Patient and public involvement in health-care research is increasingly required by funding bodies. It was also an important part of the study, and so it is discussed in depth in this penultimate chapter. The chapter describes the effectiveness of PPI in the development of the technology-based complex intervention, a topic that might seem to be especially daunting to lay involvement. The challenges and issues, as well as the successes and the adoption of values and principles core to PPI work, are described and reflected on.
Introduction
Patient and public involvement is recognised as being important to health and social care research, in enabling researchers to take into account and incorporate the needs, wishes and experiences of those people whose health and well-being their research is intended to benefit. Brett et al.,166 for example, in a systematic review, cite the advantages as ranging from more appropriate research objectives to more effective study recruitment and dissemination of findings. Staniszewska et al.167 specify the benefits of improved quality, relevance and appropriateness, and the broader democratisation of research. As a consequence, PPI increases the likelihood of the effective translation of research findings and outputs into practice and their sustained use and benefit to patients and the public, as well as enhanced transparency, accountability and public ownership of the research. These are important factors in NIHR-funded research, which is supported by public money, and so NIHR, along with other UK funding bodies and research ethics committees, actively encourages PPI.168 Significantly, in 1996, NIHR established INVOLVE as a national AG for PPI in health-care research.
Patient and public involvement was particularly important to the study, as this aimed to transform patient experience survey free text into a form that could drive health-care improvements through automated computer analyses. Including PPI throughout ensured that the study remained true to the purposes of the data, did not distort the patient voice and produced results that would improve the service experience per se.
Two projects funded by NIHR underpinned the consideration and the use of PPI in the work:
the ReseArch with Patient and Public invOlvement: a RealisT evaluation (RAPPORT) study,
169 which resulted in an inductively and deductively derived list of the features of good PPI, using NPT to structure the analysis, just as NPT runs through this study
the framework of values and principles developed by INVOLVE.
168
The study team also followed the recommendations from Staniszewska et al.,170 and the GRIPP (Guidance for Reporting Involvement of Patients and Public) checklist171 in this report.
Type of involvement
Tripp172 developed a typology of possible roles for PPI representatives, lying on a continuum from researcher in control, through shared control, to users in control. This makes plain the potential for PPI representatives to be involved in a range of activities from consenting, consulting and co-operation through to collaboration and collective action. The PPI in this study fits the consulting and collaboration roles and occurred at all stages of the research. It centred around group discussions of items supplied by the research team, preceded by updates on the progress and details of the study. On members’ request, a mailing group was established to enable communication outside the meetings.
Summary of involvement opportunities
Carers and patients were key partners in the research. Patients had cancer or had undergone major surgery for a condition different from cancer. This mix reflected both the study’s focus on a cancer survey, and the study’s commitment to making the approach transferable to other condition-specific and non-condition-specific surveys. The study team specified surgical non-cancer patients because the CPES includes comments across primary and secondary care, and this was a simple way of ensuring the same spectrum of experience across all members.
A PPI member was included as a co-investigator on the grant proposal, and two other members named in the proposal as lay PPI representation with roles on the SSC and AG. These three PPI representatives, who were all patients with cancer, helped to shape the research proposal and protocol through informal meetings and correspondence and consultation with other patients.
The study team also established a separate PPI reference group, referred to as the PPI Research Group (PPIRG), which emphasised its active role. Team members were in e-mail contact with members of the PPIRG throughout the early stages of the study. The PPIRG subsequently met approximately every 3 months over the last 10 months of the study; thus, there were four meetings in all. SSC and AG meetings always included a PPI agenda item.
The study team had intended to make use of the NIHR Research Design Service ‘Building Research Partnerships’ scheme to train PPIRG members, but the Research Design Service did not run a training event during the relevant time period. Macmillan Cancer Support had also stopped running their similar training programme, but all PPIRG members were invited to use its course content, which is available on its LearnZone web page [http://learnzone.org.uk/ (accessed 29 January 2019)] after free registration, and Macmillan Cancer Support also offered the use of experienced facilitators to optimise the use of this training. In the event, no PPIRG member wished to take up training; most already had some experience of PPI or similar tasks from work on other research projects or charity involvement.
