The primary outcome of this project is to demonstrate the feasibility of systematically developing a theory-driven, evidence-based DA for a context characterized by limited and evolving evidence—specifically, medical cannabis use for symptom management among older adults with arthritis. Feasibility will be demonstrated through (1), recruitment of at least 80% (n=8-10) of the target sample with ongoing consideration of information power, (2) retention of ≥80% (n=8-10) of participants recruited in phases 2 and 3 to provide feedback for phase 4, (3) adherence to a 12‐ to 18-month development timeline (based on comparable studies), (4) the generation of actionable user feedback to guide iterative refinement, and (5) comprehensive reporting of each development phase in accordance with the IPDAS-recommended DEVELOPTOOLS checklist [33]. Meeting these prespecified benchmarks will indicate the feasibility of developing a prototype ready to progress to beta testing.

The secondary outcome is to obtain acceptability and qualification as data that support proceeding to the beta testing phase. We will conduct the acceptability [40] and the DA qualification assessment [41] as part of alpha testing (iteration with participants). These assessments will provide structured feedback for iterative formative refinement rather than definitive validation for implementation. Measures and scoring procedures are prespecified under phase 4b. This approach will identify aspects requiring refinement before beta testing with naive users and is consistent with iterative development processes for DAs [33,38].

The advisory board was assembled under the leadership of the principal investigators, CAS (geriatrics specialized pharmacist) and EY (rheumatologist), who served as the first 2 members of the board. Drawing on their professional networks, they invited members with relevant expertise. Common interests were verified through conversations to ensure a collaborative, multidisciplinary team committed to supporting the development of a DA. The advisory board defined the scope and purpose of the DA.

Two reviews were conducted to scope the literature on (1) DSIs among patients with arthritis and (2) cannabis use in arthritis [42,43]. Both studies used systematic search and evidence synthesis approaches guided by Arksey and O’Malley’s methodological framework and the PRISMA-ScR (Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews) checklist [44,45].

Phases 2 and 3 aimed to explore patients’ and clinicians’ perspectives on cannabis use decision-making. Figure S2 in Multimedia Appendix 3 shows the participant flow throughout the study.

We adopted a qualitative descriptive approach, an inquiry method aiming at providing a comprehensive summary of participants, staying close to the surface of the data without extensive interpretation or theorizing [46-48]. This approach is particularly well suited for exploring the decisional needs of patients and clinicians, as it allows for the direct capture of participants’ perspectives in their own words—an essential consideration when the goal is to inform the development of practical, user-centered DA.

The research team acknowledges that their professional backgrounds and clinical experiences will inevitably shape how participants’ statements are interpreted. The lead researcher (HATA) is a health researcher with training in quantitative and qualitative methodologies and experience in decision-making, geriatrics, and chronic disease management, which may be associated with increased awareness of issues related to medical cannabis safety and evidence uncertainty. Coauthors’ stance aligns with the Canadian Rheumatology Association (CRA) Position Statement recognizing insufficient evidence for medical cannabis in arthritis, while calling for empathetic, pragmatic guidance that supports respectful patient-clinician dialogue and prioritizes harm reduction [49]. This stance may predispose the team toward cautious, harm reduction-oriented interpretations of participants’ experiences. To address this, we will engage in ongoing reflexive practice, including memoing and analytic discussions, to critically examine how our clinical perspective may influence theme development and framing.

For phase 2, the target population will comprise patient participants who are 55 years of age or older living with any form of arthritis, currently using, interested in, considering, or declining medical cannabis for symptom management. We initiated recruitment in a large urban center in Alberta, Canada. For phase 3, the target population will comprise clinician participants (eg, pharmacists, rheumatologists, family physicians, and nurse practitioners) involved in arthritis care and counseling about symptom management options. We initiated recruitment through advertising in a local medical newsletter.

