Understanding how children and young people with chronic non-cancer pain and their families experience living with pain, pain management and services: a metaethnography

Research aim

To conduct a meta-ethnography on the experiences and perceptions of children with chronic pain, and their families, of living with chronic pain, treatments and services to inform the design and delivery of health and social care services, interventions and future research.

Review questions

1. How do children with chronic pain and their families conceptualise chronic pain?

2. How do children with chronic pain and their families live with chronic pain?

3. What do children with chronic pain and their families think of how health and social care services respond to and manage their/their child’s chronic pain?

4. What do children with chronic pain and their families conceptualise as ‘good’ chronic pain management and what do they want to achieve from chronic pain management interventions and services?

Objectives

1. Conduct comprehensive searches to identify qualitative research literature on the experiences and perceptions of children with chronic pain and their families to address review questions 1–4.

2. Select relevant studies and synthesise them using meta-ethnography.

3. To ensure salience of findings via involvement of children with chronic pain and their families in study design, analysis and interpretation.

4. Assess how much confidence can be placed in our synthesised findings using GRADE-CERQual (Grading of Recommendations Assessment, Development and Evaluation-confidence in the evidence from reviews of qualitative research)53 in order to facilitate use of our findings for NHS decision-making.

5. Identify research gaps regarding review questions 1–4 in order to inform future research directions.

6. Integrate our findings with existing relevant Cochrane treatment effectiveness reviews22–28,46–50,54,55 in order to determine if programme theories and outcomes of interventions match children and their families’ views and preferences.

7. Inform the selection and design of patient-reported outcome measures for use in chronic pain studies and interventions and care provision to children and their families.

8. Disseminate findings to academic, clinical, lay and policy audiences to influence childhood chronic pain policy and practice.

Phase 1: Selecting meta-ethnography and getting started

Meta-ethnography is ideally suited to developing new understandings, insights and theory about the experiences of children with chronic pain and their families to inform service and intervention implementation and identification of outcomes of value to children and families. Preliminary searches indicated that no meta-ethnography existed in this area, and that a reasonable-sized and conceptually rich (in-depth explanatory/interpretive) evidence base existed, suitable for meta-ethnographic synthesis, to address our review questions. Meta-ethnography involves reinterpretation of the findings in primary studies, taking into account the contexts and meanings of the original studies;44 it therefore requires in-depth explanatory or ‘conceptually rich’ data with accompanying detailed or ‘thick’ data on the study context.

Topic of interest

Our topic focus was refined through discussion with health consumers (professionals and patients, see Chapter 3 for more details), examining relevant Cochrane intervention reviews and scoping searches of the available literature. We included studies that focused on the views and experiences of children with chronic non-cancer pain and their families towards chronic pain, health services and treatments. We excluded chronic cancer pain following common practice in chronic pain research, including Cochrane intervention effects reviews on children’s chronic pain, which focuses on cancer pain separately due to the very different care pathway and causes of the pain. 22–26 End-of-life chronic pain in the last weeks and months of life was excluded for similar reasons; children have a distinct care pathway accessing different kinds of services, for example hospices, and their pain management has different aims and considerations, for example addiction to analgesics is less of a concern. We defined a ‘child’ as a person under 18 years of age, according to the UN Convention of the Rights of a Child.

Inclusion criteria

• Peer-reviewed journal articles, published reports, book chapters, books, PhD theses.

• Contained qualitative research data on chronic pain, that is, pain lasting for 12 weeks or more, relevant to the review questions.

• Reported the views of children with chronic non-cancer pain from 3 months up to age 18 years or their family members (e.g. parents/guardians, grandparents, siblings).

• Qualitative primary research studies of any design (e.g. ethnography, phenomenology, case studies, grounded theory studies) including mixed-methods studies if it was possible to extract data that were collected and analysed using qualitative methods.

• Used recognisable qualitative methods of data collection (e.g. focus group discussions, individual interviews, observation, diaries, document analysis, open-ended survey questions) and analysis (e.g. thematic analysis, framework analysis, grounded theory).

• In any language.

• Any publication date.

Exclusion criteria

• Acute pain, that is, pain lasting for < 12 weeks, such as that caused by medical procedures.

• Cancer pain.

• Pain in neonates and babies < 3 months old.

• Focused on end-of-life pain management (in the last weeks and months of life).

• Non-empirical article, for example editorial, commentary, study protocol.

• Findings did not differentiate between children with acute and chronic pain.

• Findings did not differentiate between adult and child participants.

• Studies did not use qualitative methods for data collection and/or analysis (e.g. studies which analysed qualitative data quantitatively).

• Literature reviews.

The inclusion criteria were discussed and subsequently agreed with our PPI group (see Chapter 3).

Phase 2: Deciding what is relevant to the initial interest

A rigorous search for studies was conducted via bibliographic databases and supplementary searches, as outlined below. RT led the design and conduct of literature searches assisted by the research fellow (MSB) and Cochrane PaPaS. Initial literature searches of all information sources were conducted between August and September 2020. Bibliographic database searches were updated in September 2022 to bring the review up to date prior to publication. Studies were selected according to the methods outlined in Selection of studies. Our PPI group discussions informed which websites to search and which experts to approach for study suggestions (see Chapter 3).

Bibliographic database searches

We searched 12 bibliographic databases selected for their good coverage of qualitative research and spectrum of relevant disciplines (Table 1).

See Appendix 1 for the database search strategy for MEDLINE. The search strategy combined three key search concepts: (A) qualitative study designs; (B) population – children and their families; and (C) phenomenon of interest – chronic pain. The MEDLINE strategy was then adapted to the remaining bibliographic databases listed in Appendix 1. All databases were searched from their inception.

Supplementary searches

Supplementary searches12 for studies were conducted between 1 August 2020 and 17 December 2020 via:

• Searches of key websites, as outlined below.

• Hand-searching 24 months, from 1 December 2018 to 31 December 2020, of key journals relevant to our review questions and/or qualitative health research:

  • BMC Pediatrics

  • Clinical Journal of Pain

  • European Journal of Pain

  • Journal of Pediatric Psychology

  • Qualitative Health Research

  • Social Science and Medicine

  • Sociology of Health and Illness

• Contacting experts in the field for recommended studies, including ongoing research.

• Checking reference lists of included papers and relevant literature reviews for any further relevant studies.

Grey literature was identified by searching:

• Three bibliographic databases [Health Management Information Consortium (HMIC), OpenGrey and Electronic Theses Online Service (EThOS) – see Table 1].

• Websites of key organisations representing chronic pain health conditions including British Pain Society, Department of Health, NIHR Library, Sickle Cell Society, Versus Arthritis, CRPS UK, Fibromyalgia Action UK, Crohn’s and Colitis UK, Reflex Sympathetic Dystrophy Syndrome Association, European League Against Rheumatism network, European Pain Federation, Pain Relief Foundation, Children’s Health Scotland, children’s hospitals, Reflex Sympathetic Dystrophy Syndrome Association supporting the CRPS community, Children’s Health Scotland, The Royal National Hospital for Rheumatic Diseases, Royal National Orthopaedic Hospital, Great Ormond Street Hospital for Children, Sick Kids Hospital, NHS Lothian and Evalina Hospital in London.

Supplementary searches were not updated in 2022 due to resource limitations, with the exception of asking experts to suggest potentially relevant new studies. A Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram was used to record search results and the results of study screening for inclusion (Figure 2). Characteristics of the excluded studies can be seen in Report Supplementary Material 1.

FIGURE 2.

PRISMA flow diagram.

Literature screening and selection

Search results were exported to EndNote and duplicates removed. Then one reviewer (MSB) screened titles to remove off-topic records that were clearly not about children’s chronic pain, which was checked by a second, independent reviewer (RT). All remaining references were uploaded to Covidence systematic review management software. 63 First, titles and abstracts were screened, and clearly ineligible records were excluded. Then, the full text of all remaining studies was screened against our full inclusion criteria. At both stages, records were screened in duplicate by two independent reviewers (MSB, RT, EF, IU, LF, AJ, JN and/or KT). Disagreements were resolved through discussion and referred to a third reviewer, if necessary. Ten randomly chosen studies were used to pilot and standardise screening practices between reviewers.

Language translation

The review team are proficient in English, Portuguese, Spanish, Dutch and French. Any titles, abstracts and full texts published in any other language were initially translated through Google Translate to determine eligibility.

Sampling

Our final set of included studies was purposively sampled from all studies meeting our eligibility criteria. We used published guidance from The QIMG64–66 and input from our PPI group and Project Advisory Group (PAG) (see Chapter 3) to develop a strategy for sampling studies summarised below and in Box 1. In short, our sampling approach prioritised ‘conceptually rich’ and ‘thick’ UK studies with explanatory/interpretive findings and detailed contextual data, and then filled gaps in the data with less rich/thick UK studies and both rich/thick and less rich/thick non-UK studies. Our rationale for this is described below.

A meta-ethnography warrants studies with a good or fair amount of rich qualitative data and moderate to fair depth or ‘thickness’ of context and setting descriptions. 59 We assessed the conceptual richness of findings (i.e. whether the findings were explanatory/interpretive) and the thickness of contextual detail (how in-depth the context, intentions and meanings underlying the findings were) in UK studies selecting rich and thick studies for synthesis. This ensured that we were including studies that were the most adequate to address our review questions that were developed for a UK context. We adapted Ames, Glenton65 existing scale in assessing data richness, developed for qualitative evidence synthesis using thematic analysis, into a 3-point scale to suit meta-ethnography to focus on conceptual richness and contextual ‘thickness’, drawing on Popay et al. 67 and Cochrane QIMG guidance. 68 The scale and user guidance were drafted, piloted and revised by the research team. Table 2 presents our final richness scale (in which we used only the terms ‘thick’ and ‘thin’, whereas we would now also use the terms ‘rich’ and ‘poor’).

We first selected all the 3* UK publications for inclusion. We then explored any potential gaps in the data provided by these studies. We extracted basic information such as age and type of participants, pain condition, aims, setting and which of our review questions were addressed by their findings. These 3* UK publications only covered a limited range of pain conditions, and few provided information on review question 4, which indicated that the UK publications might not provide sufficient data to answer all our review questions.

