Exploring how people with chronic pain understand their pain: a qualitative study

Participants

Participants were adults living in the UK with any type of chronic pain, recruited through local and national networks for people with chronic pain. A sample size of 12 participants was used, within the 6–15 participant range advocated for small-to-medium samples [36] and for studies scrutinising common perceptions or experiences [37].

Design

The GEM elicits immediate associations from participants, via a free association task (described below) [34], allowing relatively unconstrained description including inconsistent accounts. TA was used to locate explicit and latent patterns of meaning in the elaborations [38, 39], consistent with the framework of socially constructed understandings of pain shared by participants.

Core PNE concepts were presented in a short online article written by Moseley [40], the main author of PNE. Its brevity increased accessibility to participants, as intended [14], and using a written rather than video format allowed participants to digest the material at their preferred speed, re-reading if necessary. After reading the article, participants were asked to identify areas of agreement and disagreement between their own explanations of pain and explanations in the article. TA was chosen again as the method of analysis.

Procedure

All interviews were conducted through video-calls by the researcher (SK) between April and August 2020. Organisers of online patient groups, acquainted with the senior researcher (AW), were invited to share an information sheet with their pain networks. No incentive was offered for participation. Those who followed up the link on the information sheet were sent key details from the information sheet and a consent form for electronic signature. Upon returning the consent form and confirming an interview time, participants were emailed a numbered, 2 × 2, A4-sized grid (without instructions) and asked to print or draw it ready for the interview. An initial pilot interview provoked no substantive procedural adjustments but was excluded from TA.

In interview, participants were emailed GEM instructions: to write a word or phrase or to draw an image in each box that they associated with ‘how [they] understand chronic pain – that is, how chronic pain works, why it starts, and carries on.’ Participants were given time to fill in the grid, but encouraged to follow their ‘first thoughts’. Participants were then invited to elaborate on each grid response via open questions (‘Can you tell me about what you've put here?’). Additional open questions were asked to elicit further elaboration (‘Can you tell me more about that?’). Similar time spent elaborating on each grid box was encouraged but flexible (‘Would you like to add anything?’). This interview stage took about 40 min, allowing time for elaboration.

Participants were then invited to read the PNE article, newly emailed to them, and introduced as a ‘standard explanation of pain written by a famous pain neuroscientist.’ After reading, participants were asked: ‘What, if anything, does this article change, or add to, or take away from the explanations you have given so far?’; ‘What parts of this article made the most sense to you, if any?’; ‘What parts made less sense to you, if any?’ Participants were allowed to answer each question without interruption before the next question was asked. If answers were brief, further elaboration was encouraged through open questions. Participants were then informed that the interview was concluded; asked brief demographic questions (though not including pain types); thanked for their involvement and given the opportunity to express queries or concerns.

Data analysis

An initial content analysis of participants' grid responses was conducted as a preliminary process of categorising the free associations and identifying their relative salience. This informed construction of the main coding frame (below) and sensitised the researcher to the overarching themes of participants' subsequent elaborations.

Interviews were transcribed from recordings, reproducing speech content verbatim but without intonational or other paraverbal information [41].

A coding frame of themes and sub-themes for TA was developed inductively from the transcribed elaborations. Coding and analysis were completed using the software package Atlas.ti.8, enabling systematic organisation and presentation of interrelated sub-themes. A reliability test was applied, whereby a third-party researcher blind-coded one sixth of the interviews and estimated coding similarity using Atlas.ti [42]. Quantifying agreement is contentious within qualitative epistemology, but comparison was made to check for notably idiosyncratic coding.

Because participants largely expressed agreement with the PNE and made no substantial elaborations on their GEM-derived pain concepts, initially planned framework analysis was replaced by a further TA, given its versatility [35].

Reflexive statement

Unintentional researcher influence on interviews and TA is inevitable [43]. The researcher (SK) is a highly educated white male in his mid-twenties with personal trust in scientistic methods. At the outset of the study, his own conceptualisations of chronic pain, from friends' descriptions, featured neurological damage and skeletal disintegration; primary care doctors as invalidating and pain experts as impotent. He was unconvinced by the rhetoric of PNE proponents concerning its potency.

Interviews

Nineteen people, from three online pain networks, responded to the invitation. Three withdrew before participating: two because of technological problems, and one whose request to amend the interview transcript prior to analysis could not be granted. Four emailed after the 12 participant cut-off had been determined. Thus, 12 participants were interviewed – remotely, via video call, as per COVID-19 regulations. All participants completed all interview phases. Interview duration ranged from 47 to 70 min, with a median of 55 min. This variance was affected mainly by differences in time spent completing and sharing the grid; reading the article and debriefing after interview.