At the final meeting, stakeholders were asked whether or not they wanted to continue being informed about the study. They stated that the study team should hold a feedback event 1 year later, so that they could determine the impact of their involvement.
Who was involved?
The study chief investigator attended all but one of the PPIRG meetings, which the lead programmer attended instead. The lead researcher attended all meetings and took notes, which she subsequently distributed to all PPIRG members for checking, amendment and finalisation. Only two research team members were included at meetings to ensure that PPIRG members were in the majority.
In addition to the research team members, the first meeting was attended by five patients, the second was attended by six patients, the third was attended by three patients and the fourth was attended by two patients. Meetings 1 and 3 took place at the University of Southampton and meetings 2 and 4 took place in central London.
Setting up the group and first meeting
The PPI co-applicant volunteered to recruit group members once the study was funded. However, because of time constraints, geographical distance and ongoing obligations to another study, this patient soon decided to withdraw his involvement. As a consequence, the recruitment strategy had to be revised. This also delayed the start of the PPIRG’s work. As he was the key PPIRG co-ordinator, for practical reasons, he had to be replaced in this role by the lead researcher. The lead researcher turned to the co-applicants, clinical networks and members of the advisory and steering groups that had already been set up, and asked for their support in reaching patients. A call was published on Macmillan’s Volunteer Village and Task Force Groups, as well as in the South East Coast Strategic Clinical Network Bulletin. Twelve patients responded, expressing their interest in joining the PPIRG. At this time, one of the steering group PPI representatives pulled out for health reasons and could not be replaced.
A Doodle poll (Doodle, Tamedia, Switzerland) was used to fix a suitable date for the first PPIRG meeting, which took place later than intended, 7 months into the project. Nonetheless, during this time, the three PPI representatives named on the proposal provided bridging input. This delay was only partly attributable to the recruitment issue and included the taking account of members’ summer holidays; some members were committed well before the first official PPIRG meeting and were invited to the study launch 3 months beforehand.
The PPIRG’s terms of involvement were agreed on at the first meeting, typed up by the research team and circulated among all members of the PPIRG for review. Further changes were incorporated.
To maximise attendance, dates were always agreed on using a Doodle poll. PPIRG member feedback on arrangements led us to alternate meetings between the University of Southampton and central London. This was agreed by members as being likely to give everyone the opportunity to attend at least half of the meetings. Skype™ (Microsoft Corporation, Redmond, WA, USA)/telephone contributions were also enabled throughout. Each meeting began with a summary of the study aims and a full update of progress from inception to ensure that all members were fully informed, regardless of previous participation. In all meetings, presentations included examples of the developing text-analytics and toolkit work, including interactive demonstrations and encouraging the input of ideas and comments.
All expenses incurred by the PPIRG members in attending the meetings were reimbursed; these were usually travel expenses and refreshments en route, but in one case it was agreed to pay for hotel accommodation. Lunch and refreshments were always provided.
Throughout the study, the team adhered to INVOLVE 2015 values.
Launch, infographic, newsletters, blogs and social media use
A talk by a cancer service user was a highlight of the study launch event and the final event, and the study team included informal workshops designed to help lay people to understand the more complex aspects of the study. This included an art workshop, a shared decision-making session in which teams were asked to design a PPI meal event and a hands-on workshop on what made for good infographics.
Prior to the first PPIRG meeting, the research team collaborated with a graphic designer to develop the study’s ‘storyline’ – an infographic resource to be used by PPIRG members, displayed during public dissemination events and placed on the website to explain the different stages of the study in an accessible way. This was circulated among the PPIRG members and discussed in the first meeting. The group emphasised the need for a clearer orientation. It was also argued that the images were trying to convey too much information at once and did not clearly indicate what was at the heart of the study or how it worked as a whole. The members also suggested a number of changes to the wording and visual improvements to particular elements of the images in order to increase clarity. It was also suggested that the word ‘toolkit’ may not be familiar to patients and should therefore be replaced by ‘website’. The feedback was shared with the designer, who worked on incorporating it into the storyline. This discussion helped to develop shared understandings between the researchers and the PPIRG members and was considered to be very rewarding.