Further sampling decisions were made in collaboration with our PPI group and PAG, to ensure that the synthesis addressed what is of greatest importance to children and their families and considered the views of other key stakeholders including healthcare professionals. The PPI members supported our preferred option not to include poor UK studies (rated 1* for richness) but felt we should include 2* as well as 3* UK studies. We therefore continued with a meta-ethnography rather than conduct a thematic synthesis, which could have incorporated less rich 1* findings. PAG members, including PPI representatives, agreed it was important to include non-UK studies (i.e. studies not conducted in the UK), as long as they fitted our aim and review questions, to try to represent a wider range of pain conditions and of participants. We completed richness assessments only for those non-UK studies which met our sampling criteria. Only those non-UK studies that were rated as 2* or 3* for richness (fairly or very rich data on our review questions) were included. We repeated the sampling process following the update of our literature searches in September 2022.

We referred to the PROGRESS-plus criteria (place of residence, race/ethnicity/culture/language, occupation, gender/sex, religion, education, socioeconomic status and social capital) when judging whether relevant populations were represented in the sample. 69 We used the four GRADE-CERQual domains (methodological limitations, adequacy, coherence, relevance) to guide the sampling process in order to develop the strongest findings for decision-making (see Assessing confidence in synthesised findings).

Phase 3: Reading the studies and data extraction

Studies were repeatedly read in full by at least two team members. We recorded study characteristics of all eligible studies (e.g. aim; country; number and type of participants, e.g. gender, age, diagnosis/type of chronic pain, patients, parents or other family members) in Microsoft Excel. Further characteristics were recorded for the final included studies such as study design, recruitment setting, sampling method, methods of data collection and analysis, ethnicity, conflicts of interest, funders. We referred to the PROGRESS-Plus criteria, described above, when extracting data on participant characteristics. 69

Full-text PDFs of the included studies were uploaded into NVivo qualitative data analysis software. 71 Each study was added as an individual ‘node’ and all of its conceptual findings were extracted as ‘subnodes’. (A ‘node’ is a container, given a specific label, into which data are gathered.) Each conceptual finding or theme was ‘extracted’ or coded, regardless of whether it appeared in the ‘findings’ or ‘discussion and conclusions’ sections, and labelled with the author, study identification number and the concept/theme name (see Report Supplementary Material 2). Authors’ recommendations and ‘interventions that worked’ were also extracted as subnodes for each paper. In total, we extracted data into 529 nodes. Some studies did not contain conceptual-level data but instead contained, for example descriptive themes or rich descriptions. For these studies, where possible, we interpreted the findings to develop new concepts and coded them at a new node. For studies identified in the updated searches, we compared the findings of included studies directly with our synthesis findings rather than extracting data in NVivo.

Phase 4: Determining how studies are related – grouping studies

In deciding our approach to grouping studies for synthesis, we considered grouping studies by type of participant (children, parents or siblings), or age of the children focused on (e.g. 0–5 years, 6–8 years), or type of pain/pain condition. We concluded that we could not classify studies into groups according to the type of participant or the children’s age group because many studies included the views of children and parents and/or focused on a wide age range. Grouping by the type of pain or pain condition seemed to be a more feasible and logical approach. Therefore, we then looked at existing systematic reviews of intervention effectiveness and qualitative evidence syntheses on children’s chronic pain to see how they had grouped studies by conditions and consulted our healthcare professional PAG and team members to agree the groupings (see Chapter 3).

We found only two Cochrane effectiveness reviews on children’s chronic pain which had grouped trials into headache versus non-headache chronic pain (all other pain conditions) for subgroup analyses; the decision was based only on the numbers of trials. 54,55 Separate Cochrane reviews had been conducted specifically for recurrent abdominal pain and for sickle cell disease pain in children. We planned to integrate our findings with existing Cochrane effectiveness reviews so separate groupings on sickle cell disease, headache and recurrent abdominal pain seemed appropriate. A WHO review 20 had categorised trials by pain condition, according to the ICD-11 classification. Most of our included studies provided a poor level of detail on pain conditions so we could not use ICD-11 to classify and group them. The only two qualitative evidence syntheses in the field at the time9,39 did not group studies by condition since one focused solely on juvenile idiopathic arthritis39 and the other included only eight studies. 9

After seeking advice from the healthcare professionals in the PAG (see Chapter 3) and team members, we grouped included studies by health condition, which resulted in a total of 11 groups:

• Juvenile idiopathic arthritis

• Abdominal pain

• CRPS

• Sickle cell disease

• Headache

• Neurological conditions

• Musculoskeletal conditions

• Skin conditions (epidermolysis bullosa)

• Dysmenorrhoea

• Mixed conditions

• Unspecified types of chronic pain

We consulted high-quality relevant systematic reviews and qualitative evidence syntheses to inform our subgroup analysis. 52,72–74

Phase 5: Translating the studies into one another

The conceptual data/findings from all NVivo nodes were interpreted in chronological order from the earliest publication date using NVivo memos. This process consisted of interpreting and capturing the key meaning of each conceptual finding while taking account of relevant contextual data. These interpretive memos were labelled with the condition, author’s name, study identification number and concept/theme name (see Report Supplementary Material 3). For the first three studies, each node was interpreted independently by two reviewers (EF and MSB). The two reviewers then discussed their interpretations and, if possible, agreed a joint interpretation recorded in one memo or recorded their alternative interpretations. For the remaining studies, one reviewer interpreted the data (EF or/and MSB) from each node and recorded this in a memo; a second reviewer (EF and MSB, IU, RT, LC and/or AJ) read the data coded at the node, then read the first reviewer’s interpretive memo and challenged, confirmed and/or added to the interpretation of the first reviewer. All authors kept an ‘analysis journal’ in a memo in NVivo to record any reflections, thoughts, issues or questions during the analysis.

Informed by our PPI group and based on previous meta-ethnographies,52 we initially analysed data in each grouping of studies separately before bringing them all together in phase 6, as we describe below. In total, 346 memos were created. In phase 5, we used the data in the memos to compare concepts systematically, study by study for each of the 11 condition groupings to identify both common and unique concepts. The translation and interpretation were completed using an inductive approach, guided by data and focused on meaning and context. At least three reviewers analysed (translated) the concepts to try to reach a new level of interpretation.

In the study-by-study translation for each grouping, all memos were downloaded from NVivo as .docx files and combined into one Word document. Then, we read the interpretive memos for each study in chronological order by publication date. We took the earliest publication first and read the memos for each concept, we compared the meaning of each with the memos for the second study looking for similar and contradictory concepts, we then compared these with the third study and so on, until we had identified the full range of concepts including common and unique concepts. To help juxtapose the concepts, we used tables in a MS Word document, one for each grouping, described below (Table 3). For groupings with only one publication, we looked for and identified any overarching concepts where possible. One reviewer (EF or MSB) carried out the initial translation for each grouping. A second and third reviewer (EF, MSB, IU, AJ, RT, LF, LC and/or JN) then read and challenged, confirmed and/or added to the translation. Common and unique concepts were identified within each grouping.

One PhD thesis77 was also published in two other included publications38,78 and contained similar data/findings, which were analysed together and considered as one publication. Britton (2002)79 and Britton (2002)80 reported different participants’ perspectives from the same study sample, so findings were also analysed as separate publications. Renedo (2020)81 and Renedo (2019)82 also used the same cohort for their studies and presented similar findings, and therefore were analysed as one publication.

Phase 6: Translation and synthesis within condition groupings

After the process of comparing concepts and looking for common and unique concepts in phase 5, we matched, merged and developed overarching third-order constructs. These third-order constructs expressed a new interpretation of the data which went beyond the findings of the original studies; we achieved this for some of the conditions for which there were sufficient in-depth data. Report Supplementary Material 4 and Appendix 2 show, for each pain condition grouping, how we progressed from included study findings to third-order constructs, illustrating the outcomes of our translation and synthesis of translations. We produced a textual synthesis in a narrative format for each of the concepts, and where possible, we also created diagrams showing how the new constructs linked to each other. We also referred to any theory or model produced by the original authors of the studies to check we had taken account of the original meanings.

All phase 5 translation and phase 6 synthesised translation information for each grouping was summarised in a table (example in Table 3) which included the author’s name and study identification number; a list of all the author’s themes or concepts labels (second-order constructs); where relevant, a list of our constructs (new second-order constructs) created for descriptive studies; a list of common and unique constructs across all studies (created from study-by-study translation); and a list of any overarching third-order constructs (our new interpretations).

Britten et al. 83 coined the term ‘third-order constructs’, thus further developing Schutz’s84 notion of ‘first-order constructs’ – lay understandings – and ‘second-order constructs’ – the researchers’ interpretations of the first-order constructs. The translation and synthesis process resulted in a total of 39 third-order constructs and 169 second-order constructs across all the groupings.

Translation and synthesis across condition groupings

To translate and synthesise data across condition groupings, we conducted a series of four research team meetings including one hybrid in-person/online and three synchronous online meetings. Prior to meeting, data from all groupings were organised into one document, which included all the narratives (the textual syntheses) created from the translation process, and data that were used to create new second- and third-order constructs. Labels (like Post-it^® notes) for each construct including the condition, the name of the construct, the contributing studies and a brief description of the construct were used during the meeting. Labels were colour-coded according to the condition and whether they were second- or third-order constructs. Six members of the team then organised the labels through a thematic analysis into distinct themes and then into broader ‘analytic categories’ (thematic headings) according to shared meaning. Four members attended in person (EF, IU, LF and MSB) and two members online (RT and AJ). Members participating online were able to follow the analysis process using a Padlet85 (a real-time collaborative web platform used to share and organise content – see Report Supplementary Material 5).

Following the hybrid analysis and synthesis meeting, the narrative document (containing all themes/constructs across all groups) was reorganised according to the new ‘analytic categories’. A further three online team meetings using Google Jamboard86 (a digital interactive whiteboard) focused on analysing the findings using all the different perspectives and expertise from the whole research team (see Report Supplementary Material 6).

