Why don't patients seek help for chronic post‐surgical pain after knee replacement? A qualitative investigation

Abstract Background Although many people are satisfied with their outcome after total knee replacement surgery for osteoarthritis, around 20% report chronic post‐surgical pain. People are often disappointed and unsure about whether their pain is normal and what can be done about it. Given the high prevalence of long‐term post‐operative pain after knee replacement, there is potentially a large hidden population with an unaddressed need for care. Objective In this study, we focus on understanding why some people choose not to consult health care for chronic post‐surgical pain after knee replacement. Methods Semi‐structured interviews were conducted with people who had received total knee replacement, at either of two National Health Service hospitals in the United Kingdom, and who had chronic post‐surgical pain (n = 34, age 55‐93 years). Data were audio‐recorded, transcribed and analysed thematically. Results We found an overall sense of futility amongst participants who believed that nothing further could be done for their on‐going pain. People's perception of their pain was often discordant with that of surgeons and physicians. Other factors that contributed to decisions not to seek help included low expectations about effectiveness and the risks involved in further treatment, treatment burden, participants' prioritization of other health conditions and views about candidacy. Many accepted their on‐going pain. Conclusion Our study indicates why some people with chronic pain after knee replacement do not seek further health care. Understanding patients' beliefs and expectations about chronic post‐surgical pain can inform approaches that might enable people to seek help in the future.

gained from social networks, relationships with health-care professionals, and socio-political and moral discourses about consultation behaviour. 22 Theoretical models of health and illness behaviour such as the Model of Pathways to Treatment can provide a useful way of understanding help-seeking behaviour. 23 The Model of Pathways to Treatment encompasses existing psychological theory to explain the events and processes that occur during help seeking, including decisions not to seek help. 23 The model describes 'backwards and forwards' bi-directional movements between a series of events and processes ( Figure 1). This begins with the patient detecting a bodily change during the appraisal process which leads to determining a reason to seek help; a decision to consult then leads to the first consultation; the clinicians' appraisal leads to a diagnosis; and finally, the planning of treatment leads to the start of treatment.
The model is underpinned by elements from social cognitive theory. 24 These include self-efficacy-the confidence people have in their ability to bring about desired outcomes, 25 and outcome expectancies-the perceived consequences of an individual's actions, which include incentives and disincentives to help seeking. 24,26 The model has been widely used to understand help-seeking behaviour in cancer populations but previously with people with chronic pain after knee replacement. [27][28][29] Our aim in this study was to understand why some people with chronic post-surgical pain after total knee replacement choose not to seek help. We draw on concepts from the Model of Pathways to Treatment in the discussion to explain these findings.

| Identification of participants
The study was conducted in the United Kingdom with patients who had received total knee replacement at either of two high-volume National Health Service hospitals in Central and South-West England. We sampled purposively, choosing a target of 40 to ensure we achieved a diverse range of participants, including patients who were between 12 months and 5 years post-knee replacement from across both the Central and South-West England regions of the UK.
We aimed to recruit equal numbers across this time frame, by posting out invitations in phases across each 12-month period. We chose this time frame because people can continue to experience some improvement in pain and function up to 12 months post-operatively 7 and we wanted to understand non-use of services for chronic postsurgical pain in the longer term. A clinical team member identified potential participants from patient lists.
Potential participants received an information pack and screening questionnaire about pain and health-care use (see Appendix S1).
The design of the questionnaire and definition for the threshold of inclusion was determined in discussion with pain clinicians, research methodologists and a patient and public involvement (PPI) group.
The questionnaire included elements of the Level of Expressed Need Scale 30 and the complete Oxford Knee Score, 31 which is widely used in national programmes to assess outcomes relating to pain and function. Both scores are validated. Patients were eligible for inclusion if they answered 'yes' to the question: 'Are you currently troubled by pain in your replaced knee, either all the time or on and off, which has lasted for more than 3 months?', and scored between 0 and 14 on the seven Oxford Knee Score pain questions, and described seeing GPs or other health-care professionals in relation to their pain as 'rare' or 'never' in the previous 12 months. The screening questionnaire was discussed with patient representatives who thought that these terms were appropriate.
Eligible patients interested in participating were contacted by a member of the research team [AJM] to arrange a face-to-face interview. All interviews took place in participants' homes. The study was conducted in compliance with the Declaration of Helsinki, and ethical approval was granted by the West Midlands-Solihull Research Ethics Committee (Reference number 15/WM/0469). All participants provided written informed consent.