To update all of the people involved or interested in the study about its progress, a PRESENT study Twitter account was created. It was used to share news about the research, publicise events, such as the study launch day, report on relevant conferences and seminars that the study team attended and link the study to other research in the area of health technology innovation or patient health-care experience. A Facebook (Facebook, Inc., Menlo Park, CA, USA; www.facebook.com) profile was created but abandoned, as it did not get any followers. The study team considered writing monthly newsletters, but opted instead for blogs about significant events, such as the participation in a group concept-mapping workshop in Sweden, the progress of the study’s workshops or talks and presentations given. It was felt that the combination of a blog and tweets would be a more rapid, engaging and manageable strategy than monthly newsletters, and blogs could still be printed out.
Impact of patient and public involvement
Two PPI members helped us to develop the study design and proposal. They:
Suggested improvements to the lay abstract, including SMOG (Simple Measure of Gobbledygook) testing.
Improved the study focus, which was initially too technology heavy; this had the potential to negatively affect the involvement of patients in the study itself and reduce the quality of outputs.
Simplified the study design; originally, group workshops were planned to process the outputs from the rule-based IR analysis using mindmapping scenarios and then more workshops were planned to undertake a separate concept-mapping task. The study team realised from PPI input that the tasks could be simplified and combined in one large workshop. This proved to be successful.
Strongly supported the plans to involve patients in the design of GATE rules because of their concerns that computer replacement of manual analysis might be detrimental, removing the patient voice. The validation stage was thus developed in accordance with a manual gold standard, partly to address their concerns.
Suggested that rule ideas (see
Chapter 3) should be developed remotely through surveys (the study team had originally planned to hold a workshop). This had the further advantage, subsequently realised, of making the process easier to adopt by other groups wishing to use the study’s model process for other topics.
Within the PPIRG meetings, the group made several important contributions, including making the research accessible to the public, recruiting study participants, keeping the study’s focus on the patient experience and collaborating on the initial development of the toolkit prototype for discussion in phase 2 of the study. They also helped us to make sense of the research findings.
During the first meeting, the overall purpose of the toolkit was discussed, along with early ideas for its features. The group emphasised the importance of using simple language and diagrams, employing illustrations and large fonts on highly contrasting backgrounds to ensure that the toolkit was visually appealing and accessible, including for people with English as their second language. Suggestions were made with regard to themes (such as emotional support) that should be included in the toolkit and useful filtering options, such as sorting the comments by gender, age and the ethnicity of the author. The group also stressed the need to include an up-to-date list of websites and charities offering information and support. A few of the members expressed that they would like to be able to input their own feedback to contribute to the comments retrieved from the surveys, in an interactive forum, and to be able to track their own CPES comments to ensure that the rule-based IR system classified them correctly. This was ultimately decided to be outside the scope of the study, as the group agreed that it would require active facilitation for which the study team did not have funding.
During the second meeting, suggestions were made about introducing a word-search box to extract comments and changing the colours on the RAG-style dials. An initial design for the toolkit theme icons was presented and the team was advised that alternatives should be presented for stage 2 research participants to consider. This early PPIRG feedback was extremely useful for preparing the materials and topic guides for stage 2, in which user preferences would be explored via group concept-mapping workshops and associated interviews and surveys.
The group played an important role during the recruitment for stage 2. This included revising the flyer used for concept-mapping workshop recruitment to make it more engaging for patients and circulating the flyer among their networks. The study team also received useful comments on how to run the workshops. As a consequence, the number of comments the participants were to sort and rate in these workshops was reduced from 80 to 60, as the former was perceived to be too cognitively demanding.