Once all data were organised into analytic categories (five categories were created), we finished updating the textual synthesis, and used Microsoft Teams87 whiteboard to create diagrams to express and understand how themes/constructs and analytic categories were related (see Report Supplementary Material 7). We used the diagrams created in Microsoft Teams whiteboard to discuss the findings/analysis virtually with the research team; subsequently, we produced short descriptions for each analytic category based on the diagrams. These descriptions were developed into detailed textual syntheses.

We continually further developed and refined our analysis and synthesis. In order to clarify our terminology, where we had been able to reach the level of new interpretation, the analytic categories were renamed as ‘third-order constructs’, for example ‘Pain organises the family system and the social realm’, most comprise ‘second-order constructs’, for example ‘adapted parenting’, which we had previously called ‘themes’. We developed lines of argument – the ‘overarching storylines’ – to explain how all the final constructs linked together with the help of diagrams in Whiteboard and team discussion. These lines of argument together enabled us to develop our theory of children’s chronic pain. We discussed and clarified ambiguous or unclear findings with our PPI and PAG groups (see Chapter 3).

Integration mechanisms for quantitative results and qualitative findings

There were various points in overall meta-ethnography production at which integration occurred. 92,95 We have integrated quantitative and qualitative perspectives during review question formulation and synthesis.

Patient and public involvement group

Patient and public involvement recruitment was conducted in three main stages. These comprised:

1. During development of the meta-ethnography research grant proposal in 2019, we involved 10 lay people including 3 CYP with chronic pain (1 was also a patient representative for a third-sector organisation), 4 parents, 2 adult patient representatives from the third sector and 2 adult members – 1 with chronic pain and 1 with a chronic illness from a university PPI Research Partnership Group (designed to assist with developing research relevant and useful to patients, carers, family members and healthcare professionals).

2. In 2020, we recruited a core PPI group of children aged 8–18 years, and parents or/and informal carers (i.e. not healthcare professionals) of children with chronic pain aged 3 months to 18 years. Additional prospective recruitment during the meta-ethnography was carried out as needed (e.g. to fill the gaps in terms of experiences with specific contexts/conditions).

3. In 2019–20, healthcare professionals, representatives of third-sector organisations and academics with relevant experience were recruited to join our PAG.

We advertised for PPI members via national pain services, social media (Facebook and Twitter) and third-sector organisations (charities) including the Sickle Cell Society, Fibromyalgia Action UK, Great Ormond Street Hospital, Pain UK, The Brain charity, CCAA kids with Arthritis, Pain Relief Foundation, Action on Pain, Coeliac UK, Guts Charity, Dystonia UK, Endometriosis UK, Sick Children Trust, Action for ME, A way with Pain, MS Trust, Diabetes UK, Fibro Awareness UK, Pain Concern and Independent Nurse. We tried to recruit a diverse core PPI group with any type of chronic pain, except for pain associated with cancer – whether primary (e.g. fibromyalgia) or secondary pain conditions (e.g. arthritis) – with a variety of experiences, ages, ethnic backgrounds and socioeconomic statuses. Other stakeholders were approached directly via e-mail.

Patient and public involvement group

We successfully recruited 12 children (10 females, 1 male and 1 non-binary from 8 to 20 years old) and 8 parents (all female). The young people had chronic pain conditions including Ehlers–Danlos, fibromyalgia, migraines, general chronic pain and chronic lumbar paravertebral muscle spasm, chronic headaches and CRPS. Parents and/or informal carers in the group were all mothers of children with pain from cystic fibrosis, CRPS or Ehlers–Danlos syndrome. All members lived in the UK, specifically in England, Scotland and Wales and were white. All PPI members were invited to join the PAG group.

Project Advisory Group

The PAG, which also facilitated some PPI involvement, comprised 27 members including 10 children with chronic pain and 7 mothers of children with chronic pain (also part of the PPI group), 6 healthcare professionals and 7 other stakeholders. The healthcare professionals/academic healthcare professionals included a consultant in paediatric anaesthesia and pain medicine who was a senior clinician and research lead at a specialist pain service; a clinical academic and consultant in pain medicine who was national lead clinician for chronic pain for the Scottish Government and vice chair of the National Advisory Committee on Chronic Pain; a clinical academic who was Chair of Pain Medicine, Honorary Consultant in Anaesthesia and Pain Medicine, and chair of the 2018 SIGN guideline development group for children’s chronic pain; one general practitioner and two physiotherapists. There were two patient representatives from the third sector including from Pain Concern and Children’s Health Scotland, a representative from Healthcare Improvement Scotland (Scottish Government), and an academic expert in qualitative evidence synthesis. We were unsuccessful in recruiting representatives of NICE or the UK government.

Methods

Initial PPI engagement during grant development involved sending PPI volunteers materials via e-mail about the topic and our rationale for the focus of the review and draft review questions.

Impact/outcome

Respondents confirmed the importance of the topic and review questions.

Methods

We finalised aspects of the literature search strategy in collaboration with our core PPI group in September 2020. Specifically, we e-mailed them materials explaining the search terms, sources in plain and simple language and asked them to suggest sources, key studies or experts (see Report Supplementary Material 8).

Impact/outcome

The PPI group suggested seven additional sources, which were included in the search strategy (see Chapter 2, section Supplementary searches). They agreed our final study inclusion criteria.

Methods

Our PPI and PAG groups informed sampling of studies. Eight PPI members (four children with chronic pain and four parents/guardians) contributed to decisions – six took part in a 1-hour online workshop on 1 April 2021 and a further two members took part via an online survey. The team presented gaps identified in the UK data and asked five questions (see Appendix 3) related to the sampling of studies in terms of their conceptual richness, study setting and pain conditions. We also asked if we should continue with a meta-ethnography or instead use an alternative synthesis methodology (e.g. thematic synthesis) which would allow inclusion of a wider range of studies, not just conceptually rich ones.

We then asked our PAG how we should sample non-UK studies. Twenty-three people attended an online workshop on 6 May 2021, including six research team members, four parents and four CYP with chronic pain, and nine healthcare professionals, academics or other professionals. We presented the gaps in populations and pain conditions covered by the conceptually rich UK studies.

Impact/outcome

The PPI members indicated that we should not include studies rated 1* for richness, but we should include richer 2* as well as 3* studies; and therefore, we should continue with a meta-ethnography rather than conduct a thematic synthesis. The PPI group also agreed that some studies conducted outside the UK should be included. Subsequently, most PAG members agreed we should include non-UK studies regardless of the country in which the study was conducted, studies for a wider range of pain conditions, and a broad range of participants, as long as studies fitted well with our meta-ethnography aim and review questions.

Methods

We e-mailed PPI and PAG members in June 2021 to ask how we should group studies for preliminary analysis by subgroup, prior to synthesising the whole sample of studies and received 16 responses. Subsequently, we e-mailed 11 clinicians and pain experts to ask about their views and suggestions on the preliminary groups of health conditions (see Report Supplementary Material 9 and Chapter 2, Phase 4: Determining how studies are related – grouping studies).

Impact/outcome

Patient and public involvement and PAG responses contributed to the decision to group the included studies into a total of 11 groups of health conditions: juvenile idiopathic arthritis, abdominal pain, CRPS, sickle cell disease, headache, neurological conditions, musculoskeletal conditions, skin conditions (epidermolysis bullosa), dysmenorrhoea, mixed conditions and unspecified type of chronic pain.

Methods

A 2-hour PPI workshop with two parents and three young people was held in December 2021 to discuss, clarify and interpret preliminary findings from primary studies. We used the software StoryboardThat (www.storyboardthat.com) to create cartoons and accompanying scenarios (e.g. Figures 3 and 4) to convey these findings to prompt discussion. Cartoons included people of different ethnicities and genders. We formulated the dialogue in speech bubbles using the language and terminology from children’s and parents’ quotations in the included studies and ensured it was suitable for a minimum reading age of 8–9 years of age. Members of the team approved all dialogues prior to the workshop. All PPI members received the cartoons and an infographic (see Report Supplementary Material 10) explaining the study findings a week prior the workshop.

FIGURE 3.

Example of cartoon used during PPI analysis workshop.

FIGURE 4.

Example of cartoon used during PAG analysis workshop.

Impact/outcome

Patient and public involvement provided a different perspective on and/or interpretation of some of the findings discussed. For instance, PPI feedback added nuance and clarification to the idea of ‘control’ found in primary studies and helped the team to reach the following perspectives:

1. The idea of ‘control’ had been used in included studies often without clear definitions and without adequate consideration of what it means to parents and CYP with chronic pain.

2. The perspective of control might vary, for example for a parent it could mean something quite different than for a young person.

3. Overall, parents and CYP perceived the term ‘control’ in relation to chronic pain negatively, for example it implied that CYP could control their pain and its impacts.

New data and insights were incorporated into our analysis and synthesis and helped us to further develop and refine the following second-order constructs:

1. – Pain organises the family system (see Third-order construct 1: pain organises the family system and the social realm).

2. – Pain’s adverse psychosocial impacts on the whole family (see Third-order construct 1: pain organises the family system and the social realm).

3. – Pain forces adjustment and adaptation (see Third-order construct 1: pain organises the family system and the social realm).

4. – Families’ striving for diagnosis and a cure (see Third-order construct 2: families struggling to navigate health services).

5. – Importance of being listened to and believed by healthcare professionals (see Third-order construct 2: families struggling to navigate health services).

The core PPI group also helped fill gaps in the data, for example around disengagement with NHS health services, and brought experiences of pain conditions not represented in the included studies, such as Ehlers–Danlos syndrome and cystic fibrosis.

Methods

In a 2-hour online PAG meeting in April 2022, attended by four parents and two young people, we discussed four further findings from included studies. No healthcare professionals or other stakeholders attended this meeting despite being invited. We also used cartoons to facilitate the discussion (e.g. shown in Figures 3 and 4). All members received the cartoons and an infographic (see Report Supplementary Material 9) explaining the study findings in advance of the workshop.