Participant characteristics are shown in Table 1. Chronic pains included diagnoses of fibromyalgia, complex regional pain syndrome and arthritis (mentioned in elaborations). Nine participants had attended prior pain management courses, and three had not.

Free associations

In a preliminary content analysis, participants’ free-associative grid responses (Figure 1) were sorted into 10 broad thematic categories. Although participants occasionally wrote multiple ideas in a single box, all boxes except one cohered into one theme.

Figure 1:

Grid responses from two participants.

The most salient themes were pain as draining (e.g. ‘tiring’, ‘heavy’, ‘overwhelming’); frustrating (‘irritating’, ‘unfair’, ‘depressing’) and needing management (‘physio’, ‘prayer’, ‘balancing’): each identified in six grid responses. The second-most salient themes were causes (‘childhood’, ‘trauma’, ‘weather’) and variability (‘erratic’, ‘up/down’, ‘boom/bust’), with five responses each, then neurological (‘rewiring’, ‘pain centre in brain’); somatic (‘muscles’, ‘joints;’); pervasive (‘constant’, ‘life-changing’) and restrictive (‘antisocial’, ‘stuck’) themes in four responses each. The smallest constellation was danger signal (‘red alert’, ‘red stop light’), in three responses. Table 2 displays this scope and prevalence.

Interview elaborations

In TA of participants' interview elaborations, three main themes emerged: communicating pain; explaining pain and living with pain. Numerous sub-themes comprised these themes, with the most salient delineated below.

Responses to PNE article

Responses to the PNE article predominantly conveyed meta-level comment on the article's style, interest and general resonance with participants' own interviews and prior knowledge. The TA of these responses produced six themes, represented below in selected quotations.

Pain conceptualisations

The primary research aim of this study was to understand how people with chronic pain make sense of their pain. TA of participants' GEM elaborations identified diverse pain conceptualisations with three main themes: explaining pain; communicating pain and living with pain. Sub-themes encompassed chronic pain experiences as both purely experiential and as more analysable phenomena. A tension was found between these two dimensions, encapsulated by a participant who prioritised impressionistic sensory terms (“pain is an experience, it is not a concept”) but also explained pain using extended conceptual models (e.g. neuropsychological metaphors). Participants made sense of pain using multiple, sometimes incompatible cognitions: pain was both experiential and conceptual; in the body and in the mind; diagnosable and inexplicable; manageable and insuperable. These representations reflected pain beliefs identified in prior qualitative research, encompassing damage-oriented models [27] and struggles with diagnostic understanding [28].

Social representations theory (SRT) [44] was used as an epistemological lens with which to understand these findings. SRT posits that people incorporate unfamiliar or challenging ideas and phenomena that disrupt their previous frame of knowledge by representing them via more familiar ideas [45], and reinterpreting in existing schemas [46]. Chronic pain is hard to understand or explain by reference to familiar acute pain, except through models that imply damage. Anchoring (classifying into known categories) and objectification (representing the abstract in concrete imagery) [47] enable the incorporation of difficult ideas and experiences. Participants anchored chronic pain in concepts of diagnoses, a nervous system that learned and adapted and ways of controlling or mitigating pain (many from medical consultations and pain management advice). Meanwhile, they objectified pain in sensory terms and images of tissue damage, drawn from lived experience.

The GEM accommodates incongruous pain conceptualisations in separate, equivalent grid boxes, allowing participants to leave them unresolved. These incongruities suggest that conceptualisations of chronic pain vary in relation to salient contexts and frames of symbolic reference.

Responses notably lacked reference to dominant tissue damage models of chronic pain, in favour of more experiential or PNE-oriented cognitions; this is discussed below. Although some descriptions were quite concrete in terms of anatomical problems, none repudiated the role of the brain as represented in PNE.

PNE engagement

The secondary research aim was to understand how participants engaged with PNE in relation to their prior conceptualisations; their responses to the brief PNE article suggested further cognitive inconsistencies. PNE-oriented themes, such as evocations of neuropsychological modulation, comprised a minority component of participants' grid elaborations. Yet, almost all participants endorsed the PNE as interesting and validating; and, notably, corroborative of their own explanations. Even when invited, participants were reluctant or unable to identify aspects of the PNE as less resonant with their experience. This agreement is striking, given the diverse conceptualisations earlier elicited. It might be that participants' other representations were decontextualised and displaced by the article's neuroscientific representations of pain: PNE fitted within their mixed conceptualisations, but may not have challenged those that were incompatible.

