Evaluating meaningful impact of Patient and Public Involvement: A Q methodology study among researchers and young people with a chronic condition

Abstract Introduction Although Patient and Public Involvement (PPI) of young people with a chronic condition (YPCC) is receiving increasing attention, evidence of impact is lacking. This is partly due to inadequate understanding of what meaningful impact entails. This study aimed to gain an in‐depth understanding of researchers’ and YPCC's perspectives on meaningful impact. Methods We conducted a Q methodology study in a group of 26 researchers and a group of 20 YPCC with experience in PPI. Participants ranked statements about impact (e.g., ‘YPCC acquire new knowledge and skills’) based on their agreement with them. During interviews, they reflected on their rankings (Q sorts). Factor analysis was conducted to identify similar patterns in the individual Q sorts. The interviews were used to determine and interpret the final factor solution. The resulting factors represented distinct perspectives on meaningful impact. Results Four distinct perspectives on meaningful impact of PPI were identified. Two were predominantly based on the Q sorts of researchers, for example improving research quality and facilitating dialogue and understanding, and two on the Q sorts of YPCC, for example achieving equality and inclusivity and doing justice to YPCC's rights. The factors were defined by 37 Q sorts (80%); 9 Q sorts did not load significantly on any of the factors. Conclusion The results indicate that researchers and YPCC can have different views about the meaningful impact of PPI. The perspectives identified here can serve as an aid when discussing these different views and formulating operational indicators of impact. Patient or Public Contribution An adolescent with a chronic condition was involved in the early phases of this study. She helped in formulating the statements and recruiting YPCC.


| INTRODUCTION
Young people with a chronic condition (YPCC) generally play a passive role in research. They are invited to participate in questionnaires and interviews. There is, however, growing consensus that YPCC should be actively involved in research that concerns them. This is also termed 'Patient and Public Involvement', or PPI. Hart defined PPI of young people as 'the process of sharing decisions which affect one's life and the life of the community in which one lives'. 1 INVOLVE-the former PPI advisory group of the British National Institute for Health Research, now replaced by the Centre for Engagement and Disseminationdevised a general and more practical definition: PPI is about 'research being carried out "with" or "by" members of the public, rather than "to", "about" or "for" them'. 2 Many researchers and YPCC are currently struggling with how to do PPI in research. 3 PPI can take many forms, with YPCC being involved in various ways and in several stages of various types of research. 4,5 The literature on PPI of YPCC provides a plethora of examples, from intensive partnerships with a few YPCC [6][7][8] to consultations with advisory panels. 9,10 The flexibility of PPI makes it possible to adapt it to the research and the people involved. 11 At the same time, it complicates PPI practice, since there is no 'blueprint' for doing it right. 5,12 PPI with YPCC is further complicated by the tendency to view YPCC as vulnerable and inexperienced and to underestimate their competence as decision-makers. 1,3,6,12 This exacerbates the general issue of power imbalances in PPI. 3 In recent years, there has been a rising demand for demonstrating the impact of PPI. The literature on PPI of YPCC shows that impact can take many forms. 4,5,11,13 It is suggested that PPI can increase the relevance and quality of research. YPCC are able to provide new insights from their lived experience, which can improve aspects such as research design, 14,15 data collection 6,14 and dissemination of results. 14,16 PPI is also thought to have a positive impact on the personal development of the YPCC involved, since they learn new skills and gain more selfconfidence. 14,17 Among researchers, PPI can increase their commitment to research 16 and evoke feelings of inspiration and pride. 14 Finally, some research suggests that PPI can enhance the position of all YPCC in society by promoting their right to be heard and supporting inclusivity. 14,18 Demonstrating the impact of PPI of YPCC can help resolve some of the complexities of PPI. 5 It provides insight into the achievements of PPI. In addition, demonstrating impact in relation to specific PPI approaches can help identify the PPI approaches that are most likely to realize these achievements. In the words of Staniszewska and colleagues: 'it requires evidence to inform best practice'. 19 However, best practice can solely be informed by evidence that is robust and of high quality. Researchers in the field of PPI are currently struggling to find ways to conduct robust evaluations of PPI of YPCC. Several literature reviews-published during a time span of 16 years-have shown that the current evidence base on PPI of YPCC and its impact is weak. 4,11,20 In recent decades, limited progress has been made to change this. 4 This limited progress is partly the result of poor understanding of what meaningful impact exactly entails. 21 A previous study reported that researchers and YPCC find it difficult to specify the impact they achieved beyond general descriptions. 22 [23][24][25] In Q methodology, between 40 and 60 participants are presented with a sample of 20-100 statements, which they rank order onto a grid. After this, they are asked to reflect on the choices they made.
Factor analysis is then conducted to reveal patterns of similarity in how statements were sorted by respondents. Unlike conventional factor analysis, in Q methodology, the individual rankings-not the different statements-are taken as variables. The resulting factors represent groups of individuals with similar perspectives. Participants' reflections on their rankings are used to interpret and describe the factors.
We chose Q methodology because it was very suitable to answer our research question for several reasons. First, it can be used to systematically study perspectives and compare their similarities and differences. 24,25 Second, it also prompts participants to carefully weigh the importance of various statements, as the method compels them to make a choice. 24,26 Third, it combines the strengths of quantitative and qualitative research. The factors retrieved in quantitative analyses are given meaning by using participants' reflections on their sorting of statements. 24,25

