Article Text

Download PDFPDF

Why colorectal screening fails to achieve the uptake rates of breast and cervical cancer screening: a comparative qualitative study
  1. Marie Kotzur1,
  2. Colin McCowan2,
  3. Sara Macdonald1,
  4. Sally Wyke1,
  5. Lauren Gatting1,
  6. Christine Campbell3,
  7. David Weller4,
  8. Emilia Crighton5,
  9. Robert J C Steele6,
  10. Kathryn A Robb1
  1. 1 Institute of Health and Wellbeing, University of Glasgow, Glasgow, UK
  2. 2 School of Medicine, University of St Andrews, St Andrews, Fife, UK
  3. 3 Usher Institute, University of Edinburgh, Edinburgh, UK
  4. 4 General Practice, University of Edinburgh, Edinburgh, UK
  5. 5 Public Health Directorate, NHS Greater Glasgow and Clyde, Glasgow, UK
  6. 6 School of Medicine, University of Dundee, Dundee, UK
  1. Correspondence to Dr Kathryn A Robb, Institute of Health and Wellbeing, University of Glasgow, Glasgow G12 8QQ, UK; katie.robb{at}glasgow.ac.uk

Abstract

Background In Scotland, the uptake of clinic-based breast (72%) and cervical (77%) screening is higher than home-based colorectal screening (~60%). To inform new approaches to increase uptake of colorectal screening, we compared the perceptions of colorectal screening among women with different screening histories.

Methods We purposively sampled women with different screening histories to invite to semistructured interviews: (1) participated in all; (2) participated in breast and cervical but not colorectal (‘colorectal-specific non-participants’); (3) participated in none. To identify the sample we linked the data for all women eligible for all three screening programmes in Glasgow, Scotland (aged 51–64 years; n=68 324). Interviews covered perceptions of cancer, screening and screening decisions. Framework method was used for analysis.

Results Of the 2924 women invited, 86 expressed an interest, and 59 were interviewed. The three groups’ perceptions differed, with the colorectal-specific non-participants expressing that: (1) treatment for colorectal cancer is more severe than for breast or cervical cancer; (2) colorectal symptoms are easier to self-detect than breast or cervical symptoms; (3) they worried about completing the test incorrectly; and (4) the colorectal test could be more easily delayed or forgotten than breast or cervical screening.

Conclusion Our comparative approach suggested targets for future interventions to increase colorectal screening uptake including: (1) reducing fear of colorectal cancer treatments; (2) increasing awareness that screening is for the asymptomatic; (3) increasing confidence to self-complete the test; and (4) providing a suggested deadline and/or additional reminders.

  • qualitative research
  • decision making
  • human factors
  • health policy

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Background

Colorectal cancer (CRC) is the second leading cause of cancer death worldwide.1 CRC screening by faecal occult blood test (FOBt) can reduce deaths.2 In Scotland, since 2007, people aged 50–74 have been mailed a self-complete FOBt every 2 years as part of the Scottish Bowel Screening Programme. The FOBt requires people to collect two samples from each of three separate bowel motions and to mail their completed kit for processing. Women in Scotland are also invited to attend prearranged appointments for breast screening using mammography and to make appointments for cervical screening using the Pap smear (table 1). All three screening tests are offered at no cost to participants through the National Health Service (NHS). Uptake of screening is 77% for cervical, 72% for breast but only 59% for colorectal among women aged 50 and over.3–5 Screening uptake rates show similar patterns in Australia and the USA with uptake of CRC screening lagging behind the participation rates of breast and cervical screening.6 7