Impact/outcome

This workshop helped us to interpret some of the findings. For instance, we discussed the stigma associated with psychological treatment, as illustrated in the cartoon (Figure 4). PPI feedback indicated some CYP had had traumatic experiences related to treatments and interactions with healthcare professionals, which helped the team to reach the following perspectives:

1. The stigma associated with psychological treatments may be related to negative experiences in health services.

2. Healthcare professionals did not explain to parents and CYP why they were being offered psychological treatments.

3. Psychological treatments were used as an alternative when nothing else worked rather than as part of a broader biopsychosocial approach.

This workshop helped us to further develop and refine the following second-order constructs:

1. – Importance of being listened to and believed by healthcare professionals (see Third-order construct 2: families struggling to navigate health services).

2. – Chasm between health services and families’ needs (see Third-order construct 2: families struggling to navigate health services).

3. – Experiences and perceptions of interventions and treatments (see Third-order construct 4: families’ experiences and perceptions of specific interventions and treatments).

4. – Discrimination and racism in health services (see Third-order construct 5: children and families experiencing prejudice and discrimination).

Meta-ethnography short title, acronym and logo

In August–October 2020, members of the PPI group and other CYP with chronic pain and their families were invited to help us choose a short title, acronym and logo for the meta-ethnography. People voted via social media (e.g. Facebook and Twitter).

Impact/outcome

The acronym and short name selected was CHAMPION (Children And faMilies’ exPerIences Of chronic paiN). A logo was chosen and used to create a brand for the CHAMPION meta-ethnography.

Conference oral presentation

In September 2022, Georgina Ferguson-Glover (a 24-year-old PPI member, with hypermobile Ehlers–Danlos syndrome) and two team members (EF and MSB) co-presented at the Royal College of Nursing International Nursing Research Conference in Cardiff, UK. Georgina decided the extent of her involvement and the content of her presentation. She chose to present her experiences participating in the meta-ethnography, her experiences living with chronic pain and how they related to the meta-ethnography findings. The presentation slides and script were produced collaboratively during multiple online meetings.

Impact/outcome

Georgina’s participation brought the presentation to life with examples of her lived experiences and her experiences of the PPI group. The audience, composed mainly of nurses and healthcare professionals, was engaged throughout the presentation and they asked many questions at the end. One of the questions related to the assessment of chronic versus acute pain and this highlighted the need for a chronic pain-specific assessment.

Animation of findings for children and young people

In 2019, CYP with chronic pain had indicated they would prefer an audio-visual output accessed via social media (YouTube) to learn about the study findings. Parents also had stated that they would watch such an animation. In October 2022, to create the short animation of findings tailored for CYP, we first conducted an online survey via social media (e.g. Facebook and Twitter) with our PPI members and the wider public with chronic pain aged between 5 and 25 years old to ascertain their preferred cartoon style for the animation. Twenty CYP, from 9 to 25 years, from three different countries (UK, Ireland and Germany) rated styles on a 5-point scale to indicate their preferences and made free-text comments. In 2023, our core PPI group chose which meta-ethnography findings should be incorporated into a script for the animation. They then reviewed drafts of the script, the storyboards and the final animation.

Impact/outcome

Patient and public involvement ensured that the animation should be engaging, appealing and relevant to CYP and families dealing with chronic pain. The animation, available here www.youtube.com/watch?v=Oz5I7xKYf5o, will be shared widely across social media platforms (e.g. YouTube, YouTube Kids).

Other

Five members of our core PPI group gave feedback on this chapter and on the lay summary. All indicated that the chapter adequately described their involvement and that the lay summary was appropriate for a lay audience. PPI members continued to be involved in collaboration on a journal article regarding PPI involvement in a meta-ethnography and future dissemination strategies via social media, beyond the official project end date.

Evidence gaps in eligible studies

We identified gaps in the evidence from eligible studies with respect to the age range, types of pain and types of participants (e.g. children, parents, siblings) covered. For example we found only 2 (1%) studies focusing on infants under 2 years old103,104 and only 11 (6%) focusing on children aged from 2 to 5 years old. 105–116 Only eight (5%) studies included the views of siblings.

Third-order construct 1: pain organises the family system and the social realm

Data supporting this third-order construct originate from 36 publications representing 33 studies published between 2000 and 2022. 34–36,38,75–77,79–82,102,105,117–126,128,131,132,135,136,138,141–143,145–149 Twenty-one of the 33 studies were conducted in the UK. These data come from studies focusing on the following conditions: sickle cell disease, CRPS, abdominal pain, epidermolysis bullosa, juvenile idiopathic arthritis, musculoskeletal conditions/pain, neurological conditions or a mix of pain conditions.

This third-order construct is composed of seven second-order constructs: pain organises the family system; adapted parenting; pain’s adverse psychosocial impacts on the whole family; pain forces families to manage uncertainty; pain forces adjustment and adaptation; pain’s influence on disclosure, social support and social relationships and school and the social realm.

Third-order construct 2: families struggling to navigate health services

Thirty-six included publications from 32 studies published from 2000 to 2022 contributed data to this third-order construct. 34,36,75–77,79,80,82,102,105,109,110,117–120,124,125,132,134–147,149 Pain conditions in these publications included: unspecified chronic pain, headache, CRPS, mixed pain conditions, neurological conditions, musculoskeletal pain, abdominal pain, juvenile idiopathic arthritis and sickle cell disease. Studies were conducted in the UK,34,76,77,79,80,82,102,109,110,117–120,124,125,132 the USA,140,143,144,147 Canada,36,105,138,142 Norway,75,137,146 Sweden,139 Australia,135,141 Cameroon,149 Brazil,145 Ireland136 and Hong Kong. 134 There was a lack of data on experiences of social care services.

This third-order construct describes the difficulties families faced when navigating and interacting with health services including getting stuck in a cycle of continually returning to services in pursuit of a diagnosis and curative treatment. Service fragmentation and numerous referrals without a clear care pathway resulted in frustration, unmet expectations, inadequate pain management and unmet needs. In contrast, a healthcare professional listening to and believing parents and CYP experiences of pain could reduce their distress and isolation and provide hope of resolution. This third-order construct is composed of four second-order constructs: families’ striving for diagnosis and a cure; family expectations of services influence experiences; chasm between health services and families’ needs; and the importance of being listened to and believed by healthcare professionals.

Third-order construct 3: families managing pain independently

Thirteen studies contributed to this construct75,76,82,105,109,110,119,123–125,134,135,144 which covered six conditions: sickle cell disease, headache, epidermolysis bullosa, abdominal pain, dysmenorrhoea and neurological conditions. There are gaps in our understanding of families’ management of children’s pain because there were few relevant data in the studies and much of the data were not very rich. This third-order construct is composed of data from two second-order constructs which are focused on how families manage pain independently: pain self-management and parents helping CYP to manage pain.

Third-order construct 4: families’ experiences and perceptions of specific interventions and treatments

Twelve studies contributed to this third-order construct77,79,105,117,118,133,135,137,139,140,143,146 covering the following conditions: juvenile idiopathic arthritis; musculoskeletal pain; CRPS; cerebral palsy (dyskinesia); epidermolysis bullosa; mixed pain conditions including rarer conditions such as osteogenesis imperfecta; unspecified chronic pain or chronic pain from any cause. This third-order construct is composed of two second-order constructs: experiences and perceptions of interventions and treatments, and adherence to and engagement with interventions and treatments.

Relevant data were limited, largely descriptive and heterogeneous, which adversely affected our ability to fully develop a coherent conceptual understanding. Overall, there was a lack of rich data in studies about family attitudes to and experiences of pain treatments. Studies focused on ACT, a cognitive–behavioural therapy;139 physiotherapy for musculoskeletal pain;137 daily physiotherapy, splinting and medication for juvenile idiopathic arthritis;79 pharmacological and nonpharmacological juvenile idiopathic arthritis treatments;105 interdisciplinary intensive pain treatment at specialist pain clinics/programmes for CYP with CRPS or a range of painful conditions;77,117,118,133,140,143,146 and complementary treatments, such as chiropractors, massage and mindfulness, used to help reduce pain associated with dyskinesia in cerebral palsy. 135

Third-order construct 5: children and families experiencing prejudice and discrimination

This is a cross-cutting third-order construct which applies across the other third-order constructs, so we present it separately here and refer to it throughout the results section, as appropriate. Seventeen studies (reported in 19 publications) focusing on sickle cell disease, epidermolysis bullosa, juvenile idiopathic arthritis dysmenorrhoea and abdominal pain contributed to this construct. 35,75–77,79,80,82,123–126,130–132,134,136,145,148,149

Children and families experiencing prejudice and discrimination was most apparent for CYP with sickle cell disease – in terms of experiencing racism in particular – and for CYP with visible differences due to their condition or with a learning disability. Visible differences were apparent in children with juvenile idiopathic arthritis, whose mobility was affected,126 and epidermolysis bullosa, who often had visible severe blisters. 123 One study focused on children with Down syndrome and juvenile idiopathic arthritis. 136 CYP with sickle cell disease, mostly from African and Caribbean ethnic groups, reported not receiving timely and adequate pain relief which they explained with reference to institutionally racist practices within the health service. 82 However, prejudice and discrimination also applied to other less ‘visible’ chronic pain conditions, such as sexist attitudes of health professionals to dysmenorrhoea. 134

Line of argument 1: dominance of biomedical perspective in how children and families conceptualised chronic pain and pain management

This line of argument pulls together data from the third-order and second-order constructs to show there were differences between how CYP and their families conceptualised (1) the impacts and experience of living with chronic pain, compared with (2) how they understood the cause of pain and (3) how they understood the role of health services in management of the pain. In terms of living with chronic pain, CYP and their families described experiencing mainly negative physical, social and emotional impacts on their lives, such as the impact of pain on family life, social life, schooling and parental employment (as described in Third-order construct 1: pain organises the family system and the social realm). Thus, it can be seen that CYP and their families expressed physical/biological, psychological and social impacts of chronic pain on their lives, which is consistent with a biopsychosocial model of pain.

