Process and impact of patient involvement in a systematic review of shared decision making in primary care consultations

Abstract Background Patient and public involvement and engagement (PPIE) in systematic reviews remains uncommon, despite the policy imperative for patient involvement in research. The aim of this study was to investigate the process and impact of collaborating with members of a patient Research User Group (RUG) on a systematic review about shared decision making around prescribing analgesia in primary care consultations. Methods Five members of an established patient RUG collaborated with researchers undertaking a systematic review with narrative synthesis, through workshops held at three time‐points. These addressed the following: designing the protocol, interpreting the results and planning dissemination. Support from a RUG coordinator and user support worker facilitated collaboration throughout the review process. Researchers reflected on how PPIE modified the review at each time‐point. Results RUG members identified factors important in shared decision making around analgesic prescribing additional to those initially proposed by the research team. Search terms and specific outcomes of interest were amended to reflect these additional factors. Thirty of the 39 patient‐identified factors were absent in the published literature. The categories of factors identified were used as a framework for the narrative synthesis and for reporting results. RUG members prioritized options for disseminating the results. Conclusion PPIE collaboration throughout the systematic review impacted on the scope of the review, highlighting gaps in the literature that were important to patients. Impact on interpretation and dissemination of findings ensured the review directly reflected patient priorities. Challenges and strategies to facilitate PPIE involvement in systematic reviews and suggestions for future researchers are highlighted.


Abstract
Background Patient and public involvement and engagement (PPIE) in systematic reviews remains uncommon, despite the policy imperative for patient involvement in research. The aim of this study was to investigate the process and impact of collaborating with members of a patient Research User Group (RUG) on a systematic review about shared decision making around prescribing analgesia in primary care consultations.
Methods Five members of an established patient RUG collaborated with researchers undertaking a systematic review with narrative synthesis, through workshops held at three time-points. These addressed the following: designing the protocol, interpreting the results and planning dissemination. Support from a RUG coordinator and user support worker facilitated collaboration throughout the review process. Researchers reflected on how PPIE modified the review at each time-point.
Results RUG members identified factors important in shared decision making around analgesic prescribing additional to those initially proposed by the research team. Search terms and specific outcomes of interest were amended to reflect these additional factors. Thirty of the 39 patient-identified factors were absent in the published literature. The categories of factors identified were used as a framework for the narrative synthesis and for reporting results. RUG members prioritized options for disseminating the results.
Conclusion PPIE collaboration throughout the systematic review impacted on the scope of the review, highlighting gaps in the literature that were important to patients. Impact on interpretation and dissemination of findings ensured the review directly reflected patient priorities. Challenges and strategies to facilitate PPIE involvement in systematic reviews and suggestions for future researchers are highlighted.

Background
The importance of patient and public involvement and engagement (PPIE) in health-care research is recognized internationally. [1][2][3][4][5] Patient and public involvement has been defined as 'doing research 'with' or 'by' the public, rather than 'to', 'about' or 'for' the public', 6 and patient engagement as 'where information and knowledge about research is provided and disseminated'. 7 In addition to ethical and political arguments for the public having a voice in health-care research, there is evidence that PPIE can impact on research questions, methods, dissemination of findings and engagement with local communities. 4,[8][9][10][11] Suggested positive effects of PPIE include prioritizing research topics, clarifying language in invitation letters, providing a wider perspective in data analysis and designing outcomes more relevant to patients. 8,11,12 In the UK, PPIE is now integral to government-funded health research, and a national advisory group, INVOLVE, 7 provides guidance and examples of best practice for involving patients in health-care research and for reporting patient involvement. 11 Methods of implementing PPIE range from consultation, when researches seek the views of the public on key aspects of the research, to collaboration, an on-going partnership throughout the research process, to publicly led research, in which lay people design and undertake the research. 9 Current policy encourages PPIE using these different methods and within all research study designs. 7,13,14 Systematic reviews 'aim to identify, evaluate and summarize the findings of all relevant individual studies, thereby making the available evidence more accessible to decision makers' (Ref. 15, p5). Systematic reviews are the highest level of secondary research (which re-examines the previously collected data 15 ) investigating the effectiveness of health-care interventions. 16 As systematic reviews inform healthcare policy and guidance, PPIE in reviews is likely to have an important influence on healthcare delivery. Despite this, examples of patient involvement in systematic reviews remain rare and of varying detail and quality. 17 Boote et al. describe a developing consensus of good practice for PPIE in health research [17][18][19] and best practice, and reporting guidelines have been developed. [20][21][22] Our study aimed to integrate PPIE into a systematic review.
The review into which PPIE is integrated explores the factors affecting shared decision making around prescribing analgesia for musculoskeletal pain. Shared decision making (SDM) is the process whereby a health-care professional and patient share information, and the patient is supported to consider management options. 23 Musculoskeletal (MSK) pain includes both pain due to chronic conditions that have a specific diagnostic label such as osteoarthritis, and pain due to symptom labels such as 'low back pain' that may be chronic, episodic or acute in nature. MSK conditions remain one of the leading causes of disability worldwide, 24 and half of those with an MSK condition feel that pain is the worst aspect of their condition. 25 Many people with MSK pain take less than the therapeutic or prescribed dose of their analgesic, 26,27 and some have inadequate analgesia. 28 Shared decision making within consultations may enable patients to be more active in prescribing decisions around their analgesic medications. 29 This research aims to explore a collaborative approach to PPIE, in the context of a systematic review, by recognizing the particular expertise that patients have. 30 Objective To describe the process and impact of involving patients in a systematic review and narrative synthesis of shared decision making around prescribing analgesia for musculoskeletal pain in primary care consultations. members in the systematic review and how the changes made to the review were recorded are described below.