Table 1

Invitation procedures in the Scottish cancer screening programmes

Reasons for the low uptake of CRC screening include lack of awareness, feeling healthy, negative views of cancer (fear, fatalism), negative attitudes towards colorectal tests, lack of motivation including other health concerns, and cultural, gender and socioeconomic influences.8–11 Many of these potential barriers also apply to breast and cervical screening,12 so it is unclear why CRC screening uptake should remain considerably lower. There have been few comparisons of barriers across these three screening modalities.13 14 We identified only one study that directly compared barriers to breast, cervical and CRC screening uptake among women eligible for all three tests.15 This British self-report survey of 890 women found that among those who participated in breast and cervical but not colorectal programmes, 23% reported not liking the idea of CRC screening test and 18% said they ‘haven’t got round to it, but intend to take part’ as explanations for non-participation in CRC screening.15 These explanations relate to both motivational influences such as dislike of the test, and volitional aspects of ‘not getting round to it’,16 the latter being particularly salient for CRC screening which, unlike breast and cervical screening, is self-completed at home. Our study adds to those data in three main respects. First, rather than using self-reported data of screening history, we linked cancer screening uptake data for the breast, cervical and CRC screening programmes for the complete population of Glasgow, Scotland—a socioeconomically diverse region with low overall screening uptake. Using this linked data set, we identified women with three different screening histories: (1) participated in all programmes; (2) participated in breast and cervical but not the colorectal programme; and (3) did not participate in any programme. Second, we invited women across these three groups to an individual in-depth interview, rather than a questionnaire, to provide the opportunity for women to speak at length about their perceptions and experiences of cancer screening.

Third, we organised our findings using the route MAP approach which is a useful tool to summarise the central tenets of multiple models of behaviour change.17 The MAP describes three routes to behaviour: (1) motivation—strategies that increase and sustain motivation (eg, information about the behaviour, reassurance); action on motivation—strategies that strengthen and elaborate skills needed to translate motivations into action (eg, setting behavioural goals, action and coping planning); and prompted or cued routes—strategies that support behaviour change without the continuous cognitive effort required by the motivation and action-on-motivation routes (eg, prompt, change the environment to facilitate the target behaviour). The MAP approach therefore provides a theoretically informed framework to identify targets for intervention.

The present study was designed to (1) identify why women (who are eligible for all three types of screening) choose to participate in breast and cervical screening but not CRC screening, and (2) gain insight into how CRC screening uptake can achieve the uptake rates of breast and cervical screening.

Methods

Participants and recruitment strategy

This study was conducted alongside a quantitative study exploring cancer screening uptake among women living in Glasgow, Scotland.18 Data on screening participation for the breast, cervical and CRC screening programmes were linked for all women aged 20–74 (n=430 591) who were registered with NHS Greater Glasgow and Clyde Health Board and invited to screening during the period 2009–2013. This linked data set was used as a sampling frame for the present study to select women who were eligible for breast, cervical and CRC screening (n=68 324). Because we expected screening experiences to differ by screening history and on socioeconomic position we aimed to interview women in each of six groups (table 2).

Table 2

Sampling frame

Purposive sampling was used to randomly identify and invite to interview women with three different screening histories: (1) women who participated in all programmes (screening participants); (2) women who participated in breast and cervical but not the colorectal programme (CRC-specific non-participants); and (3) women who participated in none (non-participants). The sample was also stratified to obtain a mix of women from areas of high or low socioeconomic deprivation (table 2). Socioeconomic deprivation was indexed by the Scottish Index of Multiple Deprivation (SIMD), an area-based measure of multiple deprivation linked to each individual’s home address.19 The aim was to obtain approximately 10 interviews across the six stratified groups (table 2). It was anticipated that recruiting non-participants and women living in deprived areas would be more challenging.20 To ensure we achieved equal numbers of participants across the six groups we significantly oversampled non-participants and women living in deprived areas. In total, 2924 women were invited (table 2).

The research team provided the Glasgow Clinical Research Facility with the sampling frame to identify 2924 women of the 68 324 eligible for all three screening tests to be sent invitation packs. Invitation packs contained a letter inviting women to participate in a study exploring views on colorectal, breast and cervical cancer screening, a participant information sheet and a response form with options to indicate interest by email, phone or using a prepaid envelope enclosed. Respondents were offered £20 for interview participation and to cover the cost of refreshments and travel to the interview. Only women responding to the invitation pack were subsequently identified to the research team.

Interview procedure and materials

The interviewer (MK) met with respondents at locations of their choice: home (n=35), the University of Glasgow (n=15), work (n=3), community centre (n=5), a parish house (n=1). MK had spoken to the participants by phone prior to the interview, but had no relationship with them otherwise. MK is female and had previous experience of conducting interviews and focus groups with women about cancer screening. All participants provided informed consent before the semistructured interview started. The interview followed a topic guide and began with an open question, ‘What comes to mind when you think about cancer?’ This was followed in turn with questions on what comes to mind in relation to colorectal, breast and cervical cancers. The next question asked how they felt about their chances of developing breast, cervical or CRC. Respondents were then asked for their thoughts and feelings about cancer screening. The interviewer asked, ‘What comes to mind when you think about bowel cancer screening?’ with supplementary questions on their understanding of what the test involves and the purpose. This was followed by the same questions in relation to breast and cervical screening. Finally, respondents were shown example invitation letters and leaflets and were asked how they felt when they received these and how they decided what to do next. The interviews took place between November 2015 and April 2016, were an average length of 43 min, audio recorded and transcribed verbatim.