Despite families recognising biopsychosocial impacts of pain, CYP and parents tended to understand the cause of chronic pain, at least initially, as solely biomedical, that is, pain has an underlying pathophysiological cause which can be identified, treated and/or cured. Families were not usually readily receptive to the idea of any psychosocial causes of, or influences on, chronic pain because they believed it implied that the pain was not ‘real’ or that parents were to blame for their child’s pain, for example because of a dysfunctional family life or poor parenting. Consequently, many families did not want or understand the point of psychological pain treatment and saw it as stigmatising and a last resort. An exception was some mothers of children with abdominal pain were open to the existence of psychosocial causes/influences, although it is not clear why (described in Third-order construct 2: families struggling to navigate health services). It was mainly families for whom there was no identifiable or known medical cause of pain (which does not necessarily mean there was no biological cause) who were asked to consider possible psychosocial causes of pain. For children whose pain was caused by an underlying health condition, such as sickle cell disease, juvenile idiopathic arthritis or cerebral palsy, families were less likely to consider psychosocial causes of, or influences on, pain; these families may not have seen the value of psychosocial treatments either, given the cause of pain was a disease.

How CYP and families initially conceptualised the role of health services in managing the chronic pain was based on a biomedical understanding which shaped their expectations of, and interactions with, services. Specifically, CYP and families sought to find an underlying biomedical cause in order to access curative or palliative treatment (as described in Third-order construct 2: families struggling to navigate health services). Families who already knew the cause of pain was a disease also sought biomedical pain relief and in some cases a cure for the underlying disease from services. Some families did, however, want support with psychosocial impacts of pain on their lives but did not expect or recognise that health services might provide such support. In interactions with health services, this desire for biomedical treatment for chronic pain created a medical focus for families, and for healthcare professionals. However, their biomedical conceptualisation of the causes and appropriate treatment of pain was not static and could change over time. For example families could realise over time that repeated clinical tests could not identify a physical cause and/or medical treatments did not resolve the pain, or they were told by healthcare professionals that no cause could be found, and/or there was no treatment or cure. (Even when an underlying physical cause had been identified, treatment could be ineffective or there could be no suitable treatment.)

Line of argument 2: parents fundamental to effective pain management

This line of argument resulted from interpreting third-order and second-order constructs (described in ‘pain organises the family system’, ‘adapted parenting’, ‘parents helping CYP to manage pain’, ‘experiences and perceptions of interventions and treatments’, ‘adherence to and engagement with interventions and treatments’, ‘pain’s adverse psychosocial impacts on the whole family’ and ‘pain self-management’ in Synthesis findings) and it explains the fundamental role of parents in implementing pain management strategies at home and the impact of these on family life.

Because of the profound impacts of pain on the child and the whole family life, parents had to adapt their parenting approach to help their child to manage pain. Families carried out pain management at home on their own most of the time which interrupted the family structures and routines. Indeed, most prescribed pain management strategies were time-consuming and disruptive of family life; some were also repetitive and painful or had other unpleasant side effects. Parents were fundamental in deciding whether and how CYP followed any prescribed pain management and treatment strategies. This decision was usually made based on the amount of impact the treatment strategy had on family life and on the child. Parents, especially parents of children who could not communicate, developed expert knowledge in assessing when their child was in chronic pain. Parents were typically the main advocate for their child in interactions with healthcare professionals and services in order to access effective pain management. Parents often had to make difficult decisions about their child’s pain management (particularly for complex conditions such as neurological conditions and juvenile idiopathic arthritis) but they were left feeling on their own to do so. Feeling that they could share this responsibility with healthcare professionals could help parents feel less alone.

Line of argument 3: the journey of living with chronic pain

This line of argument was developed with data from third-order and second-order constructs (described in the second-order constructs ‘pain organises the family system and the social realm’, ‘pain’s adverse psychosocial impacts on the whole family’, ‘pain forces adjustment and adaptation’, ‘families striving for diagnosis and a cure’, ‘family expectations of services influence experiences’, ‘chasm between health services and families’ needs’, ‘importance of being listened to and believed by healthcare professionals’, ‘pain self-management’ in Synthesis findings). It explores families’ journeys from the onset of chronic pain, their struggle to navigate services seeking for a cure, to when their needs and expectations were unmet and they then moved on either to prioritise living well with pain or they gave up hope. We hypothesise that a biopsychosocial perspective of pain management could help improve how families experience pain management and services.

Poorly managed, moderate or severe pain (which was not resolved by treatment or not treated at all) deeply restricted and shaped family life and relationships. At the onset of chronic pain, families tried to hold on to a sense of normality and sought help from services with the aim of curing pain. Families believed a diagnosis could help to determine the treatment they should take to get rid of the pain. However, receiving help from health services was not straightforward and a cure was not always possible. Services were fragmented and difficult to navigate, and families found themselves being referred from one clinical speciality to another without a clear care pathway to indicate the steps in the process or what the result would be.

In addition, communication with healthcare professionals was difficult. Families did not feel their experiences of pain were listened to and believed, and therefore, a diagnosis was difficult to achieve. In the end, families’ expectations of services (particularly services not specialising in pain) were not met. Pain was not cured and the disruption on family life not acknowledged by healthcare professionals. Families were left to deal with pain on their own. They dealt with pain on their own by employing self-management strategies and seeking complementary/alternative medicine, which sometimes increased the negative impact of pain and its management on the whole family. Some families moved on from searching for a cure as they realised they had no option but to ‘get on with it’, and so some started to prioritise living well with pain rather than ‘curing’ pain; others gave up hope and became depressed. As part of the process of moving on, families learnt as best they could how to manage the pain and how to plan their routine around it. We hypothesise that adopting a biopsychosocial perspective of chronic pain management and shifting the focus solely from curing pain to living well with pain could reduce the impact on family life and improve how families experience pain management and experiences of navigating services. To facilitate this biopsychosocial perspective in families early on in their journey, it is important for healthcare professionals to set expectations regarding the unlikely probability of a cure, if appropriate, to help families in this process of adaptation and help them navigate their chronic pain journey.

Linking the lines of argument into a model: families navigating chronic pain management

The interpretation of all data and lines of argument helped us to understand families’ journeys dealing with chronic pain and how these were affected by health services. We created a model of the non-linear nature of families’ journeys with chronic pain while they navigated services seeking a resolution for the pain which links all three lines of argument, illustrated in Figure 6.

FIGURE 6.

Families navigating chronic pain management (connects all lines of argument). Key: Reduces impact on family life; possible direction (i.e. families may follow this path); direction (i.e. usual path families took).

The diagram describes families’ journeys dealing with chronic pain and how health services can affect their journey (Figure 6). We hypothesise that there are two main paths (A and B) families can follow during their journeys. Families navigating Path A can move on to Path B with time or help from healthcare professionals and interventions. Families can also move from Path B to A, for example when pain worsens or fluctuates in severity.

On the left side of the diagram on Path A, families experiencing the impact of poorly managed, moderate or severe pain seek services in hope of a cure. When their expectations are not met due to fragmented services, referral fatigue, lack of diagnosis, poor communication with healthcare professionals and so on, they try to adapt to pain on their own. To adapt, they use many strategies, such as adapted parenting, adapting and changing the family routine according to pain levels, and so on, which can have a negative impact on their lives as the whole family focus is shifted towards dealing with the pain. Families also employ different strategies to manage pain on their own and seek help from alternative medicines which are not always successful. Because pain remains poorly managed, families are stuck in a cycle of pain affecting their lives and seeking services in the hope for a cure but without any resolution for the pain. Families and healthcare professionals understand pain and its causes through a biomedical lens/perspective which deeply affects their expectations of pain management and services. Psychological treatments are stigmatised and only used as a last option.

The right side of the diagram depicts when healthcare professionals are effective in helping set families’ expectations of treatment, which requires them to listen to and believe the families and acknowledge their lived experiences. Realistic expectations can help families to adapt to living with pain as they shift their focus from trying to cure pain to trying to live well despite the pain. Families might also realise on their own that a biomedical cure is unlikely after following Path A for a long time without any resolution for the pain. Families on Path B tend to be more receptive to biopsychosocial approaches for pain management, due to lack of other options.

Family-centred theory of children and young people’s chronic pain and its management

Based on interpretation of all evidence and PPI input, we developed a theory of what would constitute ‘good’ children’s chronic pain management interventions and services. The theory (Figure 7) can be used by, for example intervention developers, practitioners, service commissioners and researchers.

FIGURE 7.

Family-centred theory of CYP’s chronic pain and its management. Key: Resources and training needs/interventions, Team’s hypothesis based on evidence, Gaps in qualitative evidence, Based on evidence and corroborated by PPI.

Our family-centred theory of CYP’s chronic pain and its management is based on a biopsychosocial model and a ‘whole systems’ approach, that is, based on the premise that everything is inter-related and interdependent (Figure 7). Because chronic pain affects all aspects of family life and their relationships with the community, pain management strategies encompass the family (depicted in green background shading), healthcare services (depicted in yellow shading) and the community including schools (depicted in purple shading). The systems (e.g. community, school, health and social care services) work together as enablers, either facilitating or delivering families’ desired outcomes. Four key outcomes desired by families and CYP (blue shading) are at the centre of this approach to chronic pain management and services. Families’ most desired outcome is resolution of/a cure for their pain. This systems approach resolves service fragmentation by offering intuitive services and proactive ongoing support for families across all environments (family, healthcare services and community/schools) as they communicate and work together to achieve satisfactory outcomes for families.

Given that parents and CYP had to constantly assess pain levels in order to plan family life and pain management, more support to facilitate pain assessment at home is put in place via links with health services and resources/training. Families and CYP carried out most of the pain management and prevention. Parents receive resources and support on how they can help their child deal with pain and how to prioritise different pain management options at home. CYP receive support on different types of pain management strategies tailored to their needs and pain levels and how to differentiate between acute and chronic pain. Psychosocial treatment approaches are offered earlier, alongside medical treatments, as a way to address the physical, emotional and social impacts of pain holistically, and these are carefully explained to families. For instance, psychosocial approaches which improve communication among family members and help CYP to build autonomy over how and when their pain is managed can enable pain management and are facilitated by health services and the community/schools. Further resources and training with information about pain and treatments tailored for families, which could be provided by health and social care services, non-government organisations and/or community organisations working together, could enable families to achieve their desired outcomes and to support self-management of pain.