Some participants qualified their agreement, acknowledging that the article did not relate much to their particular pain; they found the neuroscience intellectually interesting and credible but doubted its utility (“What is it telling me to do differently?”). Participants also reported fatigue: whether with the interview process or with the difficulties of articulating their understandings was not clear.

The PNE thus presented acceptable ideas in a compelling style, but did not necessarily challenge some parts of participants' understanding, nor explain all their pain experiences. As described above, participants' discrete agreement with one pain model may not force change in their other representations in other contexts.

Relevance

The grid responses and elaborations resonated with previous qualitative research into patients' chronic pain experiences. They aligned particularly with the meta-ethnographic review by Toye et al. [28], which found common themes across qualitative studies of restricted, beleaguered living; grappling with the body; pursuing understanding through diagnosis and struggling with abandoning that pursuit. Responses also endorsed the sense of pain as an assault on the self [29]: pain as hostile in contrast to PNE's protective pain. Whether hostile or protective, these representations give pain its own voice, introducing a somewhat mystifying ‘homunculus’ into conceptions of volition [48]. These elaborations go further than prior research in highlighting the complexity of conceptualisations that pain education interventions seek to revise.

The lack of reference to tissue damage in the initial grid responses was unexpected, given the clinical focus on challenging patients' damage models as important [49]. Perhaps damage beliefs are becoming less dominant among patients, ceding to more central nervous system models [50] although of course it is not possible to generalise from this sample. More importantly, responses to PNE elicited here did not suggest that agreement automatically implies a reversal of habitual avoidance as implied [15].

The PNE responses are relevant to research and clinical fields. Neuroscientific models may be compatible with people's chronic pain schemata – particularly after pain education – but where they are incompatible, they do not necessarily provoke reconceptualisation. The article aims to remove fear and reverse habitual avoidance, but participants did not articulate this. Prior research into reconceptualisation after PNE [22, 30, 31, 51] has highlighted the importance of perceived personal relevance of PNE; this study showed little interaction between PNE and participants' wider pain models, and described fluid, contextually contingent conceptualisations of chronic pain [52], [53], [54], [55]. As belief in the practical helpfulness of a scientific concept is important in knowledge revision [56], responses may further clarify why some PNE trials have shown limited clinical benefits without additional physiotherapy [16, 19].

A very recent survey of pain education concepts most valued by patients after intervention [25] showed substantial integration of neuroscientific concepts by the majority who reported gains but, consistent with this study's findings, knowledge revision did not necessarily lead to behavioural change.

A recent meta-ethnography of 195 qualitative studies of chronic pain experience concluded that treatment may benefit from greater validation of patients' own conceptualisations [57]; the authors propose that interventions might usefully be more personalised and iterative. Further research might also explore barriers to reconceptualisation in those patients who do not benefit from PNE (excluded in [25]).

Strengths and limitations

A strength of this study is that it elicited chronic pain conceptualisations not directly informed by previous pain education. Although potentially informed by prior pain education, these conceptualisations are still owned by the participants, and tissue damage models can persist after PNE [27]. Further, to our knowledge, this was the first use of GEM in chronic pain research, eliciting incongruities in the conceptualisations that we addressed using SRT. However, the sample size, although adequate [22, 37], was sub-optimal for robust TA; a larger sample would have allowed comparison of data in relation to previous PNE exposure [58] and including diagnosis in the demographic questions might have enriched our interpretation. Further comparison between unprimed elaborations and PNE responses, if sufficiently detailed, might use framework analysis productively. Remote interviewing (mandated by COVID-19) improved accessibility [59], but lacked precedents for the GEM, and it was difficult to constrain participants to one image, word or phrase per box [34]. Further, while the researcher attempted to introduce the PNE through neutral tone and emphasis of its authorship, his researcher status may have inhibited criticism. Lastly, the brief written PNE to which participants were exposed cannot be equated with a clinical application of PNE, so clinical implications should be made with caution.

Implications

This study identifies how people with chronic pain make sense of their pain through diverse, inconsistent conceptualisations. It illustrates conceptual and experiential pain representations in themes of communicating pain, explaining pain and living with it. It also indicates how PNE is acceptable within these representations, but with a risk of being held as a discrete belief set detached from practical implications. As participants only acknowledged this apparent detachment when expressly questioned, it may go unarticulated during PNE interventions.

Further, since some PNE interventions have failed to show expected outcomes, this study suggests that PNE might strengthen its practical recommendations. There is little basis to expect that education alone, however intellectually appealing, will be sufficient to overcome the threat of pain and to change behaviours that are deeply wired into mammalian systems [25, 60]. Where pain has interfered across many areas of life, direct behavioural interventions and experiential disconfirmations – multiple and repeated across contexts – may be required for lasting behavioural change.