Analysis

The transcripts were analysed using the framework method, which takes a matrix-based analytical approach to facilitate rigorous and transparent data management.21 This approach permitted comparisons of accounts between women with different screening histories and living in different socioeconomic circumstances.

MK, KR, and SM read and re-read the data. A mind-mapping process was developed based on the one-sheet-of-paper method.22 From these mind maps, MK identified themes as the basis of the framework matrix. The themes were organised into the three MAP routes: motivation; action on motivation; and prompts.17 As described by Ritchie and colleagues,21 the framework matrix was organised in an MS Excel spreadsheet containing one theme per worksheet with subthemes in the columns. The rows contained individual participants, grouped by screening history and socioeconomic status. The themes and subthemes within the framework matrix were discussed within the research team. Two researchers (MK and LG) coded the transcript in full and populated the framework matrix with relevant data extracts. They discussed comments and queries using web-based collaboration software (Trello), thus creating an audit trail. MK summarised each theme by comparing patterns within the preidentified participant categories (screening history and socioeconomic status) and noting representative data extracts. KR reviewed and discussed the summaries with MK to ensure consistency with the data.

Results

Sample characteristics

Of the 2924 women sent an invitation to participate; 2629 did not respond, 129 declined, 76 invitations were returned undelivered, 4 women had died and 86 expressed an interest in being interviewed of which four were excluded (due to becoming uncontactable, residing primarily outside the UK, or having a stoma). In total, 61 interviews were scheduled and 59 women were interviewed.

Response rates varied considerably across the six groups; 55% of people who had participated in all three programmes, and who lived in the least deprived areas agreed to be interviewed, whereas only 0.9% of people who had participated in none of the programmes and lived in the most deprived areas agreed to be interviewed (table 2). The respondents’ age ranged from 51 to 64 years. The respondents’ views varied most commonly by screening participation history (screening participants, CRC-specific non-participants, non-participants), which formed our main comparison category. Comparisons by socioeconomic deprivation did not show clear differences in respondents’ views but are highlighted where differences were found.

The results were organised into the three routes of behaviour change described by the MAP approach: motivational challenges to CRC screening; action-on-motivation challenges to CRC screening; and prompts to CRC screening.17 23 The results are summarised in table 3.

Table 3

Summary of results identifying unique challenges to colorectal cancer screening in comparison to breast and cervical screening

Motivational challenges to CRC screening

Treatment beliefs

CRC-specific non-participants considered CRC treatment to be more severe than breast or cervical cancer treatment.

I know that breast cancer caught early is really treatable. {…} I think cervical cancer’s totally curable. {…} I don’t really know that much about the bowel one. I mean, I know that bowel cancer’s very, very serious. I mean, you can live without your breasts, you know, you can, you know, have a hysterectomy {…} you cannae [cannot] really live without the bowels. (P134, 53 years, CRC-specific non-participant)

Similarly, for non-participants the most threatening aspect of cancer appeared to be its treatment. They questioned the effectiveness of cancer treatments to reduce mortality and expressed concern over side effects, such as hair loss, nausea, fatigue and the quality of life that patients experienced during and after treatment. These respondents questioned whether they would accept treatment if they were diagnosed with cancer. Further, non-participants, mostly those living in the least deprived areas, did not believe early detection could help them avoid cancer treatments that would reduce their quality of life.

You just don’t want to think of bowel cancer, and getting colostomies or whatever. Just the very thought. Sometimes I think I’d rather just not know and die, rather than be diagnosed with that and having a colostomy. {…} I would rather just die than go about like that, that’s not living. (P45, 64 years, non-participant)

In contrast, the screening participants viewed cancer treatment more positively.

