cohort groups were not randomised to facilitate meaningful comparisons, firm conclusions could not be drawn here. However, a definitive trial would enable robust conclusions to be made regarding the cost-effectiveness of the RTW intervention.

The cost of training, which fed into the intervention cost (as part of the feasibility study), was based on the time spent by the RTWCs at the participating sites. It is acknowledged that this is a simplification of the costs involved, but owing to the variation in training across the sites and additional data not being available, the RTWC time was the focus for the training cost. As part of a definitive trial, more detailed information could be collected in order to estimate the training cost more accurately.

Other research recommendations

There are a number of opportunities for further research. Further research to define the optimal method of implementing the OPAL intervention would be essential before the intervention is formally tested in a RCT (see further discussion in Strengths and limitations).

Additional research could focus on how the intervention might integrate with primary care. This could allow RTW planning to start earlier and may provide support to additional patients seeking to remain in work who have symptomatic osteoarthritis below the threshold requiring joint replacement. If proven to be effective in a definitive trial, there is also the potential to investigate how the OPAL intervention could be implemented in other elective orthopaedic surgery and other surgical specialties undertaking planned surgical procedures. A significant proportion of the developed intervention has transferable content and the needs assessment conducted in OPAL is likely to be generalisable more widely to other specialties and settings, making this an important area for future research.

Intervention mapping stages 1–4

Because our evidence synthesis followed rapid review methodology, there was a restriction on the range of databases that were searched. However, our searches were undertaken by an experienced information specialist in order to capture the most relevant databases, given this restriction. Preliminary results from the rapid evidence synthesis found only four papers for the elective surgery population. This finding was discussed with the Trial Steering Committee, which advised including the musculoskeletal literature even though this included patients with a range of chronic musculoskeletal conditions, not representative of our target population. This approach yielded useful results that were applicable to our hip and knee replacement patients.

We were initially unable to set up the cohort study within the planned time scales in the three originally proposed sites. To mitigate issues posed by the delays, we opened an additional site, which helped achieve our target recruitment and provided additional support during the feasibility study. Despite approaching all hip and knee replacement patients awaiting surgery at the study sites, we were able to identify only six patients intending to retire after surgery; this was a group we intended to examine. As a consequence, this part of the analysis was not carried out. Interestingly, 10% (n = 9) of patients at the 8-week follow-up stated that they no longer intended to return to work. It may be that these patients intended to retire after surgery but, for reasons that are unclear, did not state this during the baseline assessment.

The cohort study followed all patients until 16 weeks post surgery and a subset of patients for 24 weeks. Only 50% (n = 78) returned to work within the follow-up period. Extended follow-up studies have shown improvements in the observed rates of RTW;44,45 however, we were restricted by the study duration. Extended follow-up would be useful as part of a larger trial as it could evaluate other complementary aspects of RTW. These include evaluating RTW over time to understand whether or not it was sustained, periods of secondary sickness absence and whether or not patients returning to work on phased or adapted returns get back to full duties and the time frame for this. The study participant group included only a small number of black and minority ethnic patients. Their views and experiences were therefore not adequately represented in the cohort, patient interviews and feasibility testing, and the findings of these elements may not be generalisable to these patient groups. Consideration needs to be given to how black and minority ethnic patients are represented in a future trial.

Follow-up rates for the cohort study were 75% (n = 104) at 16 weeks and were lower at other time points. Similar issues with follow-up and dropout were observed during the feasibility study. Once patients had returned to work, they disengaged with follow-up procedures, instead seeming to focus on their work role. This may be a specific issue when conducting research on working patients who are possibly least likely to have time to participate in research and needs to be considered when designing a future study. We had initially hoped to use data to stratify patients into high and low risk of failed RTW as part of a tiered intervention. However, the failure to identify predictors, alongside the results from the Delphi consensus group and advice from the Trial Steering Committee, led us to design an intervention for everyone. The advantage of this approach was that the intervention was available for all patients, allowing engagement based on their individual need (described in Chapter 9).

The interviews demonstrated that current ‘usual care’ was frequently not following best practice in relation to the use of workplace adjustments and partial RTW, and the use of the fit note to advise on this. This suggests that there is room for improvement in how patients are supported in returning to work. Unfortunately, we were able to interview only a small number of self-employed patients. This meant that we were unable to provide a comprehensive commentary on the needs and behaviour of this group and to investigate whether or not they behave differently to those in other types of employment (e.g. because of lack of sick pay, pressure to get back to work sooner than employed patients, more options for workplace adaption and phased returns for employed people). Recruiting employers and GPs to interview was difficult. However, using a variety of strategies,173 we were able to achieve the recruitment required in IM stage 1, although we were unable to replicate this in the feasibility study. Whether or not interviews are conducted on a group or one-to-one basis is likely to change the dynamics of the interaction between researchers and participants. This may have influenced the nature of clinicians’ contributions and the data collected, and may thus be a limitation of the study.