In the healthcare services, environment is where triage and pain assessment by healthcare professionals occur which are fundamental steps for pain treatment and prevention. An effective triage system is underpinned by a biopsychosocial assessment of chronic pain, including mapping the family’s psychosocial as well as medical needs; prioritises pain relief and living well with pain; and triggers appropriate action from different services. For instance, triage could trigger referral to appropriate specialised services and multidisciplinary teams, and/or the mobilisation of schools, social care and social services to evaluate the level of support the CYP and family require. Empathetic communication by professionals is a core component of this approach, which is fundamental for building trust between families and services. Healthcare professionals look holistically at the impact of pain on families’ lives and listen to the expert knowledge of parents, seeing parents and children as true active agents in the decision-making processes, while at the same time guiding and supporting parents in helping their child to manage pain. The impact of treatments on family life and relationships, and vice versa, is considered. When appropriate, services and healthcare professionals help families set expectations regarding the possibility of a cure for the pain. The front-line staff, such as children’s nurses, are trained to use simple interventions when appropriate to avoid repeated returns to services and referral fatigue. Specifically, simple early interventions are the first option when appropriate. Families who need them have access to suitable multidisciplinary teams (including, e.g. pain clinical specialists, nutritionists, physiotherapists, psychologists, occupational therapists, social workers and social care workers, etc.) according to their needs and underlying conditions. Accessibility of services is considered as part of the pain management plan. The transition to adults’ services is done gradually and planned as part of the pain management strategy.

To enable healthcare services to work efficiently as part of the system, the provision of training and resources for healthcare professionals are needed. Training and resources are aimed at differentiating chronic and acute pain and helping healthcare professionals to understand the impact of chronic pain and any associated condition including an awareness and understanding of the importance of the biopsychosocial approach.

In the community/school environment is where much self-management and pain prevention occur, in addition to that which occurs in the home environment. Social support, which could include peer support, for families and CYP is tailored according to the impact on their lives and mediated by social care, social work or non-governmental organisations. School and work policies are in place to allow adjustments to life in the community/schools allowing for pain prevention and management. For example school policies facilitate a routine for CYP to prevent pain exacerbation including flexible school hours and support for children to catch up on schoolwork. Work policies for parents of CYP with chronic pain allow flexible work hours and remote work opportunities. To enable schools/community to work efficiently as part of the system, it is fundamental that resources and training are provided to help staff at school and work to understand chronic pain and condition and how it can impact family life. The community/school environment can provide or facilitate social support and help to create a sense of belonging, which can help families to deal with the impacts of chronic pain on their family and social life. Social support interventions and policies consider how families’ and CYP’s social groups understand chronic pain and pain management, and whether CYP want to disclose their pain condition.

Managing pain is essential to enable families to live well; it is a priority and a combined effort involving multiple systems, such as health services, schools, social care and social services. Proactive support for families, for example via check-up phone calls or digital clinics, is in place to support families to manage pain across all the different environments (health services, schools/community and at home).

Matrix analysis

The results of the matrix analysis (Table 7) show that, of the 14 reviews, no reviews focused on whole-family interventions. Only one review focused on interventions for parents of children with chronic illness, and measured family functioning. 27 Only two reviews (14%) measured parental outcomes – one analysed parenting behaviours and parental mental health, including that of mothers and fathers, although they did not perform a subgroup analysis by parental sex27 and the other planned to analyse caregiver distress but none of the included studies reported this. 28 None measured sibling outcomes.

All 14 Cochrane reviews specified that their inclusion criteria included study outcome measures of child psychological/well-being outcomes such as quality of life, depression, anxiety, psychological functioning. However, in five (36%) reviews,22,24–26,23 none of their included trials had analysed and reported these outcomes. The inclusion criteria of only four (29%) reviews included measures of child social outcomes including school attendance in three reviews,46,47,50 social functioning in three reviews47,48,50 and social disruption in one review. 50

The inclusion criteria of eight (57%) reviews included pain outcome measures which measure absence of pain, such as the Numerical Rating Scale, the Visual Analogue Scale, the Wong Baker FACES pain scale and/or other pain rating scales. 27,28,46–48,50,54,154 The inclusion criteria of six (43%) reviews did not include a measure for absence of pain,22–26,49 five of which were focused on the evaluation of pharmacological therapies. 22–26

Regarding under-represented subgroups of children, specifically children with autism or learning disabilities and children under 5 years old, nine (64%) of the Cochrane reviews specified in their inclusion criteria that all children up to the age of 18 years old with the pain diagnosis of interest were eligible to be included in analyses. 22–27,49,54,154 However, in all of those nine reviews, none of the eligible trials had included children with learning disabilities/autism or aged under 5 years in their samples. The inclusion criteria of five (36%) of the Cochrane reviews, four on recurrent abdominal pain46–48,50 and one on musculoskeletal pain,28 focused only on school-aged children aged 4 or 5–18 years old and thus excluded younger children. For the abdominal pain reviews, this was due to the use of diagnostic criteria which do not apply to younger children. The inclusion criteria of eight (57%) reviews included children under 5 years, but none of their included trials had participants under 5. The inclusion criteria of two (14%) reviews specified in the protocol that children under 5 years were included but only the mean age was provided in the reviews. 27,54 Only one review49 had planned to conduct subgroup analyses by ethnic origin of participants, although there were insufficient data to permit this.

In summary, existing Cochrane intervention effects reviews on children’s chronic pain focused on the psychological impact on CYP, but have not measured, or have been unable to measure, many of the outcomes of importance to children and their families (e.g. the psychosocial impact on whole family). Few have results relevant to children under 5 years old or children with learning disabilities or autism, either because of the inclusion criteria of the review or the trials it synthesised. None of the studies had data which inform outcomes for children or parents from different ethnic origins.

Analysis of evidence for the chosen intervention (programme theories)

Programme theories explain how an intervention is thought to work to achieve an outcome. However, in all reviews, their sections about how the interventions might work to manage pain (described in some reviews as the ‘mechanisms of action’) were brief and lacking in detail. None had a clearly discernible biopsychosocial explanation of how the intervention(s) might work. Seven (57%) reviews exploring pharmacological interventions22–26,46,47 and one exploring dietary interventions for recurrent abdominal pain50 described only the biomedical ‘mechanisms of action’ without consideration of the social and personal contexts/factors affecting, for example medication adherence (Table 8).

In some reviews, it was difficult to discern how the interventions were thought to work and whether they adopted a biopsychosocial or another approach. Biological (e.g. acting on physical mechanisms) and psychological (e.g. influencing cognitions, coping) aspects were apparent in programme theories in two reviews – one focusing on psychological therapies for sickle cell disease and pain49 and one on physical activity and education about physical activity for chronic musculoskeletal pain. Nevertheless, both lacked detail or precise mechanisms:28,49 Anie et al. 49 specified expected outcomes of interventions rather than mechanisms and Nascimento Leite et al. 28 stated only that the interventions affect physical and psychological mechanisms (Table 8).

Four more reviews investigated psychological or psychosocial interventions only, and did not appear to consider any biological mechanisms. 27,48,54,154 Two of these reviews, one focusing on psychological therapies27 and one focusing on psychosocial interventions,48 covered multiple types of interventions, each with different programme theories. Abbott et al.’s48 review included very brief descriptions of how cognitive–behavioural therapy, hypnotherapy, guided imagery, yoga and written self-disclosure are thought to work, all of which were described as mainly psychological in nature, except for family therapy, which was social (in contrast to Law et al. ’s27 description of family therapy as psychosocial). Law et al. ’s27 review also included separate programme theories for a variety of interventions including motivational interviewing, which was described as having a psychological mechanism, and those described as both social and psychological: cognitive-behavioural therapy, family therapy, multisystemic therapy, problem-solving therapy and one review investigating psychological therapies appeared to describe a psychosocial mechanism about which they gave little detail. 154

In summary, programme theories in Cochrane reviews of intervention effects on children’s chronic pain were brief and lacking detail making it challenging to ascertain whether explanations were biopsychosocial or not. Nor were there any considerations about feasibility, acceptability or implementation as additional context to the programme theories.

Overall, our integration of findings has shown that outcomes measured in Cochrane reviews and trials of interventions do not always match those of importance to children and families. In some cases, this is possibly because of a lack of suitable measurement tools. We also found that a biopsychosocial approach has been largely neglected in programme theories.

Review question 1. How do children with chronic pain and their families conceptualise chronic pain?

Our findings (presented in our Line of argument 1: dominance of biomedical perspective in how children and families conceptualised chronic pain and pain management) show that there is a difference in how CYP and families conceptualised the impacts of pain versus its causes, and therefore, the role of health services in managing the pain. CYP and their families expressed biopsychosocial (biological, psychological and social) impacts of chronic pain on their lives, for all kinds of chronic pain. However, when there was no identifiable or known medical cause of the pain, families initially tended to reject the idea of psychosocial causes or influences (e.g. anxiety), because it challenged the legitimacy of the CYP’s pain or suggested that pain was caused by parental shortcomings. For children whose pain was identified as being caused by an underlying health condition, such as sickle cell disease, juvenile idiopathic arthritis or cerebral palsy, families were less likely to consider psychosocial causes of, or influences on, pain. Most CYP and parents tended to understand the cause of chronic pain as solely biomedical – having an underlying pathophysiological cause. This influenced their expectations of and interactions with health services. However, families’ conceptualisations were dynamic and could change over time with experiences of treatments, services and interactions with healthcare professionals. Thus, healthcare professionals and targeted interventions could play an important role in managing family understandings of pain and expectations of services.

