Co‐construction of health technology assessment recommendations with patients: An example with cardiac defibrillator replacement

Abstract Context The National Institute of Excellence in Health and Social Services (INESSS), which functions as the Québec health technology assessment (HTA) agency, tested a new way to engage patients along with health‐care professionals in the co‐construction of recommendations regarding implantable cardioverter‐defibrillator replacement. Objective The objective of this article was to describe the process of co‐construction of recommendations and to propose methods of building best practices for patient involvement (PI) in HTA. Design Throughout the process, documents were collected and participant observations were made. Individual interviews were conducted with patients, health‐care professionals and the INESSS scientific team, from January to March 2018. Results Three committees were established: an expert patient committee to reflect on patient experience literature; an expert health professional committee to reflect on medical literature; and a co‐construction committee through which both patients and health‐care professionals contributed to develop the recommendations. The expert patients validated and contextualized a literature review produced by the scientific team. This allowed the scientists to consider aspects related to the patient experience and to integrate the feedback from patients into HTA recommendations. The most important factor contributing to a positive PI experience was the structured methodology for selecting patient participants, and a key factor that inhibited the process was a lack of training in PI on the part of the scientific team. Conclusions This experience demonstrates that it is possible to co‐construct recommendations, even for technically complex HTA subjects, through a more democratic process than usual which led to more patient‐focused guidance.


| INTRODUC TI ON
Patient involvement (PI) is gaining prominence in the area of health technology assessment (HTA). Policymakers, health-care managers and professionals, and researchers are increasingly interested in users' experiences regarding the use of medical devices. In the field of research and evaluation, PI refers to doing research 'with patients' as opposed to 'about patients'. 1 User's experiential knowledge has recently been included in HTA, under the assumption that consideration of such information may lead to more relevant, humanistic and comprehensive consideration of the impact of new technologies on quality of life. 2,3 Scholars agree that PI in HTA should be undertaken on a case-by-case basis. 4 Different frameworks have incorporated PI, ranging from a passive role, such as receiving information during medical encounters, to being actively involved in the co-design and co-construction of healthcare-related issues and guidance. 5,6 The latter considers the patient's experiential knowledge as an invaluable complement to scientific and academic expertise; it relies on the idea that patients and health-care professionals can and should collaborate in the co-construction of medical guidelines and recommendations (Table 1). 6,7 Despite a general trend towards increased PI in HTA agencies that are members of the International Network of Agencies for Health Technology Assessment (INAHTA), PI remains limited in scope and highly variable in practice. 8 In fact, various attempts to define models to better include patients in HTA processes have been made by HTA agencies worldwide. For example, the National Institute for Health and Care Excellence (NICE) in the United Kingdom promotes fairness and inclusion of patients in decisions related to their health and well-being in its explicit public commitment policy. 4 For this reason, NICE prioritizes HTA that includes patients with (a) experience of having a given condition and receiving care for it, (b) perception of the impact of the technology and (c) expectations about the technology's risks and benefits. 9 A recent study in HTA research, designed to follow the PI standards described by NICE, found that while PI was overall a positive experience, patients valued their involvement at early stages of health technology development more, because they perceived that their contribution was greater. In the face of this evidence, the authors speculated that as patients are experts in their own illnesses, their contribution to the early development stage was key to better understanding patient needs and producing devices that responded to these needs. 10 Another study assessing NICE's processes in incorporating the views of patients in HTA decision making found that although the organization has attempted to be flexible in integrating patient views, the patient's role is still confined to representation, rather than decision making. 11 The Canadian Agency for Drugs and Technologies in Health (CADTH) also involves patients, on working groups and committees at several levels, in the assessment process for medications and medical devices, through the use of questionnaires. 12 For example, the Pan-Canadian Oncology Drug Review, an evidence-based cancer medication assessment that is part of the programmes and services led by CADTH, offers patients the chance to share their experience as participants in clinical trials 13,14 and to describe needs unmet by current therapies. 12 Likewise, elsewhere in Canada, Health Quality Ontario created the Ontario Health Technology Assessment Committee to provide guidance and advice on how to include PI in their HTA activities. This led to recommendations fostering PI in different forms. Moreover, a framework depicting PI in HTA was developed to support these initiatives. 15 Although this framework is broad and all-encompassing, it does not provide guidance regarding the selection of patients nor their means of participation in the different steps of the HTA process. [15][16][17] Since 2010, INESSS has had an important role in the delivery of health care and social services to Quebeckers. 18 The mission of INESSS is to promote clinical excellence by optimizing the use of resources when considering the incorporation and utilization of devices, medications and interventions. Thus, health-related innovations entering the market are initially assessed; based on the results, INESSS issues recommendations concerning their utilization and implementation, as well as potential reimbursement of healthcare costs to users. Until the beginning of the present project, INESSS assessed products and services by consulting health-care professionals, administrative managers and decision makers (usually recruiting the latter following consultation 19 ); no patients were included in the co-construction process of guidance, although a few evaluations were completed that included patient consultation. 20 Moving forward, INESSS has adopted a 2016-20 strategic plan looking to enhance the participation of knowledgeable users. 21 Following the adoption of the 2016-20 strategy 21 and the 2016-19 triennial activity plan, 22 INESSS decided to include patients, along with health-care professionals, in the co-construction of recommendations concerning devices, medications and interventions. 6 In the present manuscript, we describe the process of co-construction of recommendations concerning the replacement of implantable cardiac defibrillators (ICDs) from the perspectives of both patients and health-care professionals. ICDs are indicated for arrhythmia which is a cardiac disorder characterized by irregular or abnormally rapid or slow heartbeats. ICDs are devices that are placed under the skin to monitor heart rhythm and to intervene with electrical stimulation or shocks, as necessary, if arrhythmia occurs. Ventricular tachycardia and ventricular fibrillation are 'malignant' arrhythmias that can lead to sudden cardiac death, 23 particularly in the presence of low left ventricular ejection fraction (defined as <40%) 24,25 ; for this reason, left ventricular dysfunction is a major criterion for the implantation, and continuing use, of ICDs. At present, ICDs require a replacement of their battery every 5-7 years. During this period, the patient's clinical condition and treatment preferences may change. Moreover, ICDs can have significant impact on the daily lives of the people who wear them. On the one hand, they can prevent sudden cardiac death and provide a sense of security to the person, but on the other hand, shocks can be distressing and unexpected. It is therefore important, as with other implanted devices, that the patient or his/her representative be involved in the decision-making process regarding initial implantation of an ICD and its replacement.
Following the publication of the ICD replacement guidance by INESSS, 26 the aims of the present manuscript were to contribute to the discussion of PI best practices in HTA, by presenting the co-construction methodology used and the learning experience of those involved in the process. A final objective is to propose a methodology for PI.