Purposive sampling will be used, and the sample size will be guided by the principle of information power, wherein the adequacy of the sample depends on the degree to which participants collectively contribute information relevant to the study aim [50]. This approach is consistent with reflexive thematic analysis (RTA), which does not rely on saturation as a marker of adequacy [51,52]. High information power [50] is anticipated, given the narrow research aim (decisional needs regarding medical cannabis for arthritis), the high specificity of the sample, the use of an established theoretical framework (the ODSF) [30], and the expected richness of semistructured interviews guided by a decisional needs assessment workbook [53]. Based on comparable decisional needs studies, we anticipate recruiting approximately 10 to 12 participants per phase. The research team will review data relevant to ODSF constructs and assess information power after every 3 interviews to determine whether additional recruitment is required [50]. Recruitment will cease once consensus among advisory board members indicates that the accumulated data provide sufficient depth, breadth, and relevance to meet the study aim. The inclusion criteria are given in Textbox 1. Eligible participants who agree to participate will be provided with an information letter and consent form outlining the study’s purpose, procedures, participants’ rights, confidentiality and anonymity protocols, and data storage practices. A trained researcher (HATA) will conduct all one-on-one semistructured interviews using interview guides for patients and clinicians (Multimedia Appendices 4 and 5) informed by the decisional needs assessment approach, the ODSF, and probing techniques [34,53-55]. To support accessibility and accommodate participants’ preferences, interviews will be conducted virtually via a secure video conferencing platform, with telephone or in-person options available as needed. All interviews will be audio-recorded with participant consent and are expected to last approximately 45 to 60 minutes. At the end of each interview, patients and clinicians will be asked to provide demographic information (eg, for patients: age, education level, and cannabis usage status and for clinicians: years of practice, age, and area of practice). Audio recordings will be transcribed verbatim using Quirkos [56] and deidentified using unique participant codes (eg, P01 for patients, C01 for clinicians). Transcripts will be cross-checked against audio recordings for accuracy and completeness. Transcripts, audio files, and the corresponding linking key will then be securely stored on encrypted, password-protected university servers accessible only to the research team and compliant with institutional data protection policies.

To ensure transparency, rigor, and methodological clarity, both qualitative studies will be reported in accordance with the Standards for Reporting Qualitative Research guidelines [62]. The data from both phases will be analyzed using RTA situated within a qualitative descriptive approach, an analytic combination that enables the researcher to remain close to participants’ accounts while acknowledging that meaning-making is inherently interpretive [46-48,51]. RTA provides methodological flexibility suitable for exploring decisional needs in depth, while the qualitative descriptive approach ensures that findings are grounded in the language and experiences of older adults and clinicians [48,51].

A single researcher (HATA) will lead all analytic activities, consistent with RTA’s epistemological stance that positions the researcher as the primary analytic instrument and supports sole-coder analysis as a flexible, coherent, and methodologically appropriate practice [51,63]. The analysis procedure will begin with familiarization through repeated reading of verbatim transcripts and verification of transcription accuracy to immerse oneself in the data. A hybrid deductive–inductive approach will be used for coding. An initial coding framework based on the ODSF will be applied deductively to identify predefined decisional needs constructs (eg, knowledge gaps, expectations, values, decisional conflict, and support needs). In parallel, inductive coding will be used to capture unanticipated themes emerging from participants’ accounts and the cannabis context. Codes will be iteratively refined and organized into higher-order themes through RTA. Final themes will be explicitly mapped to the DA components, including informational content, values-clarification exercises, and clinician-patient discussion prompts, ensuring that the DA content is directly informed by empirically identified decisional needs of both patients and clinicians. Coding will be conducted in Quirkos [56] to organize excerpts, identify patterned meaning, and support early category development.

Throughout the analysis, HATA will generate initial themes, refining them through iterative cycles of reviewing and returning to the data, critical reflection, and agreement among the research team. This approach will serve to deepen interpretive insight by incorporating the research team’s diverse disciplinary standpoints in health services research, geriatric pharmacotherapy, rheumatology, and pharmaceutical sciences, consistent with RTA’s emphasis on reflexivity over reliability [51]. Themes will be assessed for coherence, consistency, relevance, and alignment with the research aims and the goal of developing the DA. Finally, themes will be defined and named by refining their boundaries and ensuring clear definitions.