It is important to highlight that the above conceptualisation stems mainly from research in the UK and other high-income countries and their ethnic populations. Nonetheless, our included studies did represent a fairly wide range of ethnic backgrounds. Families from other countries and ethnic and cultural backgrounds may have very different conceptualisations to that described above. Furthermore, menstrual pain may be conceptualised differently. Girls and young women, their families and healthcare professionals may not at times perceive pain as a pathological/pathophysiological issue, for example severe menstrual pain was perceived as ‘normal’ pain not requiring medical intervention.

Review question 2. How do children with chronic pain and their families live with chronic pain?

The impacts of poorly managed, moderate and severe chronic pain on the child, their whole family and their wider social connections and relationships were profound (described in Third-order construct 1: pain organises the family system and the social realm). There were serious negative impacts of children’s chronic pain on family systems. Pain affected family dynamics and relationships; family members’ emotions, well-being, autonomy and sense of self-identity; parenting strategies; friendships and socialising; children’s education and future employment prospects; and parental employment. Children and family members experienced adverse physical/biological, psychological and social impacts of chronic pain. We found that most pain management happened at home and in the community, for example at school, and was supervised or implemented by parents, especially mothers who were often the main caregivers. Thus, the demands of caregiving disproportionately affected women (mothers).

Even though there were many serious psychosocial impacts on families, there was very little support provided by health and social care services or in the community (e.g. school health services) to address these. Families needed greater family-centred psychosocial support, which could include social support in the form of peer support. Children benefited from awareness about pain and its associated impact on CYP within schools and school health services, for example some school policies and practices flexibly supported CYP with chronic pain in self-care, while avoiding stigmatisation, and avoided negative impacts of disruption to their education. Health providers, schools and families working together to achieve outcomes valued by CYP and families is the basis of a ‘whole systems’ approach which, in turn, is based on strong connections and communication.

Review question 3. What do children with chronic pain and their families think of how health and social care services respond to and manage their/their child’s chronic pain?

A key challenge identified by CYP and families concerned their experiences of health systems and individual health services as difficult to access and navigate, resulting in challenges to achieve satisfactory pain management (as described in Third-order construct 2: families struggling to navigate health services, and Line of argument 3: the journey of living with chronic pain). Families often entered a lengthy cycle, which could last many years, of attending then reattending health services, and awaiting referral to secondary or tertiary services for tests and/or treatment. In the meantime, pain was not usually well managed, and families were left largely unsupported to find their own solutions. Being referred to a pain clinic could ‘validate’ the pain (i.e. confirm it was a real illness), be a source of hope, and emotional and practical support. One of the reasons for families’ difficulties in navigating services was the lack of a clear care pathway for CYP with chronic pain.

Interactions with healthcare professionals were frequently negative and stressful. CYP and their main caregivers, usually mothers, felt at times they were not being listened to and believed and thus they felt unsupported. For instance, CYP’s experiences of living with pain were ignored or their pain was disbelieved, and mothers felt blamed for their child’s pain. Parents of younger children and of children with communication difficulties perceived that their expertise in assessing their child’s pain was not always recognised by healthcare professionals. Open, empathetic healthcare professional communication could help CYP and their families to develop a trusting relationship with healthcare professionals. Children and families wanted healthcare professionals to explore their experiences of the impacts of chronic pain and their priorities for pain management.

Current tools to assess pain in CYP do not seem to incorporate parental assessment and advice regarding children with special needs (e.g. non-verbal/unable to communicate), and there appear to be no tools suitable for all CYP, tailored for chronic as opposed to acute pain. CYP may hide or minimise their pain to avoid causing their parents’ distress which is likely to pose challenges for pain assessment. Parents wanted healthcare professionals to recognise and make use of their parental expertise in assessing and responding to their child’s pain, particularly for children who have communication difficulties. CYP wanted healthcare professionals to use the child’s expertise and knowledge of their pain, especially for those with underlying chronic conditions like sickle cell disease and juvenile idiopathic arthritis. Family and child’s preferences, needs and lifestyle can have implications for pain management and pain management can affect family life.

Our evidence suggests that a scattergun rather than a co-ordinated approach to treatment decisions was common, with few children referred to specialist pain clinics when it was appropriate to do so. Many CYP did not appear to be offered the full range of appropriate treatment options; hence, parents tended to seek out one option after another in an unco-ordinated way that was not based on evidence of potential appropriateness, benefits or potential harms. In the data, there was an apparent lack of use of shared decision-making frameworks by healthcare professionals that could help them signpost families to all the available and appropriate treatment options.

There are difficulties diagnosing certain types of chronic pain, especially for generalist healthcare professionals, which would require diagnostic input from tertiary services. There is currently no pathway for children’s chronic pain, including chronic pain of unknown cause, for CYP and families and healthcare professionals to follow. Services were fragmented and some children who needed it waited a long time to be referred to or access tertiary specialist pain services, even for pain conditions like CRPS, which might benefit from early referral to such services. There was a lack of data regarding screening and assessment tools for children’s chronic pain which could have facilitated triage of children, for example to specialist services or use of primary care interventions, such as signposting families to educational pain management resources.

Lastly, but importantly, CYP and their families, particularly CYP with sickle cell disease, experienced racism and discrimination in health services. Children with chronic pain and learning disabilities, such as Down syndrome, also encountered discrimination. Direct and indirect prejudice and discrimination towards patients adversely affected the care they received and their pain management.

Review question 4. What do children with chronic pain, their families conceptualise as ‘good’ chronic pain management and what do they want to achieve from chronic pain management interventions and services?

As described in Third-order construct 2: families struggling to navigate health services and line of argument 1 (see Line of argument 1: dominance of biomedical perspective in how children and families conceptualised chronic pain and pain management), our findings indicate that because of the way families initially conceptualised the cause of chronic pain as biomedical, they first approached health services looking for medical treatment and/or cure and they usually rejected psychological treatments or approaches. Families with diagnosed underlying health conditions (such as juvenile idiopathic arthritis, sickle cell disease) which cause chronic pain also might not see the value of psychosocial pain treatments, given the cause of pain is a disease. Families did want help with the psychosocial consequences of pain, such as parental distress, but did not think health services could assist with that. This perceived medical focus for families and healthcare professionals could act as a possible barrier to biopsychosocial care.

Families’ reluctance to accept there may be psychosocial influences on, or causes of, chronic pain (or them not seeing the role of psychosocial support in managing pain) means healthcare professionals need to carefully manage how and when they offer psychosocial support to avoid alienating them. Healthcare professionals could play a role in sensitively explaining biopsychosocial explanations of pain management to families – since these apply even when there is a biological cause – by building on the family’s prior experiences of the biopsychosocial impacts and aspects of pain. However, when not approached sensitively, children and their families can feel healthcare professionals do not believe their pain is real and parents can feel they are being blamed for any social and psychological causes of pain. Biopsychosocial approaches are important for all CYP and families managing pain whatever the cause of their chronic pain. Parents and CYP highly value healthcare professionals who listen to and understand the impact chronic pain has on them and their family life.

Families ideally wanted interventions that reduced or removed the pain; if that is not possible, then strategies to help them live well with pain would be beneficial. Interventions that do not worsen the impact on the whole family and its well-being would be more acceptable to families. Consequently, it is important to tailor interventions to families’ preferences and needs so that they are not boring or excessively time-consuming, not painful and have minimal side effects.

An important role for healthcare professionals is in managing families’ understanding of chronic pain and expectations of pain management. Families needed clear communication about the purpose of any tests, examinations and treatments and to have their expectations managed about the possibility of a cure (which will be condition dependent) from the outset of their contact with health services. CYP needed to have their pain managed while awaiting test results and diagnosis.

Existing health and social care service models are adult-orientated and not child-specific or child- and family-centred. Biopsychosocial health and social care tailored for CYP, which is family-centred considering needs of the entire family and the impact on the family system would potentially meet families’ needs. Multidisciplinary care/treatment is currently recognised as the ‘gold standard’,18 but there is limited availability and accessibility of specialist multidisciplinary pain management for children’s complex pain problems. 8,18 There is a lack of a clear clinical pathway or specialist condition-specific pathways for particular pain conditions, for example juvenile idiopathic arthritis, headache.

Situation of the findings in relation to wider literature

Two qualitive evidence syntheses, published prior to conducting our meta-ethnography, had focused on children’s experiences of living with juvenile idiopathic arthritis39 and the impact of chronic pain on adolescents’ interpersonal relationships. 9 Our findings concur with, but greatly extend, theirs which found predominantly negative impacts on CYP’s self-identity, social participation and adolescents’ relationships. 9,39 Since we conducted our meta-ethnography, two new qualitative evidence syntheses on children’s chronic pain have been published focusing on specific pain populations and topics: a thematic synthesis of adolescents’ experiences of fluctuating pain in musculoskeletal disorders,155 and an aggregative synthesis of the experiences of children and young adults (up to 21 years) and their parents of living with juvenile idiopathic arthritis. 156 Khanom155 found that pain flares restricted adolescents’ activities leading to them feeling isolated and different from peers. The fluctuating nature of the pain was difficult for others to understand and adolescents tried strategies, including medical treatments, social support and concealing the pain, to try to regain a feeling of control over the unpredictable pain. 155 Min156 found adverse effects of juvenile idiopathic arthritis on young people’s emotional and physical well-being and on their relationships with parents, teachers and healthcare professionals. Young people and parents needed more self-management knowledge and skills. 156 A further mixed-methods review synthesised three qualitative studies on the impact of pain on adolescents’ schooling found mainly negative impacts of pain on school functioning with a need for better pain management in schools. 157 However, these five syntheses had narrower focuses and populations and different aims from our meta-ethnography. Their findings confirm but do not substantively add to ours and none produced a theory. We developed a new mid-range theory of children’s chronic pain and its management (see Family-centred theory of children and young people’s chronic pain and its management), which can be tested. Consequently, our meta-ethnography still adds valuable, key knowledge to the growing literature, but on a much wider range of chronic pain conditions and how they affect all aspects of the lives of CYP with chronic pain and their families.