| Aims of data collection
The aims of the data collection carried out for the present project were firstly to continuously improve the co-construction methodology process while it took place and secondly to assess the added value of PI to HTA, [27][28][29] as appreciated by the various participants involved. Using a formative evaluation approach, we qualitatively assessed the feasibility and acceptability of our process, during which patients and health-care professionals actively collaborated. 30,31 We combined data collection, logical analysis and implementation of changes 31 in this project.

| Informants
Four types of participants were queried: the patients (n = 8) and the health-care professionals (n = 11) who were part of the co-construction process, INESSS scientists (n = 4) and INESSS specialists in PI and ethics (n = 3). Prior to participating on the various committees, all persons external to INESSS received information about the purpose and rationale of the evaluation of ICD replacement.

| Data collection
We followed the procedures of data collection for a formative evaluation by Brouselle and colleagues. 31 Data were collected throughout the evaluative process and consisted of the following:

| Document collection
All documents produced during the project (eg minutes of meetings documents prepared for meetings, questionnaires and feedback forms from both patients and health-care professionals) were collected to assist with assessing the co-construction process.

| Participant observation
The first author (MPP) attended all meetings with the INESSS scientists and led all the committee meetings, as well as the co-construction meeting, during which all experts (patients and health-care professionals) were gathered. After each committee meeting, the first author (MPP) completed a logbook specifically designed for recording observations on the patients' and health-care professionals' involvement. In total, 8 hours of meetings were held.

| Semi-structured individual face-toface interviews
In total, 23 interviews were conducted: 12 with patients (soon after their participation in both the patient (n = 6) and the co-construction committee meetings (n = 6)), two with health-care professionals and nine with INESSS staff members. All interviews were conducted by professionals trained in evaluative research (MPP and IG), from January to March 2018. Interviews varied in length from 30 to 90 minutes and were all digitally recorded; the topics discussed are presented in Appendix S1.

