This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
Many intervention development projects fail to bridge the gap from basic research to clinical practice. Instead of theory-based approaches to intervention development, co-design prioritizes the end users’ perspective as well as continuous collaboration between stakeholders, designers, and researchers throughout the project. This alternative approach to the development of interventions is expected to promote the adaptation to existing treatment activities and to be responsive to the requirements of end users.
The first objective was to provide an overview of all activities that were employed during the course of a research project to develop a relapse prevention intervention for interdisciplinary pain treatment programs. The second objective was to examine how co-design may contribute to stakeholder involvement, generation of relevant insights and ideas, and incorporation of stakeholder input into the intervention design.
We performed an embedded single case study and used the double diamond model to describe the process of intervention development. Using all available data sources, we also performed deductive content analysis to reflect on this process.
By critically reviewing the value and function of a co-design project with respect to idea generation, stakeholder involvement, and incorporation of stakeholder input into the intervention design, we demonstrated how co-design shaped the transition from ideas, via concepts, to a prototype for a relapse prevention intervention.
Structural use of co-design throughout the project resulted in many different participating stakeholders and stimulating design activities. As a consequence, the majority of the components of the final prototype can be traced back to the information that stakeholders provided during the project. Although this illustrates how co-design facilitates the integration of contextual information into the intervention design, further experimental testing is required to evaluate to what extent this approach ultimately leads to improved usability as well as patient outcomes in the context of clinical practice.
Only a fraction of intervention development projects is able to bridge the translational gap from scientific research to clinical practice [
An opportunity to increase the emphasis on these factors is to incorporate co-design methods. Co-design not only is characterized by an incremental knowledge over multiple development cycles [
Although co-design is increasingly adopted in the development of health care interventions (eg, [
In the present project, called the SOLACE project (grant number: SIA RAAK 2014-01-23), we developed a relapse prevention intervention for patients with chronic musculoskeletal pain who participate in an interdisciplinary, multimodal pain treatment program. The primary reason for adopting a co-design approach was that, despite high prevalence rates of relapse after successful treatment, there is a paucity of available research to explain relapse for this particular population [
To increase understanding of how co-design can be successfully applied in the development of interventions in the health care domain, more examples of good practice are needed [
We performed an embedded single case study [
The SOLACE project consortium consisted of 2 interdisciplinary multimodal pain treatment centers, the Royal Dutch Society for Physical Therapy, The Dutch National Pain Patient Advocacy Organisation, and 4 research groups with a respective interest in chronic pain treatment (2 groups), co-design, and behavior change. All consortium partners assisted with the recruitment of participants when this was required at specific co-design activities, including patients and their spouses, HCPs, designers, researchers, and students. The core team was composed of 3 researchers, each from a different research group. This team was responsible for the planning and preparation of the co-design activities. To monitor overall progress, a steering committee was formed, which included representatives of all consortium partners. Ethical approval for this study was granted by the local ethics committee (Medical Research Ethics Committee Atrium 15-N-120).
In interviews and co-creation sessions, the core team adopted various co-design methods, including generative techniques, contextual interviews, system mapping, and prototyping. These methods were adopted to facilitate stakeholder participation during key moments in the design process: generative techniques to elicit tacit knowledge and latent needs, contextual interviews to increase empathy, system mapping to develop a comprehensive overview of the acquired data, and prototyping to make ideas tangible and possible to experience. Co-creation sessions included multiple co-design methods and were specifically employed to empower a variety of stakeholders to participate in the design process.
At various time points in the project, we interviewed patients and HCPs. The interviews were performed by 2 researchers and were conducted in the everyday context of the HCPs (treatment facility) and patients (at home). To activate prior knowledge and experiences, all participants received “sensitizers”—assignments that stimulated thinking about relevant topics—before the interview (see page 4 in
To explore participants’ ideas, needs, and values beyond their first response, various generative techniques were employed during interview and design sessions. These techniques aim to bring up “tacit knowledge” by addressing social, emotional, and functional elements related to a topic of interest [
A key element of PAR is to increase insight by reflecting on actual interactions with prototypes. As Step 1.3 on page 5 of
System mapping is a method for creating a visual representation of interacting variables that facilitates the understanding of complex systems [
A specific way to represent the data as a coherent “whole” for usage throughout co-design activities is by creating personas: fictitious archetypes of users, each reflecting a distinct pattern in goals, attitudes, and behaviors based on empirical research among potential users. With personas, it is possible to highlight certain areas of tension or to facilitate discussion of important patient characteristics [
We used co-creation sessions at key moments during the project to discuss and reflect on the collected data, to generate new ideas, and to make decisions regarding future development directions (see page 10 in
The dataset for this case study consisted of 4 different sources. To capture the results of the design methods, researchers documented each design and research activity, using observation notes, pictures, audio files, or video clips. In addition to the session documentation, researchers also organized reflective sessions directly after a co-design activity to summarize the output of co-creation sessions (eg, notes or post-its) into system maps. These maps served both as a descriptive analysis of the data as well as for input during subsequent co-design sessions. The dataset also consisted of minutiae of steering committee meetings and a retrospective project journey. This journey was the result of a reviewing session, where researchers and members of the steering committee chronologically described and discussed critical incidents.
