Sage Journals: Discover world-class research

Abstract

Background. In primary care, general practitioners (GPs) and community pharmacists (CPs) play pivotal roles in guiding patients’ drug choices. Nonprescription drugs (NPDs), which are dispensed with or without a prescription, represent significant health care costs. NPDs are most often used for symptomatic relief and may be subject to shared decision making. We designed shared DECIsion in hEalth (DECIdE), a patient decision aid for NPDs in primary care, via a 3-step user-centered approach. Design. In the first step, a nominal group composed of potential future users (patients, GPs, and CPs) reached a consensus on the prototype’s specifications. In the second step, GPs, CPs, and their patients tested the prototype in user tests during simulated consultations based on real-life scenarios, and clinical psychologists conducted individual interviews to improve the prototype. In the third step, international experts used the eDELPHI consensus method to validate DECIdE as a shared decision aid. Results. Sixteen participants in the nominal group reached a consensus on 18 specifications. The user tests involved 16 patients and 4 health care professionals across 2 cycles. The prototypes were improved according to an analysis of 20 individual interviews. Fifteen French-speaking experts, searchers, physicians, and pharmacists from 4 countries reached a consensus on 11 of the 13 propositions inspired by the IPDAS criteria and the French HAS criteria in 2 rounds to validate the final prototype of DECIdE. Conclusion. Interprofessional collaboration at each step of development is the main strength of this user-centered design. DECIdE is currently being evaluated for its effects on decisional conflict in GPs’ practices and community pharmacies in a randomized controlled trial. Implications. DECIdE could encourage more rational use of NPDs, particularly in self-medication.

Highlights

DECIdE is the first patient decision aid in French for usual drugs in primary care, based on the latest scientific data in line with an evidence-based medicine approach for general practitioners (GPs) and community pharmacists (CPs).

DECIdE’s user-centered design includes interprofessional collaboration at each step between searchers, GPs, CPs, and social and clinical psychologists.

The original 3 steps of the user-centered design of DECIdE combined user testing with qualitative analysis and 2 successive consensus methods to create and validate the patient decision aid using the nominal group and eDELPHI methods.

The free availability of DECIdE on a dedicated Web site will enable GPs and CPs to make widespread use of shared decisions about nonprescription drugs daily.

Keywords

shared decision making decision support techniques primary health care nonprescription drugs

In primary care, general practitioners (GPs), as prescribers, and community pharmacists (CPs), as dispensers, play pivotal roles in guiding patients through their drug choices. To encourage patient involvement in health-related decisions, the World Health Organization advocated for the practice of shared decision making (SDM) in 2008¹ and urged governments to make SDM a fundamental element of health policies. To promote the practice of SDM, the French State Health Authority (HAS) published a methodological guide in 2018 on the development of patient decision aids (PDAs).² These tools support patients by making their decisions explicit, providing information about options and associated benefits/harms, and helping to align decisions with personal values.^3,4

In France, nonprescription drugs (NPDs) have marketing authorization and can either be prescribed—partially reimbursed by national health insurance—or purchased directly at the pharmacy without a prescription and fully paid for by the patient. Among NPDs that are reimbursable by health insurance, 65% of those dispensed in pharmacies have a prescription, and nonprescription sales accounted for more than 2 billion euros in 2023.⁵ While they are not alternative or complementary products, the clinical evidence supporting their effectiveness is sometimes lacking. Less than 12% of drugs have proven efficacy with a high level of evidence⁶ for important patient outcomes, and many drug approvals are based on uncertainty. For example, in France, acetylleucine (NPD) prescribed for vertigo⁷ and phloroglucinol⁸ (NPD) for pelvic pain in gynecology have been approved by health authorities, even though there is no evidence of clinical benefit. In addition to these situations of uncertainty, the use of NPDs often reflects a symptomatic approach, in which decisions must balance potential benefits, risks, and personal preferences. This makes them a relevant focus for SDM based on scientific data, patients’ values, and clinical expertise: deciding whether to take the drug or choose an alternative, including doing nothing, ensuring that NPD use is a choice rather than an automatic response to a symptom. Numerous sources of information on drugs are available for health care professionals on the internet, including institutional Web sites such as the French public drug database (Base de données publique du medicament), the database of the Prescrire® journal, drug registries such as Vidal®, and specialized Web sites such as the Reference Centre for Teratogenic Agents (CRAT). To date, in France, there is no scientifically based platform intended for health care professionals and their patients with respect to the risk–benefit balance of the most used drugs, including NPDs.