| Data analysis
Each interview was transcribed by one member of the research team (MPP). Transcripts were imported into QDA miner software 32 for coding purposes. Data analysis was carried out concurrently with data collection to allow the integration of information from each step of the process. We used the framework approach to analyse data, 33 a strategy frequently used in the context of policymaking.
This strategy employs five analytic stages, namely: (1) familiarization with the data through reading; (2) identification of a thematic framework that reflects the ideas discussed; (3) indexing data, that is identifying patterns across the transcripts; (4) charting data, that is comparing data across identified patterns; and (5) mapping and interpretation of data, that is making sense of the data as a whole.
A logbook was kept to ensure the reproducibility of the analysis.
After each test of the PI process, internal reports were shared with INESSS staff members to discuss methods of improvement; a final report sent to those involved in the evaluation of the process was unanimously confirmed by all. In addition, in May 2018, we presented our findings at both a CADTH meeting 34 and an INESSS Forum, 35 during which we profited from feedback from the HTA community on which dimensions of our work were the most interesting for others in the field.

| RE SULTS
In this section, we present the four steps that were carried out to involve patients and to assess the PI process, as well as the perceived added value of our process according to others in the HTA community.

| The choice of level of PI
At the beginning of its mandate, the INESSS scientific team, in collaboration with INESSS specialists in PI, aimed to develop a PI strategy to support the organization's initiative to develop the recommendations through co-construction with patients. Firstly, a literature review on patient experiences, involvement in decision making and quality of life with respect to ICDs and particularly their replacement was conducted by the INESSS scientists. Information research strategies, adapted from a prior systematic review of high quality, 36 were used and applied to the PubMed, EBM Reviews and EMBASE databases ( Table 2).  Table 3).

| The co-construction of guidance
A modified Delphi method was subsequently carried out with the members of the joint committee through two consultation rounds by electronic mail. 40 Comments and agreement regarding proposed wording for recommendations on ICD replacement were collected from the patients and health-care professionals. This feedback was shared anonymously with all participants in the second round.
The revised recommendations and the comments from the Delphi rounds were then presented during the in-person co-construction committee meeting to discuss: (1) the recommendations that engendered the most comments related to shared decision making and

| The perception of the expert patients on their involvement
After each expert patient committee meeting, patients mentioned that these allowed for a discussion that helped them validate and contextualize the literature review, especially the sections concerning the patient's experience. They shined light on the themes they

| The perception of the added valued of PI by the INESSS scientists and by the expert healthcare professionals
For the INESSS scientists, the contribution of the patients was evident at several levels. First of all, it allowed most of the team members to be in contact with patients with cardiac problems, which really brought home the point that there were people behind the  As for the health-care professionals, they appreciated the transparency of the process which reduced their uncertainty about having patients act as partners, rather than mainly as advocates: '[I was not sure what to expect, but I was pleasantly surprised to see patients being able to react to complex issues and give us as many solutions as possible to improve the organization of services]'.

| D ISCUSS I ON AND A PI FR AME WORK PROP OSAL
This study has two main limitations. The first relates to the difficulty interviewing the persons on the expert health-care professional committee. Only two of the 6 clinician experts who were present during the co-construction committee meeting provided feedback on their experience, despite multiple efforts to contact them. This is probably best explained by a lack of availability on the part of the clinicians, but may also signal less interest in taking part in this kind of reflection. The second limitation concerns the fact that many of the authors of this article work for INESSS, which could lead to difficulty in being objective in the evaluative process. 31 We also made sure that the participation of the members of the patient committee was acknowledged, firstly by thanking them individually not only by telephone and by e-mail, but also by including their names in the reports that were published as a result of their work, with their permission. We ensured that lessons learned from the experience were translated into action to improve the process of developing HTA advice and practice guidelines. As a result of this project, a dedicated team, comprised of three partnership experts and two patients, was put in place to support the team methodology of patient participation in HTA, and a PI framework was adopted (see

| CON CLUS ION
This project, conducted at INESSS, contributes to future development in optimizing the incorporation of PI in HTA through a co-construction process, perhaps most importantly when the topics are complex. The evaluation of PI in the assessment of ICD replacement thus highlights the strengths and areas for improvement, as well as the challenges that have been encountered. INESSS strives to continue improving its methods to engage patients in partnership with health-care professionals in a more systematic way throughout the entire assessment process, even if time constraints are a limiting factor. This study shows that it is possible to carry out co-development of recommendations that combines patient and health-care professional experiences. This allows a more democratic process than usual and contributes to a real commitment to PI in HTA, which results in recommendations with more patient focus and ultimately more potential impact on health services delivery and population health.
Through this project, INESSS was able to propose a framework for meaningful PI in its various activities, a framework that could also help other HTA agencies to better structure their own approach and promote optimal IP in their evaluation work.

ACK N OWLED G EM ENTS
The authors acknowledge the contribution of all the patients and health-care professionals that participated in the different committees.

CO N FLI C T O F I NTE R E S T
The authors declare that there is no conflict of interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available in