The work of return to work. Challenges of returning to work when you have chronic pain: a meta-ethnography

Aims To understand obstacles to returning to work, as perceived by people with chronic non-malignant pain and as perceived by employers, and to develop a conceptual model. Design Synthesis of qualitative research using meta-ethnography. Data sources Eleven bibliographic databases from inception to April 2017 supplemented by citation tracking. Review methods We used the methods of meta-ethnography. We identified concepts and conceptual categories, and developed a conceptual model and line of argument. Results We included 41 studies. We identified three core categories in the conceptual model: managing pain, managing work relationships and making workplace adjustments. All were influenced by societal expectations in relation to work, self (self-belief, self-efficacy, legitimacy, autonomy and the meaning of work for the individual), health/illness/pain representations, prereturn to work support and rehabilitation, and system factors (healthcare, workplace and social security). A mismatch of expectations between the individual with pain and the workplace contributed to a feeling of being judged and difficulties asking for help. The ability to navigate obstacles and negotiate change underpinned mastering return to work despite the pain. Where this ability was not apparent, there could be a downward spiral resulting in not working. Conclusions For people with chronic pain, and for their employers, navigating obstacles to return to work entails balancing the needs of (1) the person with chronic pain, (2) work colleagues and (3) the employing organisation. Managing pain, managing work relationships and making workplace adjustments appear to be central, but not straightforward, and require substantial effort to culminate in a successful return to work.

The introduction spells out the rationale for the study leading to clearly defined aims. The concept and use of the metaethnography methodology is well described as are the methods used to conduct the systematic review. The findings are well presented with conceptual categories derived from the papers identified through the systematic review, with a conceptual model developed to illustrate the relationship between these categories. There is a good discussion, setting the findings in the context of other literature with acknowledgement of the limitations of the current work and consideration of the implications.
The authors state that the search terms used "included...", it would be helpful if the search strategy could be included either in a supplementary file or a note added to the text that the strategy is available on request.
The analysis of papers was principally carried out by one author, with two others checking the first ten papers (alphabetically). It is unclear whether this was a percentage of the papers identified or the rational for selecting just 10 papers, could the authors justify this approach please?
It would also be helpful to have a sense of how much agreement there was in identifying the third order concepts, it is reported that this took place over three meetings but was there any overlap in the concepts initially identified?
The CASP tool was used to appraise the included papers, it is noted that all included studies passed the first two screening questions, were papers excluded if they did not pass these screening questions?
There is a short section on the implications of the paper, I feel that this section would benefit from some further consideration. The authors suggest developing a return to work intervention to explore ways of addressing and managing chronic pain, managing work relationships and making workplace adjustments, but who would deliver such an intervention? Should it be located in healthcare, with the employer, another agency, or a combination? What are the likely implications of such an intervention on healthcare services and employers (improving health/changing workplace cultures)? Line 144 -how was the focus endorsed by the representative, when reading that it makes me want to have more information. Starting line 168 -There needs to be more information given here on the relevance of the three key conceptual categories as this is missing. You have a great explanation of them in lines 260-266 at the beginning of the discussion but I think a bit more is needed earlier. This paragraph from the discussion would even fit well in this first section of the overarching conceptual categories to give the reader a good explanation of the relevance and importance of the three key conceptual categories.