The shared DECIsion in hEalth (DECIdE) project aims to develop a PDA on NPDs for GPs, CPs, and their patients. As per HAS,⁹ SDM involves 2 key steps: the exchange of information and deliberation, aiming for a decision that is mutually agreed upon by the patient and the health care provider. DECIdE should help clarify the decision whether to take the NPD or opt for therapeutic abstention, during the medical consultation, when a prescription decision is being made, or at the pharmacy counter, when a patient requests an NPD (Appendix I). In line with the evidence-based medicine approach, DECIdE provides data about the risk–benefit balance of NPDs and incorporates the notion of uncertainty by including information about the level of evidence from the Rebuild the Evidence Base (REB) method. REB aims to improve the reading and interpretation of randomized clinical trials and meta-analyses to produce high-quality data and to help GPs focus on evaluating the risk–benefit balance in the best interest of patients.^10,11 To ensure that patients are not just informed but also actively engaged in the decision making, DECIdE should help health care professionals guide patients in identifying their individual needs, preferences, constraints, and resources. The aim of this study is to design DECIdE via a user-centered approach.

Methods

DECIdE was designed using a 3-step user-centered approach. First, a nominal group of potential users defined the prototype’s specifications; second, user testing with GPs, CPs, and patients in simulated consultations, followed by interviews with clinical psychologists, refined the prototype; and third, international experts validated DECIdE as a PDA using the eDELPHI consensus method.

Step 1: Nominal Group Consensus on the PDA Specifications

A nominal group consisted of potential future users of DECIdE: GPs, CPs, and patients. The GPs and CPs were selected internationally among French-speaking health care professionals from teams at the Université Clermont Auvergne, France; Université Laval, Québec, Canada; and Université catholique de Louvain, Belgique, collaborating in the DECIdE project due to their interest in SDM or evidence-based medicine. Patient volunteers were invited from members of patient associations belonging to a regional group, France Asso Santé AuRA.

Each participant received an individual e-mail that explained the aims of the PDA and asked a single question: “What are the elements or features (content and form) that should make up a tool to help GPs/CPs and patients make shared decisions about a drug?” All participants were free to answer individually. The answers were collected by 2 investigators (EC and TM) who individually analyzed the proposed elements of each participant and grouped them by theme before pooling the analyses. The participants were brought together by video meetings with the investigators to check that each of their proposed elements had been classified under the correct theme and to clarify and validate any necessary rewording to summarize the answers. When there were conflicting elements, a debate was held between the participants of the nominal group to retain only the useful elements, including those opposed, for the subsequent vote.

At the end of the meeting, a list of elements grouped by themes was created. Each proposal was put to a vote by each participant in the form of a questionnaire, which was sent individually using LimeSurvey software. The participants answered a list of questions that, depending on the wording of the proposal, enabled them to give a yes/no/do not know validation, to provide a level of agreement on a 5-point Likert scale, or to vote for a range of elements among items to be defined as “essential” or “to be rejected.” A social psychologist (MM), expert in cognitive biases, individually received the questionnaires and added comments to guide the design of the PDA while avoiding cognitive bias as much as possible. The retained elements were those that had received an absolute majority of “yes” votes for the binary questions, an absolute majority of “tend to agree” and “agree” votes for the level of agreement questions, and an absolute majority of “essential” votes.

The elements accepted by consensus were used to draw up a set of specifications for a graphic designer specializing in graphic facilitation (i.e., helping people understand through illustration and graphic representation). Based on these specifications, 3 prototypes of DECIdE for 3 different NPDs were used as “examples.” These 3 drugs, for which REB reviews and grading data were available, were selected by the scientific committee to reflect frequent requests for NPDs in general practice or pharmacies. These prototypes were reviewed by the social psychologist (MM) to limit potential risks of cognitive bias.

Step 2: User Tests

Recruitment of health care professionals and patients

Two GPs and 2 CPs were invited by the investigators from university lecturers in the 2 disciplines. Care was taken to balance the variables of age, sex, and place of practice (i.e., urban or rural) as much as possible.

Each health care professional recruited 2 patients from his or her usual patient base. Eligible patients were adults (>18 y old) with an established relationship with the GP or CP and sufficient understanding to discuss medications. This level was judged clinically by the professional, who adapted explanations to the patient’s understanding. We excluded patients who were suffering from a severe acute or unstable chronic pathology.

User test cycle

For each user test cycle (Figure 1), the same 4 GPs and CPs conducted simulated consultations with the 2 patients at their place of practice. Each cycle involved 8 new patients.

Figure 1

User-tests cycles.

To ensure that the tool could be evaluated under conditions close to real-life practice, the user tests were based on simulated consultations using standardized scenarios and conducted within the framework of an already established patient–professional relationship. Two standard scenarios were written for the use of each of the 3 DECIdE prototypes (Appendix II). The patient randomly drew 1 of the 6 scenarios and read it. The investigator (EC or SC) checked that the patient understood the scenario, answered any questions, and then left the room to make way for the health care professional. The health care professional conducted the simulated consultation on the basis of the scenario played by the patient using the corresponding DECIdE prototype to make a shared decision about whether to use the NPD.