We used a deductive content approach to identify information within the dataset that relates to our main themes: stakeholder involvement, generation of insights, and incorporation of stakeholder input. We defined stakeholder involvement as the commitment to participate in the development project, to collaborate with other stakeholders during design activities, and to actively participate during these sessions. Generation of insights referred to the extent by which co-design activities resulted in an increased understanding of the problem of interest that could inform subsequent development activities. Incorporation of stakeholder needs was defined as the extent by which prototypes were based on stakeholder perceptions, judgments, and evaluations from co-design activities. Furthermore, we adopted the Double Diamond model to describe the design process along 4 development stages [
Overview of the co-design development process.
In the “Discover” phase, we aimed to generate a deeper understanding of factors influencing relapse after successful rehabilitation. The primary activities took place over a period of 11 months and consisted of 3 kick-off sessions, 20 stakeholder interviews (12 HCPs, 8 patients), and a student design project. In the first kick-off session, representatives from all consortium partners were present to discuss the project planning and to decide how co-design would be implemented throughout the project. Representatives also participated in co-creation by using their professional and personal experiences to formulate initial ideas on relapse (see page 3 of
In phase 1, we were able to create a large qualitative dataset. This dataset not only contained experiences and ideas of stakeholders but also included specific feedback in response to multiple provotypes on a wide array of topics. The consecutive planning of the 3 key activities enabled us to iteratively expand our insights on relapse after pain treatment: Interviews were prepared by using the insights from the kick-off sessions, and the student teams could build upon the preliminary analysis of the available interview data. The participating stakeholders responded positively to the co-design approach and cooperated actively during the sessions and interviews. Despite their inexperience with co-design, the sessions were considered accessible, pleasant, and relevant. However, medical ethical screening procedures and personnel deployment planning limited the possibility for last-minute requests or invitations for including HCPs and patients. The obtained dataset of patient and HCP responses also contributed to a deeper understanding of relevant factors related to relapse, which provided a solid base for further intervention development. For example, the interviews revealed important contextual information such as a “feelings of emptiness after treatment,” “difficulties sharing treatment experiences with friends and family,” and “the different context between the rehabilitation center and the personal environment.”
The “Define” phase lasted for 1 month and started with thematically organizing the interviews by means of open coding by the core team (see page 6 of
The final system map that included both posters and the card set provided a complete overview of the collected data. This presentation form stimulated participants to combine various insights to develop concept interventions. With respect to stakeholder involvement, the number of patients and HCPs was lower than originally planned. The duration of the session and traveling distance required participants to block a full day, which turned out to be difficult to organize. In line with our findings in phase 1, the co-design methods successfully engaged nonexperts in the design process. The assignment to create concept intervention ideas was concrete and tangible. The resulting 5 concepts were associated with earlier identified patient needs, were grounded in contextual information, and contained relevant insights on relapse prevention. For example, one concept idea focused on monitoring and recognizing early signals of relapse, which was based on stimulus cards (eg, a research insight related to difficulties in unbiased self-monitoring of behavior), interview data (eg, a quote from HCP on the possibility of daily feedback via eHealth), and newly added notes (eg, patient feedback should always be related to patient-specific goals). However, only a fraction of the possible combinations of cards and system maps was explored during this session. Limited time and resources prevented organizing additional sessions to cross-validate the results and achieve saturation.
During the 4 months of the “Develop” phase, students held 5 focus groups to regularly test their ideas with patients and HCPs (see page 11 in
This phase was characterized by a shift from “what” to “how” to design. Accordingly, presentation form, usability, and implementation into existing treatment practice became increasingly relevant. To engage stakeholders, the students visited the treatment centers on multiple occasions. In contrast to other phases, the patients and HCPs could provide feedback on ideas, but were not involved in the decision-making process regarding the final design of the rudimentary prototype, which potentially influenced their commitment. Moreover, their reduced involvement in this phase resulted in limited information regarding the applicability of the rudimentary prototypes in clinical practice.