REVIEWER
Figure 2 -conceptual model -This is a good model but it does seem unnatural going anti-clockwise and then this needs to be identified as you have done as anti-clockwise.
Line 303 -instead of saying 'carefully reading and re-reading abstracts and then full texts' it might have more consistency if you said for example 'by applying the seven steps of the metaethnography process'. Otherwise there is a potential confusion over what methodology was used. Line 340-341-the sentence on the influence of health and pain is overly wordy and confusing, but this isn't my subject area of expertise so it might be correctly pitched.
Thank you for the opportunity for reviewing this interesting metaethnography. in an open access format in several journals. I will draw on the eMERGe Reporting Guidance in writing this review. The authors of the current paper (to be referred to as 'synthesis authors' to avoid confusion later) may be helped by the explanations of the 19 reporting criteria given in the Guidance. The paper under review also used the GRADE-CERQual guidance with which I am less familiar to assess the strength of evidence in different conceptual categories. The rationale and context for using the method of metaethnography is clearly described as are its aims and focus. In particular the synthesis authors cite knowledge gaps in relation to return to work by people with chronic non-malignant pain, and the perspective of employers. However readers of BMJ Open may also be interested in the perspectives of doctors: perhaps the synthesis authors could at least indicate if there is any qualitative literature on this topic. In relation to the aim of the synthesis, the synthesis authors have gone beyond 'understanding obstacles' (abstract, page 2) and have provided a conceptual model of return to work. Thus it might be appropriate to revise the aim of the synthesis to reflect its more ambitious outcome (eg 'to provide a conceptual model'). The synthesis authors explain why they considered meta-ethnography to be the most appropriate qualitative synthesis methodology, saying that they wished to 'identify any other overarching concepts that would explain the data' (page 5). The synthesis authors list the databases searched and the search terms used but they do not describe the rationale for their search strategy (for example, why these databases were considered the most appropriate ones to use). They seem not to have used a search strategy specifically developed for qualitative studies (as discussed by Booth 2016 for example), which might perhaps have identified other qualitative papers. They do not say who conducted the screening and selection of papers although possibly these were the same people who did the searching. Some of the exclusion criteria are given in the results section.

REVIEWER
The results of the searches and screening are clearly and fully set out in figure 1. The synthesis authors explain who read the 41 included papers, how the data were extracted and the conceptual categories identified. The synthesis authors state that 'the concepts are ideas drawn from the findings of the original papers' (page 6) though they do not distinguish between participant quotes and authors' findings (often referred to as first and second order concepts, see for example Britten et al, 2002). The first column in table 2 seems to contain the synthesis authors' summaries (often referred to as third order concepts) and the second column contains participants' quotes (first order concepts). Thus it is not clear if the synthesis authors have worked with second order concepts. The characteristics of the included studies, and the data extracted from each study, are clearly set out in Table 1. However there is no discussion of which aspects of the included studies were compared, nor how the studies were compared. Noblit and Hare (1988), whose approach the synthesis authors cite, refer to reciprocal and refutational syntheses, but it is not clear that the synthesis authors considered the methods of translation in these terms. Thus, point 13 of the eMERGe Reporting Guidance referred to above asks about the steps taken to preserve the context and meaning of relationships between concepts within and across studies; how the reciprocal and/or refutational translations were conducted; and how potential alternative interpretations and explanations were considered. It might have been helpful, for example, to group the 5 papers covering employer perspectives separately; or to consider if the differences between employees and employers merely reflect different perspectives or if they constitute refutational translations. The three key conceptual categories are described in the text while the remaining 13 categories are described in the supplementary material, no doubt for reasons of space, as well as briefly in table 2. The synthesis authors describe the ways in which they developed overarching concepts and in particular how three of them independently developed their own conceptual models before comparing them and eventually agreeing a final (11th) version. It is not stated if this process involved the consideration of potential alternative interpretations or explanations (point 15 of the eMERGe Reporting Guidance). The synthesis authors say that they worked in a safe team environment in which they were able to disagree with each other: if there were disagreements about alternative interpretations, these might be usefully reported. The new conceptual framework is clearly set out in figure 2 and described in the section entitled 'Line of argument' on page 31. The main findings are summarised and compared to the existing literature in the Discussion section. Some of the limitations of the synthesis are described but perhaps more could be said about reflexivity and the impact of the research team on the synthesis findings. The recommendations are succinctly described, although they provide no guidance for clinicians who manage patients with chronic pain and have to sanction their return to work. Overall, the synthesis is reasonably well reported. As the eMERGe Reporting Guidance has only just been published, it may be unfair to expect the synthesis authors to address all 19 points in the Guidance. The reporting guidelines are not themselves directly concerned with the quality of the synthesis, although a reasonably well reported synthesis such as this one is likely to have been conducted well. In terms of quality, the synthesis does seem to have been well conducted. However I think that the synthesis authors could have gone further in their analysis.

Reviewer 1 Comment
The authors state that the search terms used "included...", it would be helpful if the search strategy could be included either in a supplementary file or a note added to the text that the strategy is available on request.

Response
We have added the search strategy as Supplementary File 1.

Comment
The analysis of papers was principally carried out by one author, with two others checking the first ten papers (alphabetically). It is unclear whether this was a percentage of the papers identified or the rational for selecting just 10 papers, could the authors justify this approach please?