If you are unlucky enough and you lose the whole bowel and you have to have a colostomy bag then are you not better having that than dying? (P58, 59 years, screening participant)

Beliefs about the value of CRC screening

CRC-specific non-participants questioned the value of CRC screening. These respondents doubted the efficacy of FOBt after having heard of others who had false-negative FOBt results. They also believed that self-sampling reduced the credibility of CRC screening and completing the test in their own bathrooms seemed incongruent with the respondents’ schemata for medical tests.

If somebody [in the health care team] had said … you know, ‘Come along and sit on the toilet and we’ll collect your poo,’ somehow it would have felt a little bit more detached, a little bit more kind of clinical {…} they’d probably have worked out some sort of system that they can do that more easily without needing plastic bags and God knows what else. (P121, 53 years, CRC-specific non-participant)

Disgust and embarrassment

The only test described as disgusting was FOBt. CRC-specific and non-participants spoke about disgust more often than screening participants. Participants’ disgust related to FOBt being a self-completed test, the involvement of faeces and its association with bowel functions. In each group, participants described CRC screening to be embarrassing, with embarrassment relating to handling their faeces, storing the test kit and concern about other people (such as grandchildren) finding the FOBt kit, or postal workers having to handle the envelopes containing completed kits. Breast and cervical screening was perceived to be more acceptable and easier to discuss than CRC screening.

You would think getting your tits out or, you know, opening your legs for some speculum, you would feel as though both of those things ought to be more intimate [laughs] but actually, I think it’s probably because the way that we are brought up not to discuss bowels in this country. (P95, 55 years, CRC-specific non-participant)

CRC-specific non-participants reported talking about cancer screening less often and specifically avoided talking about CRC screening, which respondents living in more deprived areas reported as off-limits for discussion. Respondents tended to describe talking about FOBt as immature, likening it to children laughing about ‘farts and smells’ (P121, 53 years, CRC-specific non-participant).

You’re told when you’re young [laughs] not tae talk about stuff like that {…} except for when you’re a boy when all bodily functions are, you know, extremely funny in your mind. (P134, 53 years, CRC-specific non-participant)

Symptom beliefs

A theme among CRC-specific and non-participants was that they felt they would be better able to detect CRC symptoms, like indigestion or blood in their faeces than breast or cervical cancer symptoms. Similarly, non-participants also reported that they would ‘know’ if they had cancer making screening unnecessary, particularly in the absence of symptoms.

It [smear test] picks up any issues if you’re not aware of issues {…} I think the bowel cancer one, you kind of know, most people know what the symptoms would be and therefore you assume that if you—since it’s easy to see, that you’re okay {…} ‘cos you’re not aware of any symptoms. (P165, 54 years, CRC-specific non-participant)

Screening participants believed that by the time cancer caused symptoms it would be advanced and less treatable. In contrast to CRC-specific and non-participants, screening participants thought they would have difficulty detecting CRC without FOBt.

Comorbidities and previous colorectal investigations

Some CRC-specific non-participants reported having had colorectal investigations and therefore felt less need for bowel screening. Non-participants also described how psychological and physical comorbidities made them less willing to complete screening. Physical comorbidities meant non-participants felt unable to cope with the prospect of additional investigations and/or treatment. Anxiety prevented some non-participants from deciding for or against cancer screening. A few reported depressive symptoms and thoughts of suicide; in this context cancer seemed to be an ‘easier’ (or, at least, a blameless) way to die—negating any perceived need for cancer screening.

Action-on-motivation challenges to CRC screening

Lack of health professional involvement

CRC-specific non-participants and non-participants expressed worry about completing FOBt incorrectly without the support of a health professional. This set CRC screening apart from breast or cervical screening where health professionals conduct the tests.

I suppose it’s different, it’s like a self-kit, you know? … compared to the other two are being done by professionals. That’s their job and maybe you just feel like they’re doing it right. (P150, 54 years, CRC-specific non-participant)

To complete FOBt correctly, CRC-specific non-participants felt they needed to be fully engaged and ‘pay attention’. In contrast, with breast and cervical screening they could disengage to some extent as the health professional completed these tests for them.

When you go and get a cervical screen you don’t have to do anything, you just turn up. For bowel screening, you’ve got to go that extra step. {…} you’ve actually to make the effort to do it and collect the sample and seal it up and all whatever, and send it away. (P165, 54 years, CRC-specific non-participant)

CRC-specific non-participants described using disengagement during breast or cervical screening to cope with physical or psychological discomfort, but disengagement to overcome disgust or displeasure was not possible with self-completed CRC screening.