Systematic review method
The systematic review addressed the question 'What factors affect shared decision making around prescribing analgesia for musculoskeletal pain in primary care consultations?' Further details of the methods of the review are below in Box 1, and the full protocol is available from the lead author.

Involving patients in the review process
Five members of an established patient research user group (RUG) collaborated with the researchers in the review process. RUG members of both genders, with a range of ages and musculoskeletal conditions, were recruited. Further information on the RUG and structure of support at the Research Institute has been reported in a case study. 38 A RUG coordinator (RC) and a user support worker (USW) supported RUG members and liaised with the research team. The RUG support team were involved prior to preparation of the study protocol, giving advice on the suitability of the study for the RUG, agreeing aims and design, and providing written training material. The support team suggested having one of the support team and one researcher coordinate activities, as informed by their experience of different ways of working with researchers. The USW (Author AH) therefore acted as a point of contact for RUG members and facilitated meetings throughout the systematic review.
To inform the systematic review, the researchers and RUG support team planned workshops at three key points in the review process: when designing the protocol, interpreting the results and planning the dissemination. The workshops were planned by the researchers and reviewed by the RUG support team. Each three-hour workshop included breaks and used a mixed format of presentation with discussion and small group techniques 39 to ensure all RUG members could make a contribution. All authors agreed that having all three researchers in meetings with five RUG members would Box 1 Details of methods of the systematic review of factors affecting shared decision making around prescribing analgesia for musculoskeletal pain in primary care consultations The search covered four broad topic areas, adapted from existing filters where possible for: primary care, 31 musculoskeletal, 32 shared decision making 33 and prescribing or analgesia. Ten electronic databases were searched with follow-up searches. Key inclusion criteria were as follows: (i) primary care physicians and non-pregnant adults (over 18 years old) with MSK pain; (ii) primary research studies focusing on consultations; (iii) sharing the decision of prescribing analgesia.
Outcomes of interest were factors that did or did not affect shared decision making around prescribing analgesia for musculoskeletal pain. Quality of studies was assessed using existing criteria for quantitative and qualitative studies. 34,35 Studies were not excluded on the basis of quality, but methodological limitations informed the narrative synthesis. Results from full manuscripts were integrated into a narrative synthesis. 36,37 Box 2 Results of the systematic review of factors affecting shared decision making around prescribing analgesia for musculoskeletal pain in primary care consultations Eight studies met the inclusion criteria (see Box 1 for details of inclusion criteria). Key messages suggested by the review were that (i) Patients' requests increase the likelihood of that medication being prescribed, and (ii) GPs attitude to the prescribing of opioid analgesics is controlling, due to legal and ethical concerns around prescribing opioids such as potential addiction. Several other factors were investigated but the quality of the studies and strength of evidence was such that it was not possible to conclude whether those factors had an effect or not on SDM. For example, patient age and relationship with the GP may be important in SDM. Thirty of the 39 factors identified by RUG members were absent in published studies included in the review (see Table S1, Additional Supplementary Information). Included studies did not investigate factors identified by RUG members in two categories of factors: 'Emotion' and 'Condition'. Three factors from eight identified within the category 'Impressions of the GP' were found in the included studies: this was the most well investigated category. likely have been restrictive. Therefore, one researcher (Author CH) worked with the RUG members and support team throughout the systematic review process. The first workshop explored RUG members' understanding of systematic reviews. RUG members were asked to discuss the research question and draw on their own experiences, share stories and describe key factors they thought were important to patients in sharing the decision of prescribing analgesia for musculoskeletal pain in primary care. At the second workshop, CH presented the preliminary results of the search and synthesis. RUG members were invited to critically discuss and plan how to share the results. At the final workshop, RUG members were asked to discuss and agree the final results of the systematic review and plans for further research and dissemination. After each workshop, notes were written up, checked by the USW and distributed to the RUG members with a comment sheet for them to return. RUG members were updated on the impact of their contribution via a newsletter and through discussion at each subsequent workshop. For all workshops, the RUG members were offered reimbursement for their time and travel as recommended as good practice for sustainable patient and public involvement in research. 40,41 Recording the impact of patient involvement To record the impact of PPIE, the researchers documented how the review changed at each stage using a before and after technique. Prior to the workshops, researchers discussed and documented their viewpoints and preliminary decisions. In the workshops, RUG members' perspectives were discussed, and similarities and differences between these and researcher perspectives were debated. Following each workshop, CH liaised with the research team (Authors CCG and KD) to discuss the results and implications of the discussions and how these could affect the review. RUG members were updated about changes to the review and given opportunities to respond.