Response
We have added justification for our approach (lines 134-6).

Comment
It would also be helpful to have a sense of how much agreement there was in identifying the third order concepts, it is reported that this took place over three meetings but was there any overlap in the concepts initially identified?
Response This is addressed in lines 129-130.

Comment
The CASP tool was used to appraise the included papers, it is noted that all included studies passed the first two screening questions, were papers excluded if they did not pass these screening questions?
Response This sentence has been rephrased to clarify meaning (lines 164-5).

Comment
There is a short section on the implications of the paper, I feel that this section would benefit from some further consideration. The authors suggest developing a return to work intervention to explore ways of addressing and managing chronic pain, managing work relationships and making workplace adjustments, but who would deliver such an intervention? Should it be located in healthcare, with the employer, another agency, or a combination? What are the likely implications of such an intervention on healthcare services and employers (improving health/changing workplace cultures)?

Response
This section has been developed and lengthened to answer the above questions (lines 372-385).

Reviewer 2 Comment
Line 136 -not sure if identifying the 11th version is necessary or relevant. You have identified that there were several revisions and I believe that is adequate.

Response
We have deleted the words "(the 11th version) which"(now line 146).

Comment
Line136-139 -the sentence describing the methodology used by Toye et al doesn't read well, even just removing the 'and this' would help.

Response
We have deleted the words "and this" (now line 148).

Comment
Line 144 -how was the focus endorsed by the representative, when reading that it makes me want to have more information.

Response
This sentence has been edited to provide more information (now line 157).

Comment
Starting line 168 -There needs to be more information given here on the relevance of the three key conceptual categories as this is missing. You have a great explanation of them in lines 260-266 at the beginning of the discussion but I think a bit more is needed earlier. This paragraph from the discussion would even fit well in this first section of the overarching conceptual categories to give the reader a good explanation of the relevance and importance of the three key conceptual categories.
Figure 2 -Conceptual model -This is a good model but it does seem unnatural going anti-clockwise and then this needs to be identified as you have done as anti-clockwise.

Response
An anticlockwise reading of the model was chosen so that concepts flowed, ending in a left to right direction. So we would like to keep it clearly identified as such, as we have done.

Comment
Line 303instead of saying 'carefully re-reading abstracts and then full texts' it might have more consistency if you said for example 'by applying the seven steps of the meta-ethnography process'.

Response
We agree this sentence is unclear so we have deleted it (now lines 346-7) as steps of metaethnography process are outlined elsewhere.

Comment
Line 340-341-the sentence on the influence of health and pain is overly wordy and confusing, but this isn't my subject area of expertise so it might be correctly pitched.

Response
We have discussed this sentence and we feel it is clear and we would prefer to keep it (now lines 403-4).

Comment
In particular the synthesis authors cite knowledge gaps in relation to return to work by people with chronic non-malignant pain, and the perspective of employers. However readers of BMJ Open may also be interested in the perspectives of doctors: perhaps the synthesis authors could at least indicate if there is any qualitative literature on this topic.

Response
A sentence has been added to introduction referring to qualitative research on this topic (lines 64-5).

Comment
In relation to the aim of the synthesis, the synthesis authors have gone beyond 'understanding obstacles' (abstract, page 2) and have provided a conceptual model of return to work. Thus it might be appropriate to revise the aim of the synthesis to reflect its more ambitious outcome (e.g. 'to provide a conceptual model').

Response
The aim has been revised as suggested in abstract (line 19) and in paper (line 75).

Comment
The synthesis authors list the databases searched and the search terms used but they do not describe the rationale for their search strategy (for example, why these databases were considered the most appropriate ones to use). They seem not to have used a search strategy specifically developed for qualitative studies (as discussed by Booth 2016 for example), which might perhaps have identified other qualitative papers.

Response
The text has been edited to describe the rationale (lines 97-9, 104-5).

Comment
They do not say who conducted the screening and selection of papers although possibly these were the same people who did the searching.

Response
The text has been edited to specify (lines 100-1) who did the searching.

Comment
Some of the exclusion criteria are given in the results section.

Response
This sentence has been rephrased (line 164) to clarify it is results.