| Participants
The participants in the study were researchers and YPCC between 15 and 30 years old. Participants were required to have at least some experience with PPI of YPCC. Researchers were recruited through the authors' networks in the academic field. YPCC were recruited through the network of an adolescent with a chronic condition who was involved in the preparation of this study and an announcement placed on the website of JongPIT, a Dutch foundation for and led by YPCC. Snowballing was used to contact additional researchers and YPCC.
After recruitment, 50 participants sorted statements and were interviewed. However, four participants were excluded from the analysis.
During data collection, it became clear that one researcher and one YPCC VAN

| Data collection
Statements were collected about various types of impact, such us 'Young people acquire new knowledge and skills' and 'PPI contributes to a society in which everyone can participate'. A recently conducted literature review was used as the primary input to formulate statements 4 by scanning the data on reported motivations and benefits of PPI. This was performed by the first author. During this process, she merged some statements that were very similar. This resulted in a list of 39 statements, which was discussed by all authors and the adolescent with a chronic condition who was involved in the preparation of this study. In an iterative process, it was decided to remove some additional statements to eliminate repetition. In addition, statements were clarified and shortened to improve comprehensibility. The final selection (also called the Q sample) consisted of 33 statements (Table 2).
Participants were asked to rank order the Q sample using a Q sort table (Figure 1). Due to the COVID-19 circumstances, they were invited to do this digitally. Participants were sent a link to the ranking exercise, which was programmed using the VQMethod. 27 The ranking exercise started with an overview of all 33 statements. First, they were asked to sort the statements (options: agree, neutral, disagree) guided by the question 'What are your motivations for doing PPI?' The second step was to rank all statements in the Q sort table. Based on the first sorting exercise, participants were instructed to rank the statements based on how much they agreed with them. The more they agreed with a statement, the more they placed it to the right of the Q sort table. The more they disagreed with a statement, the more they placed it to the left. The statements they agreed or disagreed with most were placed at the extreme right (+4) and left (−4). After finishing the Q sort table, participants were asked to answer a few short questions about their background and previous experience with PPI of YPCC ('little', 'some' or 'much').
After the ranking exercise, telephone interviews or video calls were conducted with participants to reflect on their individual ranking of statements. Participants answered questions about the two or three statements that they placed at the extreme right (+3 and +4) and left (−3 and −4) sides. Some additional questions about successful PPI and impact were included to ensure that all relevant topics regarding meaningful impact were addressed (Box 1). All interviews lasted a maximum of 45 min. They were audio-taped and transcribed.

| Analysis and interpretation
The individual rankings of statements-also termed Q sorts-were analysed using principal component analysis, followed by Varimax rotation.
Analyses were conducted in Stata 15.0 using qfactor. 28 The selection of a factor solution was based on both quantitative and qualitative criteria. 23,29 Some statistical features were examined. For example, only factors with an eigenvalue in excess of 1.00 were selected. Another requirement was that at least two Q sorts must load onto each factor.
Ideally, the percentage of explained variance in the chosen factor solution is 35%-40% or higher. 29 In our analyses, factor solutions with 2-7 factors fulfilled these requirements.
Since factors should be interpretable and represent coherent and comprehensible narratives, the final factor solution was based on the qualitative interviews. 29 As a part of Q methodology, the views expressed during the interviews were compared with the idealized Q sort of each factor, which are called the composite sorts. In a composite sort, 'the Q sorts of all participants who define a given factor are merged together to yield a single (factor exemplifying) Q sort'. 23 The choice for the final factor solution was based on the extent to which composite sorts were consistent with the qualitative reflections of the participants whose Q sorts defined each factor.
Interviews were analysed in MaxQDA. The statements of the ranking exercise were translated into codes. All fragments reflecting on specific statements were coded within the corresponding code.
This resulted in an overview of how participants interpreted different statements and how they reflected on them. The principal author

| Validity
All participants provided informed consent before they started the ranking exercise. The study was conducted in accordance with the General Data Protection Regulation. Formal ethical approval of this study was not required under the prevailing Dutch legislation. Table 1 presents a summary of the characteristics of the study participants. On average, YPCC were 23 years old (range 17-29 years).