Overall participation in the Delphi consensus process decreased from round 1 to round 3. Attrition through the Delphi process is well recognised;151,174 hence, various strategies were employed that were known to enhance response rates. 175,176 By round 3, fewer than 20% of our invited participants remained, and employers and GPs were poorly represented in the final two rounds. Round 3 responses comprised feedback from only one employer and one GP. To mitigate the potential response bias introduced by a stakeholder group being not represented,177 an employer representative (Unison, London, UK) was approached, who provided structured feedback on the intervention outside the Delphi process. This information was used alongside the comments from other participants to finalise the intervention prior to feasibility testing.

Intervention mapping stages 5 and 6

OPAL is a complex intervention that required considerable planning for its implementation and sufficient time to put in place the facilitators to embed it into practice and to remove potential barriers to its effectiveness. 178 Having developed the intervention using the IM framework, it was extremely challenging to effectively implement it within the time scale of the feasibility study. For it to be successfully embedded at each site, there was a need to involve all members of the HOT. However, in reality, owing to the limited time available for implementation and feasibility, it was difficult to train all staff and implement the intervention as intended. Consequently, some staff had no training or did not receive the training as intended.

The feasibility study suggested that patient experience of the intervention was positive and there was high adherence to the patient POs. However, some of the staff POs showed lower adherence. This may reflect that data sources were unable to ascertain if these POs had been adhered to but also may reflect the challenges of incorporating new behaviours and procedures into current health-care professional roles. It suggested that some of the implementation processes presented challenges and barriers to effective adoption. For future implementation and research studies, these barriers to implementation could be viewed within the Consolidated Framework for Implementation Research,179 which could help with understanding and overcoming them.

Throughout OPAL, and particularly in the Delphi study, we found evidence of reluctance among health-care professionals to take on the role of the provider of occupational advice: an ‘anyone but me’ attitude. This reluctance suggests that sufficient time needs to be provided to change the attitudes, norms and behaviours necessary to embed the roles and responsibilities for occupational advice in the OPAL intervention. Disappointingly, we were unable to interview GPs and therapy team members during the feasibility stage, which might have provided greater detail about their attitudes towards the OPAL intervention and might help to identify potential improvements. Time is also needed to prepare patients’ expectations with respect to the provision of RTW advice as part of routine health care.

The RTWC role was not fulfilled in all settings in the way we had expected. It is a key role in the OPAL intervention, delivering or facilitating a number of the POs. It would take a considerable time for someone to adapt and learn the competencies required for this role unless they had been recruited to specifically fulfil the required criteria for the role. There were contrasting experiences of the recruitment of the RTWCs. In one centre, there was uncertainty about whether or not funding was available to backfill posts, meaning that the local team found it difficult to predict how much time and resource would be required. By contrast, at another centre an experienced and enthusiastic senior nurse with a background in patient experience could be identified. Interviews suggested that some RTWCs accepted the role with reluctance. RTWCs were also required to be good clinical practice trained because the role was in the context of a research study, causing delays in their appointment. The feasibility study demonstrated that some members of staff were not fully committed to the intervention. However, evidence from the feasibility interviews suggested that if the occupational advice intervention were to be an agreed and funded component of routine treatment, a greater level of acceptance and adherence to delivery would be expected.

During the feasibility study, some of the trained staff rotated to other departments (junior doctors and AHPs), and others went on leave. This highlighted the need for ongoing training if the intervention was to become embedded and sustained in a department. In addition, some departments (e.g. outpatient departments) were reluctant to take on additional duties as they were already ‘overstretched’.

The barriers described above relate to the ‘readiness for implementations’ in the inner settings constructs of the Consolidated Framework for Implementation Research,179 suggesting that increased commitment and engagement among staff need to be ensured prior to implementation of the OPAL intervention. Given the time scales for implementation in the feasibility study, it is not surprising that the intervention was not fully embedded and that there were signs of a lack of commitment, resource issues and lack of awareness in some teams. However, these are all factors relating to the inner setting for implementation that could be improved on for a larger trial or full adoption of the intervention in a service. There was also variation among patients in understanding and perception of the aims of the OPAL intervention, which suggests that the intervention may need to be presented more clearly. This links into the Consolidated Framework for Implementation Research179 concept of intervention design quality and packaging, and is an area for further consideration beyond the OPAL study.

In hindsight, it may have been beneficial to pause the project after IM stage 4 once the final intervention had been drafted. This would have allowed time to develop a more robust implementation strategy that provided the necessary training and support to deliver individual, organisational and cultural change in the local orthopaedic teams. However, this was not an option given the protocol and the need to make recommendations about the feasibility of a future clinical trial as per the original commissioning brief. Other studies have similarly reported difficulties with applying IM stage 5. In a systematic review of interventions to prevent work disability developed using an IM approach,63 reviewers were unable to report the IM stage 5 outcomes because they were insufficiently reported in the studies they reviewed.

It is important to consider how a future study would evaluate the cost-effectiveness of a RTW intervention in terms of the study design and data collection considerations. The OPAL study found challenges around obtaining accurate cost information (e.g. cost of training) and achieving sufficient participant numbers to allow meaningful conclusions around the cost-effectiveness findings. Rather than the feasibility and cohort groups that were summarised here, a future definitive study should aim to randomise participants using a robust RCT design to enable a full comparison to be made.

Search strategies for systematic reviews