Comment
The first column in table 2 seems to contain the synthesis authors' summaries (often referred to as third order concepts) and the second column contains participants' quotes (first order concepts). Thus it is not clear if the synthesis authors have worked with second order concepts.

Response
We did work with second order concepts and this is clarified by editing text (lines 184-6). Also punctuation in column of Table 2 has been changed to differentiate second and first order data and the latter are in now in italics.

Comment
The characteristics of the included studies, and the data extracted from each study, are clearly set out in Table 1. However there is no discussion of which aspects of the included studies were compared, nor how the studies were compared. Noblit and Hare (1988), whose approach the synthesis authors cite, refer to reciprocal and refutational syntheses, but it is not clear that the synthesis authors considered the methods of translation in these terms.

Response
Text has been added to clarify what we did (lines 140-3).

Comment
The synthesis authors describe the ways in which they developed overarching concepts and in particular how three of them independently developed their own conceptual models before comparing them and eventually agreeing a final (11th) version. It is not stated if this process involved the consideration of potential alternative interpretations or explanations (point 15 of the eMERGe Reporting Guidance).

Response
Text has been added to clarify and illustrate this process (lines 129-30).

Comment
The synthesis authors say that they worked in a safe team environment in which they were able to disagree with each other: if there were disagreements about alternative interpretations, these might be usefully reported.

Response
This was clarified in response to a comment from Reviewer 1 (lines 150-3).

Comment
Some of the limitations of the synthesis are described but perhaps more could be said about reflexivity and the impact of the research team on the synthesis findings.

Response
Text has been added to provide further explanation (lines 393-396)

Comment
The recommendations are succinctly described, although they provide no guidance for clinicians who manage patients with chronic pain and have to sanction their return to work.

Response
Text added in response to a similar comment from Reviewer 1 (lines 372-385).

Comment
As the eMERGE Reporting Guidance has only just been published, It may be unfair to expect the synthesis authors to address all 19 points in the Guidance.

Response
The importance of the eMERGE Reporting Guidance going forwards is acknowledged (lines 397-9).

Comment
However I think that the synthesis authors could have gone further in their analysis. In particular, I would have liked to see a more systematic comparison of all 16 conceptual categories which might have provided explanations of the differences between upward and downward spirals, referred to in the last point of the line of argument. Are there ways in which the different conceptual categories work together to enhance return to work, or to inhibit it? Could separate lines of argument be formulated for upward and downward spirals, which might help clinicians in the same way that Malpass et al's (2009) 'decisive moral junctures' provide insights for those managing people taking antidepressants?

Response
We have reflected on and discussed this and feel the line of argument and conceptual model demonstrate the extremely complex interaction between conceptual categories. It demonstrates that people with chronic pain need to be understood, believed, not judged and expectations need to be managed between employees and employers. It also shows that support can be given to manage pain, relationships and make work place adjustments, negotiate obstacles and navigate change but sometimes the person's thinking (health and pain representations), their level of self-belief, what work means to them, family expectations or system factors (e.g. type of job) may mean that even with this support it is not possible to return to work. Our additional section from lines 372 highlights the way in which it is highly individual how different factors can work together to enhance or inhibit return to work, and how these factors need to be assessed by the clinician working with the person with pain and the employer. Having thought about this carefully, for the above reasons we would like to keep our conceptual model diagram as it is.

Comment
The synthesis authors worked with a patient and public representative to develop their funding proposal, but this person seems not to have had any involvement with the synthesis itself or in the writing of the paper. Was such a contribution considered or discussed with the individual involved?

Response
We didn't involve the PPI representative in the synthesis, but would want to do this in future work.
Comment I cannot follow the comment about CERQual made on page 34 and possibly other readers would not be able to follow it either.

Response
We have decided to delete this comment (lines 364-6).

REVIEWER
Gwenllian Wynne-Jones Arthritis Research UK Primary Care Centre, Keele University REVIEW RETURNED 13-Mar-2019

GENERAL COMMENTS
I have enjoyed re-reading this paper, the authors have taken on board, and given full consideration, all the comments received making changes where necessary.

REVIEWER
Hazel Keedle, Lecturer Western Sydney University Australia REVIEW RETURNED 28-Feb-2019

GENERAL COMMENTS
I think the changes you have made to the manuscript have enhanced the paper and that it is now ready for publication.