Issues and challenges, with Patient and Public Involvement Research Group feedback and research team reflection
The withdrawal of the co-applicant patient and the subsequent need for a new recruitment strategy led to considerable delay in the setting up of the group. One of the participants withdrew before the first meeting as a result of ill health. There was a core group of seven patients who were very active and contributed to most of the meetings. Of the remaining five, one could not get involved until the last 5 months for personal reasons. The rest made no contact or contributions beyond their initial agreement to sit on the PPIRG, despite receiving meetings invitations, agendas and minutes. One issue was around compensation for time – all members were offered the INVOLVE recommended rate of £125 per day. For some potential PPI contributors, this was too big a stumbling block, as it would affect their disability allowance or other benefits. On the other hand, had the study team developed a suitable alternative and emphasised this to potential contributors, engagement might have been better.
The SSC and AGs were instrumental in facilitating recruitment for the PPIRG. Members of both groups circulated the call among their networks. Macmillan Cancer Support published the call on the Volunteer Village and forwarded the information to the leaders of its task forces and strategic lead. The AG offered useful advice on how to run the PPIRG meetings effectively. The group recommended that the materials shared with the PPIRG were rich in engaging visuals and clear, jargon-free language. It also asked the study team to carefully consider how the PPIRG might contribute to the study and to ensure that the tasks and questions for the group were set out clearly.
The majority of the members were already sitting on various other PPI or charity groups, and were thus a select group. The two members who were not affiliated with other groups were incidentally the ones who withdrew for health-related reasons and a change of circumstances. One of these patients doubted that she would be able to make a valuable contribution, given her lack of experience in PPI. She was reassured by the research team, but lack of confidence could have been a potential barrier in the recruitment of patients not already involved in PPI. Nevertheless, the feedback from the group members was that the group included people with a wealth of different experiences, knowledge and backgrounds.
It was found that some of the members had priorities based on their involvement with other organisations, and on their personal, sometimes upsetting, experiences as patients. They would sometimes pose expectations that lay outside the scope of the project. The group felt let down by the fact that patients will not have access to the original free-text comments on the toolkit. The members felt strongly that the public must have access to the feedback patients offer in PESs and made recommendations that appropriate consent procedures should be put into place to enable this.
Practical challenges were experienced in ensuring inclusion (explaining jargon, talk time in meetings and help with technical aspects). One of the members pointed out that some operational details of the toolkit became clear to him only at a late stage. Although the team members always tried to ensure that they explained the study in lay language, using visual materials to support the discussions, this feedback highlighted that the team members could have improved on this by undertaking a more adequate review of PPI members’ understanding of the study earlier on. It was sometimes easy to forget that PPI representatives were not colleagues but volunteers.
The overall feedback that the study team received verbally, on feedback forms and as solicited e-mail feedback, was very positive. The members felt that the meetings were well organised, with the materials circulated well in advance. Most importantly, they also pointed out that they felt that their contribution was always valued and acted on, as the group saw changes that they suggested being incorporated into new materials from meeting to meeting.
Discussion
Given the importance of co-design and the development of the toolkit with its users at the centre, it was felt that it was key that the research was understandable to the wider public. The PPIRG was imperative in helping us to achieve this aim by supporting us in the preparation of the public-facing materials and in the preparation for the group concept-mapping workshops. The group also played a key role in the recruitment of both patients and professionals for the study’s workshops and interviews, and designing the toolkit in such a way that the patient voice remained. Overall, it was found that PPI was useful in the development of this technology-based complex intervention. The PPI members also evinced appreciation of the efforts to determine their needs and satisfy these, the information provision and the transparency with which their contributions and inputs were used.
Engagement and contributions were good, and the study team followed the core values and principles of good PPI work. More could have been done, however, as some patients who expressed an interest failed to engage. The original facilitator, who had to drop out, might have helped us to resolve some of the challenges ().
Summary of the study’s PPI experiences