To ensure trustworthiness across both qualitative descriptive studies, multiple strategies will be used to establish credibility, dependability, confirmability, and transferability [64,65]. Credibility will be enhanced through advisory board review of thematic findings, negative-case analysis to identify divergent patterns, and triangulation across patient and clinician datasets to explore convergent, divergent, and complementary insights [66-68]. A comprehensive audit trail with a versioned codebook documenting analytic decisions and coding iterations will be maintained throughout analysis to transparently track coding evolution and methodological reflections supporting dependability, while confirmability will be supported through extensive reflexivity practices, including researcher journaling and reflexive memoing [69-71]. Transferability will be established through thick description of study context, procedures, and participant characteristics, enabling readers to assess applicability to other settings [72]. The advisory board will provide a final review of the thematic structure to evaluate its relevance and usefulness for informing the DA prototype development.

Phase 4a focuses on generating the initial prototype, while phase 4b constitutes a formative, embedded alpha-testing phase situated within development rather than a summative validation intended for implementation. The purpose of alpha testing is to obtain structured, actionable feedback to refine the prototype prior to subsequent beta testing with naive users. To enhance methodological transparency and reproducibility, all outcome measures, instruments, scoring procedures, and a priori success thresholds are prespecified below.

Each participant will independently review the DA and complete two questionnaires:

Think-aloud and semistructured cognitive interview will be conducted with all participants to contextualize and corroborate questionnaire results [33,38,89]. Participants will elaborate on reasons for high or low ratings; perceived ambiguities or biases; navigation challenges or areas of cognitive burden; and suggestions for improving clarity, neutrality, and integration into clinical encounters.

Interviews will be audio-recorded, and field notes will capture contextual observations. Interviews will be analyzed to identify usability concerns, refine content, and guide iterative improvements before beta testing.

Quantitative and qualitative findings will be analyzed and used together to identify issues to be addressed through iterative refinement. All changes will be logged in a structured decision log, aligning with DEVELOPTOOLS recommendations for transparent reporting [33].

This alpha-testing stage ensures that the DA is evidence-based, clinically accurate, and responsive to end user perspectives before beta testing. By integrating evidence synthesis, structured content governance, explicit accessibility and readability standards, and predefined acceptability and qualification thresholds, the protocol offers a transparent and user-centered development process [33,38,90]. All modifications will be documented in a design-decision log to ensure transparency and traceability of development decisions.

Phase 1a is completed as the advisory board is assembled, and it includes 1 rheumatologist (EY), 1 pharmacist (CAS), 1 clinical and experimental science researcher (TKLK), 1 patient-oriented research expert (YNA), 1 graduate student researcher trained in qualitative research (HATA), and 1 patient living with arthritis. In alignment with the IPDAS development process, the advisory board collaboratively defined the scope of the DA. Following IPDAS guidance to clarify the target decision, user population, and context, the board identified the decision as preference-sensitive, involving trade-offs between potential benefits, risks, and uncertainties. Accordingly, the DA was scoped to assist patients and clinicians in engaging in SDM regarding the use of cannabis for symptom management.

Phase 1b is completed. Two scoping reviews were conducted to inform the development of the DA. The first review examined the development and evaluation processes of DSIs in rheumatology [38], identifying 36 DSIs classified as educational tools, DAs, or adaptive conjoint analysis aids. Development was primarily based on frameworks, such as the IPDAS (53%) and the ODSF (16%). Common evaluation measures included the IPDAS qualifying criteria and acceptability assessments. Decision tools varied in format (eg, paper based, web based, interactive) and timing of use (before, during, or after consultations). The second scoping review investigated evidence for cannabis-based medicines in osteoarthritis, revealing mixed evidence with only 60% of studies reporting improvements in pain [58]. Larger studies with longer exposure durations showed no significant benefits. The review highlighted limited high-quality evidence and the need for tailored information addressing patient concerns about safety, efficacy, and potential interactions. Based on the findings from both reviews, we decided to follow established standards (ODSF and IPDAS Development Process Model) emphasizing user-centered methods. These reviews provided actionable insights for developing DSIs tailored to rheumatology patients’ needs regarding cannabis use decisions.