| Data collection
Interviews were conducted by an experienced qualitative methodologist who was previously unknown to the participants [AJM].
Interviews lasted between 32 minutes and 105 minutes (mean average 57 minutes). Data collection took place between May 2016 and August 2018. A semi-structured topic guide was used to guide discussion, and to enable the researcher and participant to explore and reflect on different areas. Topics included experience of chronic pain after knee replacement, characteristics of pain, comorbidities, selfmanagement and use of formal and informal health services. The guide was designed in collaboration with members of the study's patient and public involvement group (Appendix S2). transcripts. Developing codes were discussed and refined and then applied across the data set and reviewed. Those that shared unifying features or patterns were grouped into themes (Appendix S3). At this point, a thematic map was constructed to illustrate the themes and the presence of an overarching 'core theme' (See Figure 2).

| Data analysis
Our approach was inductive and the codes were developed from and grounded within the data. 33

| RE SULTS
Thirty-four people took part (18 women), all of whom had received a total knee replacement between 14 and 68 months before interview (Table 1). Average age was 74 years, ranging from 55 to 93 years.
A single core theme encapsulated the overall sentiment: 'nothing more can be done' (see Figure 2). Seven related subthemes explain why participants had rarely or never sought health-care for their pain. We include illustrative quotes relating to these subthemes, and all names used are pseudonyms.

| Nothing more can be done
Overall, participants described a sense of futility and a belief that nothing further could be done for their pain. This belief was informed by a combination of experiences and expectations about health care and whether further care was appropriate or possible. These are contained in the seven inter-related subthemes: the response of health-care professionals; expectations about outcomes or risks involved in further treatment; the burden of treatment; prioritizing other health conditions; the moral question of candidacy; characteristics of pain; and acceptance.  Initial consultations with GPs were also seen as unhelpful, as patients described receiving offers of pain medication which they did not want.

| Health
Well nobody seemed to want to help me do anything about it. The GP didn't particularly want to help. He just wanted to give me stronger painkillers (Eric) Others did not consult their GP because they expected that nothing further could be done or offered, having already had a joint replacement.

| Expectations about the outcomes or risks of further treatment
Some participants decided not to engage further with health care because they felt that doing so was risky. Some thought that they might be referred for further surgery, which they wished to avoid for fear it could worsen their outcomes, and in one case even result in amputation.

| The burden of treatment
As well as concerns about the outcomes of further treatment, some participants wanted to avoid the personal disruption caused by surgery and recovery, including incapacitation, being away from home or enduring months of waiting before an operation.
Others felt they were too old to go through the physical exertion of further treatment, especially when they were coping with other health conditions. Oscar had undergone thirteen operations altogether, for his knee and other problems and felt that 'enough is enough'.
You've got to wait probably two or three months

| Prioritizing comorbid conditions
Some participants were also living with conditions that took priority over their post-surgical pain. Benjamin explained that his other knee had become more painful and that he was also living with a terminal lung condition, which was his main focus.
I'm more anxious about my lungs when I'm short of breath. I'm anxious, then I think to myself, 'This is it' […] That, that, to me is my -at the moment, for me, it's the lungs that are priority.
Similarly, Peter's heart problem caused breathlessness and limited his ability to mobilize.
As I said to you, it takes me all the time to get to the kitchen, breathless. Now that is my biggest concern, breathlessness, not me knee. My knee is purely in the background. (Peter) One patient felt that although her knee was painful, she could manage it, and the pain in her foot concerned her more. Clint also suggested that the pain from his replaced knee was 'bearable' compared to that of his other knee which was 'taking over the top spot'.

| Pain characteristics
All participants reported a high level of pain severity during screening before they were invited to interview. During interviews, they were asked about the characteristics of their pain and whether it had changed over time. Some reported that although they still experienced episodes of intense pain, it was considerably better than pre-operatively, which made it easier to cope with.
For some these improvements, although marginal, made on-going pain more acceptable, and were a disincentive to seek further help.

| Acceptance and getting on with it
Many participants spoke of 'just getting on' with their lives, describing an acceptance of their situation because they felt that there was no other choice. With different levels of acceptance, all participants were trying to self-manage their on-going pain.

| Bringing the themes together in the context of the model of pathways to treatment
In the context of the Model of Pathways to Treatment (Figure 1