Results
Involving patients in the systematic review was feasible, and we successfully conducted three workshops with the RUG members. The aim, process and impact of the RUG members' involvement are summarized in Table 1. Impacts resulting from each workshop are described with specific details in the text below. Results of the systematic review itself are described in Box 2. Challenges and strategies for facilitating PPIE in systematic reviews are summarized in Table 2.

Workshop 1: Refining the scope of the review
The RUG members influenced the scope of the review by identifying factors that may affect SDM around prescribing analgesia for MSK pain, additional to those identified by the researchers. The researchers had highlighted factors potentially affecting SDM through preliminary literature searching and their clinical experience (CH and CCG). The researchers categorized these into 'GP' factors, 'Consultation' factors and 'Patient' factors. An example of a factor from the 'GP' category is that a GP's past experience of using a specific analgesic can influence SDM around prescribing that analgesic.
RUG members identified factors important to patients in sharing decisions about prescribing analgesia. They developed categories which were similar to the researchers, with additional categories of 'Medication' and 'Emotion'. Rather than a category of 'GP' factors, patients discussed factors both of their 'Impression of the GP', such as the GP's ability to listen and explain, and their 'Impression of external influences that affect the GP', such as guidelines. The full list of 39 factors in seven categories identified by RUG members is available as supplementary information. The additional factors identified by the RUG members resulted in two changes to the review search strategy. The first change was that search terms for SDM were amended to reflect additional factors such as 'Emotion'. The second change was that the researchers specifically looked for additional factors identified by RUG members in the full-text studies included in the review. The categories of factors identified by RUG members were used as a framework for the narrative synthesis and for reporting results. The patient perspective was integrated into the implications of the review by translating concerns RUG member raised into suggestions for further practice. For example, RUG members' surprise at the lack of studies of SDM with patient populations emphasized gaps in the literature for further study.

Workshop 3: dissemination of findings and reflections on engagement
Initial discussion of dissemination strategies focused on reaching patient audiences to  empower patients when consulting a GP, to suggest medications that the patient had heard about or thought might be useful. RUG members discussed the finding that a patient suggesting medication to a GP increased the chance of that medication being prescribed, which may empower some patients. However, the RUG members raised concerns that patients suggesting medication to a GP may lead to conflict within the consultation. Therefore rather than aiming to disseminate findings with patients, we prioritized dissemination and engagement with GP audiences. RUG members and the support team planned their own roles in engagement including reviewing abstracts, presentations and publications, attending and giving presentations and contributing the patient's perspective to discussions at conferences. Gaps in the literature identified in the review and factors identified by the RUG members informed a subsequent phase of research: a secondary analysis of digitally recorded GP consultations exploring how decisions are shared in practice. PPIE continued into this phase of the research.

Challenges of, and strategies facilitating, PPIE in systematic reviews
A number of challenges of involving PPIE in systematic reviews such as time pressures and managing power dynamics have been identified in a review of case examples. 17 Strategies used by the researchers to manage these commonly identified challenges and suggestions from their experience of integrating current recommendations for good practice around PPIE in a systematic review are described in Table 2. Key strategies for addressing challenges included continuity with a PPIE lead in the PPIE support team and researcher team, and applying for specific funding to support PPIE time and resource requirements.