CRC-specific non-participants and non-participants reported CRC screening to be burdensome. Having to read instructions was considered to be a ‘hurdle’ (P165, 54 years, CRC-specific non-participant) and FOBt to be complicated and effortful. Having to take three samples added to FOBt seeming burdensome. In contrast, screening participants rarely reported that FOBt was time consuming or complicated, but they agreed that taking faecal samples by themselves was awkward to do.

If somebody was to {…} make it [FOBt] easy for me, I would have done it because I approve of the principle. (P166, 57 years, CRC-specific non-participant)

CRC screening requires planning

CRC-specific and non-participants rarely described making plans to do screening. In contrast, screening participants described detailed planning strategies to overcome practical barriers. They reported dealing with CRC screening invitations promptly and planned specific days to do the FOBt. Screening participants living in the least deprived areas also described routines for cleaning after FOBt completion and using gloves and wipes to make FOBt less disgusting to them.

Impact of comorbidities

Illnesses and other conditions were an additional challenge to the self-completion of FOBt. Non-participants most often reported unpredictable bowel movements, diarrhoea, irritable bowel syndrome, coeliac disease, multiple sclerosis and other health conditions as impediments to CRC screening although such conditions were also mentioned by CRC-specific non-participants. A further two respondents with visual impairments reported abandoning half-completed FOBt kits or waiting for a support worker to organise help with doing FOBt.

I’ve got coeliacs, so, it’s very, very seldom my bowel, my my my toilet is… what’s the word? Solid. {…} so it’s quite difficult that way. So, maybe once I get the coeliacs and everything under control then it might be different. (P130, 54 years, CRC-specific non-participant)

Prompts to CRC screening

Postponing and forgetting

Unlike breast and cervical screening, CRC screening is completed at home and does not require an appointed time. CRC-specific and non-participants reported that CRC screening could be more easily delayed or forgotten than cervical screening which only required them to make an appointment, and even more easily than breast screening where the appointment is prearranged. Some CRC-specific non-participants and non-participants living in the least deprived areas explained that they would put their FOBt invitation to one side and, as a result, forget about it.

If it’s a bowel screening one, yep, put it somewhere and think ‘Yes, I’ll do that’ and then forget about it because it doesn’t have an appointment date. I think if something has an appointment date, you’re forced to act. (P165, 54 years, CRC-specific non-participant)

Screening participants living in the least deprived areas described creating CRC screening reminders: leaving the FOBt material near their bathroom or within their view as a cue.

I just usually take the pack, read the instructions again leave it in the loo until the next time I have to go {…} I just usually take all the bits in and just leave them there to remind me what I’ve to do. (P172, 57 years, screening participant)

Discussion

Our findings show that women who participated in breast and cervical but not CRC screening (CRC-specific non-participants) differed in their barriers to CRC screening compared with screening participants and to a lesser extent, non-participants. CRC-specific non-participants reported that treatment for CRC was more severe than for breast or cervical cancer, and colorectal symptoms were easier to detect oneself than breast or cervical symptoms, which influenced their motivation to complete CRC screening. CRC-specific non-participants also worried about incorrectly completing the test without the support of a health professional, and that they felt the home-based CRC screening test could be more easily delayed or forgotten than breast or cervical screening, which challenged the translation of their motivation into action.16

A key strength of our study was in achieving a sample of women whose screening histories were objectively established by linking three cancer screening programmes’ data for the entire population of Glasgow, Scotland. To our knowledge, this has not previously been done. Among the CRC-specific non-participants and the non-participants the response rate to the invitations to be interviewed was extremely low (0.9%–5.3%) reflecting the difficulty of engaging all population groups in research, and the value of the data that has been obtained. The study has limitations; the SIMD measure used to assess socioeconomic deprivation was an area-based measure which offers a relatively blunt assessment and may offer one explanation for the limited number of socioeconomic deprivation differences noted in the analysis. The study focused on women due to the design, and so it is yet to be determined if the same specific CRC challenges apply to men. It is also noted that this study focused on increasing uptake of CRC and does not consider explicitly the role of informed choice principles. Some women make an informed choice not to engage in cancer screening which is entirely appropriate.24 Our approach does not conflict with the principles of informed choice. For example, providing more information reflecting advances in CRC treatment would increase knowledge.25 However, interventions to address the identified motivational challenges would aim to improve knowledge and understanding and so support informed choice. The action-on-motivation targets would be aimed at supporting people who intend to screen to put their intentions into action and so would not compromise informed choice.