| Description of participants
Their experience with PPI varied from little or some (50%) to much (50%). The researchers who participated were on average 39 years (range 22-63) of age. Their experience in doing research varied, with interns, PhD students, postdocs, senior researchers and professors taking part in the study. They had little or some (62%) or much (38%) experience with PPI of YPCC.

| Perspectives on meaningful impact
Four factors and composite sorts were extracted from the data, that is four distinct perspectives on meaningful impact of involving YPCC in research. These factors were defined by 37 Q sorts (80%); 9 Q sorts did not load significantly onto any of the factors. Table 2

| Factor interpretation
Factor D focuses on YPCC's right to make their opinions heard. YPCC have the right to be involved in research that concerns them (25: +4). it is above all a right, and the focus should be on providing YPCC with sufficient space to make their opinions heard rather than improving the research as such.

| DISCUSSION
In the present study, we examined the perspectives of researchers and YPCC on meaningful impact of PPI in research. The resulting insights offer valuable opportunities for improving the evaluation of PPI of YPCC. This enables advancement in a field that has shown limited progress in recent decades.
Using Q methodology, four distinct perspectives on meaningful Conducting research that will improve the lives of YPCC is considered the most important type of impact, from this perspective. In perspective B, the highest value is placed on improved understanding of research outcomes and formulating correct conclusions and implications. One similarity between perspectives A and B is the emphasis on improving the usefulness of research. In perspective C, the most important type of impact is a more inclusive society. In contrast to perspectives A and B, the personal development of the YPCC involved in research is also an aim. Perspective D prioritizes the right of YPCC to be involved above all other impacts. We also noticed during the interviews that questions about impact were answered with caution and many qualifications. Researchers in particular feared that impact would become a condition for doing PPI. We would suggest, however, to look at impact evaluations as a learning mechanism and a way to obtain insight into the best ways to do PPI. 5 According to Lundy's line of reasoning, imperfect PPI is not the end but rather the beginning of a learning process. 32  The adolescent provided input on the statements for the sorting task.
She suggested that we add 'YPCC feel heard'. This appeared to be a relevant addition, since many study participants rated this statement as important. Based on the suggestions of the experience expert, we also reformulated statements, such as 'YPCC are given the opportunity to do something useful for others', which was changed to 'YPCC are given the opportunity to do something useful for other YPCC'. The adolescent also helped in recruiting YPCC as study participants via her network.
To make sure that the perspectives of YPCC were also included in the final stages of the study, we informally discussed our findings with several other adolescents with a chronic condition who were involved in another research project.

| Strengths and limitations
To the best of our knowledge, this is the first study that has examined the impact of PPI in so much detail. Q methodology enabled us to systematically study perspectives and compare their similarities and differences. An important strength of this method was that it compelled participants to choose which impact matters to them most.
This ensured that they carefully considered their motivations in a way they had not done before.
Due to COVID-19 circumstances, we had to conduct the study online. Although the programme VQMethod 27 was very suitable for doing the sorting task online, one limitation was that participants could not complete the sorting tasks in the presence of the researchers. This resulted in the exclusion of a few participants who did not fully understand the sorting task. Also, it would have been valuable to hear their thoughts and considerations during the sorting tasks rather than retrospectively during the interviews. Another limitation is that recruitment of participants through the networks of the authors may have resulted in some selection bias. However, by applying snowballing, it was ensured that researchers and YPCC outside the networks were also reached. Finally, it could have been valuable to discuss the sorting tasks in focus groups rather than interviews to enrich our understanding of overlap and differences between arguments for doing PPI. The interviews, however, enabled us to hold more in-depth discussions with all participants about what motivated them.
The current study provides extensive insight into perspectives on meaningful impact of researchers and YPCC. It would be interesting for future studies to explore these perspectives in different age groups, such as children, adults and the elderly, since it is conceivable that desired impact may differ according to age.

| CONCLUSIONS
In recent decades, limited progress has been made in evaluating the impact of PPI with YPCC. In the study presented here, we aimed to change this by clarifying the concept of meaningful impact. Using Q methodology, we identified four distinct perspectives on meaningful impact among researchers and YPCC: improving research quality, facilitating dialogue and understanding, achieving equality and inclusivity and doing justice to YPCC's rights. Researchers and YPCC generally have different perspectives on meaningful impact. Evaluations should therefore take into account the perspectives of everyone involved in a PPI process. Our study also highlights the importance of predetermining indicators for meaningful impact in PPI evaluations.
The perspectives retrieved in this study can serve as a starting point for this and for operationalizing them into measurable indicators.

ACKNOWLEDGEMENT
We would like to thank Claire Aussems for her advice on the quantitative analyses.

CONFLICT OF INTERESTS
The authors declare that there are no conflict of interests.

ETHICS STATEMENT
All participants provided informed consent. The study was conducted in accordance with the General Data Protection Regulation. Further ethical approval of this study was not required under the applicable Dutch legislation.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.