Discussion
This paper reports PPIE collaboration at three time-points in a systematic review and reflects on the impact on the review. Altering the search strategy to reflect additional factors identified by PPIE members highlighted gaps in the published literature around issues important to patients which would have otherwise gone unreported. PPIE throughout the review ensured that the patient's perspective was integrated into the protocol design, interpretation of the results and planning and prioritizing dissemination.
This study applied previous research on good practices for PPIE in health-care research [18][19][20]22 and in systematic reviews. 17,21 We used case examples described by Boote et al. 17 to anticipate and report common challenges in PPIE. We found that the work by Boote et al. 17 provided a comprehensive framework for the challenges we faced. Placing the patients' perspective alongside the scientific paradigm can increase 'credibility' with a multilayered understanding in research. 43 Concerns that PPIE could be a source of research bias have been described. 9 This study demonstrates how the patients' perspectives both agreed with and differed from the researchers' perspectives. Two approaches towards PPIE in this study address concerns of bias. Firstly, clear description of the review before and after each time-point at which PPIE occurred so that the influence of PPIE was recorded and reported clearly. A second approach is linking aims of PPIE with the process and impact of PPIE so that the intention, application and influence of PPIE in the research is apparent. These transparent approaches to PPIE may prove useful for further enquiry into the impact of PPIE on research. PPIE has been performed in several other systematic reviews in a health-care context, but details of impact of PPIE on systematic reviews are described in separate reports 8,44,45 rather than in the reviews themselves. [46][47][48] It has been suggested that the structure of reviews change so that the process and impact of PPIE are routinely reported. 8 This paper provides an example for reporting PPIE within a systematic review, and the PIRICOM report is a further illustration. 8 Previous research has reported PPIE in dissemination through patient-focused publications and the popular media in addition to traditional conferences. 44,48 Our study is to our knowledge the first to report the PPIE in prioritizing dissemination and engagement activities in a systematic review within health care. Involving patients in prioritizing the dissemination and engagement plan resulted in early identification of challenges in implementing findings from the review with patient audiences.

Strengths and limitations
Strengths and limitations of the study are focused around two issues, the 'representativeness' of PPIE and the extent of PPIE integration within the review. How well those involved in PPIE 'represent' the wider community is a recognized tension. 9,49 Lay experience is 'located within personal history and narrative logic' and contains a 'diversity of world views' [43 p513]. In order to capture this diversity, both who is involved and how they are involved influences the representativeness. Strengths of the PPIE representativeness in this study are that RUG members with different research and condition experiences were involved, supported by a team and working with one researcher. These RUG members could therefore potentially voice a wider range of experiences than individuals who are very well integrated into the research community. Secondly, the researchers used strategies throughout to facilitate sharing the experiences and narratives from all RUG members. Limitations in representativeness are that the RUG members had similar white British ethnicity and were all over the age of 45 years. Different groups such as carers or stakeholder organizations 10 were not involved and including individuals from different populations and roles may have yielded differing factors and priorities. However, the factors RUG members identified were similar in number and scope to those identified in a systematic review of patient reported barriers and facilitators to SDM in health-care consultations. 50 This suggests that the RUG members identified factors that were broadly relevant to patients.
The extent of PPIE integration into the review process also raises strengths and limitations of the study. It is a strength of the study that PPIE occurred at the early stage of question development, and continued throughout the review process. Having one of the PPIE support team and one researcher acting as points of contact is a strength as it facilitated clear communication and clarity of roles. Tokenistic PPIE in healthcare research is a concern, 12 and in this study, we aimed to truly value PPIE and took practical steps to manage challenges of PPIE and facilitate involvement throughout (see Table 2). However, lack of understanding of how PPIE members themselves felt about the extent of their integration into the study is a limitation. Although members did not drop out between workshops, which has been discussed as proxy for PPIE satisfaction, 51  experience of the PPIE members themselves and this highlights a gap in methods and tools for understanding PPIE member experiences.

Recommendations
Future research exploring PPIE should use best practice guidelines and our recommendations when planning, implementing, evaluating and reporting PPIE in systematic reviews. This approach will enable researchers to build on what is already known in this area, anticipate and manage expected challenges, and document the impact of PPIE clearly. It will also facilitate comparison and synthesis of PPIE across reviews. The impact of PPIE reported in this study is on the systematic review itself. Further work remains to be done on evaluating the impact of involvement in reviews on PPIE members themselves.

Conclusions
PPIE collaboration refined the scope of this systematic review and influenced interpretation and dissemination of the findings. Gaps in published literature important to patients were highlighted and the systematic review and its dissemination more clearly reflected patient priorities. PPIE in this systematic review built on existing good practice guidance and examples in design, implementation and reporting. Future researchers may draw on strategies discussed to integrate good practice into their work, including transparent description of PPIE impact and describing how common PPIE challenges are managed. As researchers respond to policy imperatives to integrate PPIE into their research, PPIE in systematic reviews should utilize best practice as described so that patient perspectives and priorities are clearly represented.