Understanding why CRC screening fails to achieve the uptake rates of breast and cervical has been explored in one previous self-report survey study.15 We are able to expand on the survey’s results, as our findings explain that screening participants also dislike the self-completed FOBt, but manage these feelings; our findings show that medical reasons for non-participation in CRC screening can include comorbidities that impede self-completed CRC screening, but also that women with previous colorectal investigations feel less need for CRC screening. Other studies have considered the relatively low uptake of CRC in isolation, and while they have identified similar barriers to the present study around, for example, fears about treatment, being asymptomatic, concerns about self-completion,8–11 and so on, the present study adds a more nuanced understanding, which informs potential targets for future interventions to increase CRC screening uptake. While it is acknowledged that people are fearful of cancer treatment,11 26 the present study identified that women fear treatments for CRC more than breast or cervical cancer, which may partly explain their reluctance to engage in CRC screening. Similarly, being asymptomatic is a commonly recognised barrier in the screening literature.8 27 Surveys have established that awareness of CRC symptoms is low,28 but it was previously unknown that there may be a misunderstanding that colorectal symptoms are more easily self-detected than breast and cervical symptoms. Furthermore, the present study has illustrated the unique challenges of self-completion of CRC screening in the absence of a health professional, and the greater chance of procrastinating or forgetting the test in the absence of a specified appointment time. We note that similar barriers have been identified for human papillomavirus self-sampling for cervical screening.29 30

We have identified potential targets for interventions to increase CRC screening uptake and drafted example policy recommendations (table 4). For example, to reduce fear and misconceptions of CRC treatments, we recommend a concerted information campaign reflecting advances in CRC treatment and success stories.31 To increase awareness that CRC screening is for people who are asymptomatic, we recommend a concerted information campaign to reiterate and reinforce existing messages that CRC screening is for the asymptomatic, and symptoms may only appear at an advanced stage. To reduce postponing and forgetting, we recommend providing a further reminder and potentially a suggested deadline for kit return. It is important to note that Scotland replaced FOBt bowel screening with faecal immunochemical testing (FIT) in November 2017 and FIT will be introduced in England and Wales in 2019. FIT requires only one faecal sample and, based on evidence from pilot studies32 and the initial months since its introduction in Scotland,33 it is likely this easier to complete test will increase uptake. Nonetheless, FIT alone is not sufficient to address the other identified challenges to CRC screening uptake, and complementary interventions are recommended.

Table 4

Potential targets to increase colorectal cancer screening

This study represents the first step in a process by identifying potential targets to increase CRC screening uptake.23 It will be necessary to test in a randomised controlled trial whether an intervention strategy increases uptake by affecting the specified targets.23 This systematic approach is in line with current best practice guidance on the development of complex interventions,34 35 and will build on and expand existing knowledge of effective strategies to improve cancer screening uptake.36

Acknowledgments

We gratefully acknowledge the patient involvement support from Ann Muir and Tom Haswell.

References

Footnotes

  • Contributors CMC, KR, SM and SW conceived the study and planned it along with MK, EC, CC, DW and RJCS. MK, LG, KR and SM carried out the data collection and analysis. All authors contributed to the writing of the paper. KR is the guarantor of the study. All coauthors had full access to all of the data (including statistical reports and tables) in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis.

  • Funding This research was supported by a National Awareness and Early Diagnosis Initiative Grant (C9227/A17676) awarded to co-PIs KR and CMcC.

  • Disclaimer The funder and sponsor had no involvement in the study design, data collection or analysis, or the writing and publication of the report. All researchers involved in this study were independent of the funder.

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Ethics approval Ethical approval was obtained from the NHS Health Research Authority (NRES Committee North West–Liverpool Central, REC reference: 4/NW/1300). Participants provided informed consent to take part in the study which was performed in accordance with the Declaration of Helsinki.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data availability statement Data are available upon reasonable request.