In the final phase, the core team merged both rudimentary prototypes into one final prototype intervention over a period of 2 months. To do so, the core team organized a final co-creation session, where the students presented their concepts. The aim of this session was to receive final feedback on the potential value and function of both rudimentary prototypes and to formulate a recommendation to the steering committee with respect to the final prototype design. To facilitate this process, stakeholders (n=14) were instructed to reflect on the concepts by taking various patient perspectives into account. For this purpose, 4 personas were created with variation on 2 characteristics that were often discussed during previous patient interviews. Each persona had either a high or low level of social support and a high or low tendency to protect personal boundaries. In
Based on this advice, the steering committee decided to merge both rudimentary prototypes into 1 prototype workbook. The core team composed a list of individual intervention components from each rudimentary prototype (eg, a prompt to set calendar reminders after a goal-setting procedure) and coded these according to the Behaviour Change Technique Taxonomy V1 (see page 15 in
Previous difficulties with recruiting sufficient patients for co-creation sessions caused us to search for alternative ways to include their viewpoint. The personas proved a useful method to incorporate various patient perspectives by proxy during the evaluation of the rudimentary prototypes. Furthermore, the validation check indicated that the majority of the intervention components could be traced back to the original stakeholder themes from the interventions in the “Discover” phase and vice versa. This illustrates that stakeholder input has been incorporated in the design. However, the decision to combine both prototype ideas into one intervention was unexpected, which resulted in last-minute planning and consequently in limited stakeholder involvement during the design of the workbook. This may threaten the usability of this prototype in clinical practice.
The primary aim of this study was to reflect on the value and function of co-design methodology during the development of an intervention that prevents relapse after successful pain treatment. In the analysis, we focused on idea generation, stakeholder involvement, and the incorporation of stakeholder input within the development process. Overall, the generative techniques that were employed supported patients and HCPs with sharing their perspectives on pain treatment and relapse, which was in line with our hypothesis. Moreover, the techniques steered the conversations beyond stakeholders' primary responses, often resulting in a detailed account of their personal experiences with the treatment program and of their attempts to integrate treatment insights into their personal environment. In addition, system maps, personas, and prototypes enabled nonexperts to actively participate in design activities. A possible explanation for the successful engagement of stakeholders during the project is that experienced co-designers constantly translated hypotheses and abstract ideas into provotypes or prototypes. This method is particularly useful to provoke user reactions or to rapidly visualize an idea, which evokes interactions with an actual object rather than reflections on past experiences of hypothetical situations [
With respect to stakeholder involvement, many different patients, HCPs, researchers, students, and designers participated during the study, which was also in line with our hypothesis. The stakeholder interactions mostly consisted of independent design activities that required low commitment and little effort. In contrast, the members of the core team remained active throughout the project, which increasingly created an imbalance in knowledge and involvement between the core team and other participants in co-design activities. This may explain why the role of the stakeholders gradually shifted from “user as partner”—where all participants within the sessions contributed as equals in the design activities—towards “users as subject”—where participants mainly provided expert opinions or performed delimited tasks (eg, usability testing) [
This project shows similarities to the experience-based co-design (EBCD) approach, which aims to improve health care services by actively involving stakeholders to collect knowledge and experiences, to set priorities, and to develop solutions [
The extensive documentation of the co-design activities allowed for a detailed reconstruction of the development process. Furthermore, during co-creation sessions, steering committee meetings, and the construction of the retrospective journey, representatives from all research groups were present, which resulted in a continuous integration of various perspectives during the project. However, we did not film or record any of the co-creation sessions. Although analyzing audio or video would have been time consuming, it would have provided further possibilities to observe stakeholder discussions during design activities and to include additional insights that we did not record.
During the project, we experienced a tradeoff between validating the outcomes of co-design activities and analyzing the results for the next iteration. For example, an additional co-creation session during the “Define” phase with different stakeholders could have cross-validated the outcomes of the initial session. However, given limited resources, this would have resulted in fewer development iterations in the remaining period. A key argument in favor of more iterations is that quickly integrating stakeholder input into subsequent sessions directly visualizes the value of their input [
This study adds to the increasing number of initiatives that use co-design to structurally integrate contextual factors into the development of health care interventions (eg, [
To acquire a better understanding of how co-design may benefit the development of interventions in the health care domain, examples of good practice are necessary
Overview of the co-design development process.
experience-based co-design
health care provider
participatory action research
We thank the researchers Jan Pool, Manon Beetstra-Huszar, Albère Köke, and Roos Tigchelaar for their continuous support with collecting and analyzing the data. We also thank Rianne Schaekens, Deloryan Hommers, and Yasmijn Baas for their assistance in the development phase during their graduation projects.
None declared.