| D ISCUSS I ON
The overarching theme that 'nothing more can be done' reflects a common experience. Based on their initial experience of followup appointments or seeking support for chronic post-surgical pain many participants had stopped seeking help, for multiple reasons leading to a sense of futility. This experience may be conceptualized as a 'futility loop' (Figure 3), whereby patients become trapped between the appraisal and help-seeking interval, in which they think that that there could be a reason to consult for pain but that nothing further can be done. Toye reported that people with chronic musculoskeletal pain felt that health-care professionals disbelieved their pain leading to them feeling delegitimized. 39 However, patients in our study report that while clinicians acknowledged their pain, it was often normalized.
We also found that some patients avoided further consultations for fear that it might lead to more surgery and worse outcomes.
Previous research highlights that health-care professionals in the NHS report an absence of clear access points into care for people with post-surgical pain. 16 This finding perhaps resonates with participants' experiences in this study and may explain their view that nothing further could be done.
Some participants felt that help seeking involved a personal burden and suggested they were too old to expend further efforts to redress the balance of health and illness. This is supported by previous research which suggests that as people get older they are less inclined to seek help for chronic pain. 40 The sense of personal burden and the context of ageing also resonate with literature that describes the 'hard work' of managing osteoarthritis before surgery and shows how people may reduce self-management efforts as they age. [41][42][43] Our work extends these findings, showing them to be as relevant even after surgical intervention for osteoarthritis.
Many participants suggested they were 'getting on with it' be- are also used to managing pre-surgical knee pain due to osteoarthritis in this way, 42,45 so when surgery fails to alleviate pain, then reverting to similar self-management strategies makes sense. Patients also prioritized some comorbidities, such as pulmonary and cardiac conditions, which were felt to limit function or represent more of a future risk than chronic post-surgical pain, similar to findings from other studies. 46,47 We used the Model of Pathways to Treatment, 23 to provide an explanatory framework for how people seek health care. The appraisal and help-seeking intervals are particularly relevant to participants in our study as all participants had cycled through the pathway as they sought treatment for osteoarthritis pain, ultimately leading to their knee replacement. Following this, some cycled through once or twice as they sought help for pain, but most had either never sought, or had stopped seeking help and had be- that unless patients believe their actions will produce desired effects, for example relief from post-surgical pain, they will be less inclined to seek help because of these outcome expectancies.
Bandura specified three forms of outcome expectations 48,49 : Physical outcome expectancies could disincentivize help seeking if people believe that help seeking could result in further pain, ineffective or unpleasant treatment, or disruption to their lives, such as the patients in our study who did not want further surgery or medication.
Social outcome expectancies involve social sanctioning and reactions to help-seeking behaviour. For example, some participants in our study did not seek help partially for fear they would be viewed as time-wasting or ungrateful, or placing further burden on the NHS.
When participants asked about their on-going pain, their concerns were delegitimized, by radiographic evidence, or health-care professionals' assertions that symptoms were normal, thus taking away their reason to seek help in the future.
Self-evaluative outcome expectancies describe how people self-evaluate and self-sanction according to their own personal standards and moral values. For example, individuals want to be seen as 'good citizens' and participants felt that nothing further can be done, and made a decision to self-manage, and not to overburden the NHS, this helps to maintain a positive self-identity.
Barriers to health-care use have been the focus of this study. The barriers described provide some explanation for why people do not proceed through the appraisal interval and on to help seeking. Scott et al 23 acknowledge that other events, beyond the stages that they already identify, may occur throughout the pathway and beyond.
However, they caution that identifying further events would not be useful as they would be too acute, ill-defined or varied to measure across different contexts. On the contrary, we found that after completion of treatment (knee surgery), patients may well return to the appraisal and self-management process, but then fail to move on to help-seeking because of a sense of futility and an overall belief that 'nothing more can be done'. We describe this as becoming stuck in a 'futility loop'.

| Strengths and limitations
A qualitative approach has enabled us to explore the complex experiences and views of people living with chronic post-surgical pain after knee replacement. Our broad purposive sample of thirty-four participants provided sufficient data to achieve data saturation, the point at which the collection of new data becomes unnecessary, 50 and we are confident that our sample was sufficiently diverse and that our use of an inductive approach has provided thematic areas of greatest relevance and salience to participants. Our work also identifies that some patients think there is a risk in further consultation which might lead to unwanted treatment and worse outcomes. We suggest that health care professionals ensure that patients have information about all local health care services and treatments to help them achieve better pain management, with information on how to access these. Patients who experience a moral conflict concerning further consultation for their replaced joint and over-burdening the NHS should also be helped to understand that asking for support for their pain is appropriate and may be helpful. There is a need for further research to identify other kinds of support and treatment that might prevent people with pain from becoming stuck in the futility loop. The identification of the futility loop has particular value for people living with long-term pain and who have undergone a number of, often major, interventions over time.

ACK N OWLED G EM ENTS
The authors would like to thank those who participated in the interviews. We also thank Dr Fiona MacKichan (FM) for early input into study design and analysis; members of the Patient and Public Involvement group for advice on study materials; and the study administration and management team: Kristina Lewis, Wendy Bertram and Gemma Munckenbeck.

CO N FLI C T O F I NTE R E S T
The authors declare no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.