In our meta-ethnography, we found that families in high-income countries mainly conceptualised pain as having a biomedical cause. This perceived medical focus can act as a possible barrier to biopsychosocial care. Shortcomings in healthcare professional education, training and experience about chronic pain could also contribute to the biomedical focus. 8,30,158 However, families from other countries and ethnic and cultural backgrounds may have very different conceptualisations. For example in Cameroon, families often believed pain was caused by witchcraft or their ancestors,149 in Jordan, many families believed that pain is given by God and should not be treated,159 while in Morocco, pain was seen as a sign of weakness. 160 Furthermore, in our meta-ethnography, severe menstrual pain was perceived as ‘normal’ pain not requiring medical intervention. This finding originated from a single study, conducted in Hong Kong; however, it is also supported by the UK and global literature on dysmenorrhoea. 160

Our meta-ethnography indicates serious negative impacts of children’s chronic pain on family systems. We have added to the scant knowledge of the impact of childhood chronic pain on the whole family system. 161 Quantitative cross-sectional research10 shows family dysfunction in children’s chronic pain but cross-sectional research cannot indicate the direction of causal relationships. Notably, our meta-ethnography shows that pain can cause family dysfunction and family-system-wide changes. Family systems interventions for CYP with chronic pain could be beneficial for the whole family. 162 The demands of caregiving can have negative impacts on the whole family but also can contribute to gender inequalities between women (mothers) and men (fathers); therefore, health and social care policies and interventions should take care not compound this inequality by increasing the burden on women. 163

Families lacked family-centred psychosocial support. The UK clinical SIGN guidelines on children’s chronic pain do not emphasise that care should be holistic and family-centred, perhaps because it is based on use of existing NHS services which are designed mainly to treat individual patients’ medical needs. 8 The guidelines do specify that ‘early biopsychosocial assessment and psychological intervention should be considered’8 (p.12). While the SIGN guideline fits with some of the families’ needs we identified, there is a lack of family-centred focus in the guideline as well as a lack of healthcare professionals with the knowledge and expertise in chronic pain assessment and management. 8,18,30,158,164,165

In palliative end-of-life care (in the last weeks and months of life), the concept of ‘total pain’ is commonly used, that is, pain is not just biological in cause but can be caused by emotional distress, existential concerns and so on, and the physiological impacts cannot be separated out from the emotional and social impacts of the pain; therefore, they need to be managed as connected. 166,167 This concept of total pain has not been applied to chronic pain that is not at end of life; it may be a useful concept to facilitate biopsychosocial management of children’s chronic pain. Adopting a ‘total pain’ approach to service delivery would help to knit together the biopsychosocial influences on and consequences of pain. 166,167 A barrier to biopsychosocial approaches to managing pain is that health services also largely still operate on a biomedical model of understanding, assessing and treating pain. 18 It is accepted that healthcare professional chronic pain education urgently needs improving, for example by adapting medical education curricula for undergraduates, postgraduates and qualified healthcare professionals. 8,18,30,31,164,165

The recent Lancet Commission identified the importance of healthcare professionals developing a more detailed understanding of pain which provides an understanding of children’s experience of living with chronic pain. 18 Denying a child’s subjective experience of pain is likely to result in child and family distress. Children and families needed healthcare professionals to believe the child’s pain was real, regardless of the (possible or actual) cause; this synthesis finding supports a key recommendation in the Lancet Commission. 18

An important role for healthcare professionals is in managing families’ understanding of chronic pain and expectations of pain management. Appropriate existing high-quality information resources for CYP and families exist, such as leaflets, storybooks for CYP, peer support groups, online modules, websites, for example ‘Kids In Pain’,168 ‘It Doesn’t Have To Hurt’169 websites. 8 In addition, healthcare professionals could make greater use of shared decision-making frameworks (e.g. the MAGIC framework),170 alongside appropriate clinical pathways, which could aid families to discuss and choose from all the available and appropriate treatment options to facilitate better decision-making.

It is not clear if and how assessment tools specific to chronic as opposed to acute pain are used in practice; there were no data on this in our included studies, nor that we can identify in the healthcare professional literature. SIGN guidelines8 recommend screening tools, for example the Paediatric Pain Screening Tool (PPST),171 which are specifically for chronic pain, to identify those children at risk of adverse outcomes. In addition, there are some other chronic pain paediatric assessment tools for specific age groups, such as self-report questionnaires for adolescents and their parents172 and a measure patient-reported pain behaviours for school-aged children. 173

Current barriers to improving children’s pain management in the UK include the lack of education of healthcare professionals in chronic pain and its management8,18,30,31,164,165 and the lack of specialist pain management services. 8 Furthermore, one of the reasons for families’ difficulties in navigating services is the current lack of a clear care pathway for children with chronic pain. In the UK, only Scotland has a clinical guideline for children’s chronic pain management;8 the pathway it contains mainly describes how services are currently organised rather than a step-by-step guide to pain assessment and treatment with clear decision points for healthcare professionals and families. There is no NICE guideline for England, Wales and Northern Ireland specifically for children’s chronic pain management. 29 NICE guidelines for sickle cell disease174 are not specific to children and the headache guideline175 is only for over 12s; yet, children’s needs may differ from adults’ needs and younger children’s needs may differ from those of older children. Other painful conditions such as juvenile idiopathic arthritis, dysmenorrhoea, epidermolysis bullosa, Ehlers–Danlos syndrome and osteogenesis imperfecta have no NICE clinical guideline, although some (e.g. dysmenorrhoea, juvenile idiopathic arthritis) have clinical summaries which provide some guidance. Child-specific chronic pain clinical guidelines could draw on existing guidelines8,19 and robust synthesised evidence. 20,40

We found that CYP and their families, particularly CYP with sickle cell disease, experienced racial prejudice and discrimination in health services. A possible alternative interpretation is that the inadequate care of CYP of mainly African Caribbean origin was related to the widespread suboptimal care of children with chronic pain in general rather than due to racism. However, in their recent comment in The Lancet Commission, Inusa et al. 176 also identified inequalities in access to care for CYP with sickle cell disease across various high- and low-income countries. Furthermore, an editorial literature review in the Journal of Pain Research177 concluded that there was systemic racism in chronic pain management, although in the US context. Children with chronic pain and learning disabilities, such as Down syndrome, also encountered discrimination, a finding which is supported by the wider literature on the poorer health care provided to people with learning disabilities. 178

Our new comprehensive theory goes beyond existing theories to inform design of chronic pain management and services by adopting a family-centred, biopsychosocial, whole-systems approach. In contrast, existing theories have either too wide a focus to derive clear testable hypotheses or too narrow a focus, such as on the causes of pain, a developmental perspective concerning how children understand their chronic pain,41 a biopsychosocial approach to clinical assessment and management of children’s chronic pain,43 or the role of parent and family factors in children’s chronic pain. 179,180 The implications for practice, described in Review question 1. How do children with chronic pain and their families conceptualise chronic pain?, Review question 2. How do children with chronic pain and their families live with chronic pain?, Review question 3. What do children with chronic pain and their families think of how health and social care services respond to and manage their/their child’s chronic pain? and Review question 4. What do children with chronic pain, their families conceptualise as ‘good’ chronic pain management and what do they want to achieve from chronic pain management interventions and services?, and for future research are summarised in Table 9.

Participant representation

Our sample of included studies specifically sought to represent diverse populations. However, we were restricted by the available primary qualitative study publications in terms of the study population characteristics and how inclusive the populations were. We identified and have described evidence gaps and under-represented groups (see Included studies and Table 9). The views and experiences of boys, and children under 10 years old with chronic pain, CYP with learning disabilities, fathers and siblings are less well represented.

We have described our PPI recruitment strategy and engagement in detail (see Chapter 3). In short, for our core PPI group, we successfully recruited CYP with different kinds of chronic pain ranging in ages from 8 years old upwards and mothers of children with chronic pain, but we did not have any fathers directly involved despite recruitment attempts. Group members came from England, Scotland and Wales. We tried to recruit children and parents from a variety of ethnic backgrounds through third-sector organisations but were unsuccessful. We also used social media and online surveys for wider PPI engagement, which may have reached a more diverse group of children and families, but their feedback was anonymous.

Reflections on our research team and wider involvement

Our core research team is made up of adults (all female) some of whom have lived experience of chronic pain, which started in childhood or adulthood. Six team members are parents, some of whom have children with long-term health and/or developmental conditions. Our meta-ethnography focuses on CYP with chronic pain and their families; they were not part of the core research team but were members of our PPI and advisory groups. Our core team has diversity in nationalities (Belgian, Brazilian, British, French) and professional backgrounds in nursing children with chronic pain (JN), psychology (EF), health psychology (AJ, LC), sociology (IU, RT), family therapy (LF) and physiotherapy with children with chronic pain (MSB), children’s pain research (AJ, LC), occupational therapy (KT) and health services research (all).

As a result of our diverse backgrounds, there were some divergent views on aspects of the findings, which for transparency we record here. Two team members with extensive expertise in children’s pain research were uncomfortable with use of the term ‘cure’ in relation to chronic pain because for most young people there is no cure for their chronic pain. Therefore, it might be more scientifically accurate to use terms like ‘alleviate’ or ‘reduce’ pain. However, we retained the term ‘cure’ because it was used by research participants in a large proportion of the included primary studies and refers to their focus on absence rather than reduction of pain. There was also a difference in team members’ opinions about whether we should refer to ‘potential racism and prejudice’ or ‘racism and prejudice’ in health services because of difficulties in knowing whether poor treatment of CYP with chronic pain was related to their protected characteristics. We decided to refer to ‘racism and prejudice’, given that included studies reported racism and there is further supporting evidence in the wider literature, as reported above.

There was a range of experience and expertise across the research team including a new post-doctoral researcher, early-career and mid-career researchers, as well as senior academics. In terms of development opportunities, the junior member of the team was provided with mentorship from more senior members; accessed institutional and external training, including writing workshops and a grant writing development group; Cochrane systematic review training seminars including use of GRADE-CERQual; and was supported to present at conferences and lead journal articles.

Our wider team included a PAG including clinicians with pain expertise, a general practitioner, physiotherapists, Scottish Government representatives, an academic expert in qualitative evidence synthesis, voluntary sector representatives from patient organisations and a PPI group of parents and CYP with chronic pain.