The simulated consultations were audio recorded and directly observed by a clinical psychologist (CD-M or MC) who were not involved in the tool’s initial design. In line with the UCD 11 criteria, detailed in the DEVELOPTOOLS Reporting Checklist,¹² the observations provided an external perspective on user interactions with the PDA. Patients were encouraged to verbalize their reactions and questions about the tool and its content to the health care professional while using DECIdE, helping to capture spontaneous difficulties and misunderstandings. This instruction was inspired by think-aloud protocols without enforcing a strict concurrent think-aloud protocol, allowing for natural interactions and avoiding disruptions in decision making or complex exchanges.¹³ After the consultations, the psychologists conducted individual interviews with each participant (patient and GP or CP) to collect their direct feedback on their experience, usability, and areas for improvement. We calculated the average length of the simulated consultations, an important factor of acceptability for professionals. The psychologist wrote up a nonverbal communication analysis and information about the user’s immediate experience, communication difficulties, and intellectual pathway. After the consultation, the psychologist conducted semi-structured interviews with each user (patient and GP or CP) on 6 themes: general feelings, PDA representation, encountered difficulties, positive aspects, suggested improvement, and the health care professional–patient relationship. The recordings were fully transcribed by an investigator (SC).

User test analysis

To ensure a rigorous qualitative analysis, 2 investigators (EC and TM) independently coded and analyzed the transcriptions of the consultations and interviews before pooling their findings.¹⁴ A general inductive approach¹⁵ was used to identify key themes emerging from the data, focusing on aspects influencing the user experience, including the clarity and comprehensibility of the information, as well as the overall usability and acceptability of the PDA. Coding discrepancies were discussed through data triangulation with the 2 clinical psychologists, who had directly observed the user tests, to ensure a comprehensive and reliable interpretation. The identified areas for improvement guided an iterative refinement of the prototype, ensuring that the tool evolved based on real user feedback until no further significant modifications were necessary.

Ethics considerations

The protocol received approval from the ethical committee of the GPs’ national college (CNGE) (approval No. 101122415). Prior to simulated consultation, an investigator (EC or SC) provided a study information leaflet to the patient, explained it, and obtained written consent for audio recording and participation in the study. The consultation was free of charge and did not involve any cost to the patient or reimbursement by health insurance. The study was reported to the French National Commission informatic and liberties (CNIL) following the MR004 reference method via the Clermont Auvergne University.

Step 3: Experts’ Validation by Consensus Using the eDELPHI Method

Construction of the questionnaire (Appendix III)

A list of 13 proposals was created on the basis of the International Patient Decision Aids Standards instrument (IPDASi) v3¹⁶ and the HAS criteria.² Two researchers (EC and LM) independently analyzed and selected criteria based on their relevance to NPD decision making in primary care, particularly in situations of uncertainty. This selection was guided by the clinical context and the tool’s intended function, focusing on key elements such as uncertainty levels, scientific evidence, and the assessment of patient preferences. No specific scoring system was used in this process. In the event of disagreement, consensus was sought with a third researcher (HV-R). Eleven questions were selected based on the IPDAS criteria. A proposal was created based on the HAS criteria. This question about the assessment of patient values and preferences was the only one that differed (in its wording) from the IPDAS criteria. Finally, a general question on the validation of the tool was asked at the end of the questionnaire.

The experts were invited to provide feedback on each proposal and share their views on the PDA at the end of the questionnaire.

Experts’ recruitment and diffusion of the questionnaire

International French-speaking SDM experts were recruited by e-mail, including researchers, GPs, and pharmacists (not only CPs), to balance the number of representatives per profession. If they accepted, they received an abstract about the previous steps containing results from the nominal group consensus, the 3 final prototypes of DECIdE, and the questionnaire. The entire procedure was conducted anonymously by e-mail without any contact between the experts.

Analysis

The participants were asked to rate their level of agreement with the 13 proposals on a 9-point Likert scale (from 1 = strongly disagree to 9 = strongly agree) and to systematically comment on their answers (mandatory for each ranking under 4). A proposal was either accepted (median from 7 to 9 inclusive), rejected (median from 1 to 3 inclusive), or left in doubt (median from 4 to 6 inclusive). Disagreement between the experts indicated that at least 30% of the experts rated their agreement from 1 to 3 and at least 30% of the experts scored their agreement from 7 to 9. Consensus on acceptance (or rejection) of a proposal was reached when the proposal was accepted (or rejected) without disagreement between the experts. Any other situation (e.g., proposal left in doubt and/or disagreement between the experts) led to a new round with possible changes in the proposals and the DECIdE prototypes according to the experts’ comments. For the next round, the experts received the new questionnaire with the reworded questions for which there was no agreement and the new DECIdE prototypes.

The funding source had no role in these 3 steps.

Results

Step 1: From the Nominal Group Consensus to a Set of Specifications for DECIdE Prototypes

Sixteen participants answered the question of the nominal group (Appendix IV). There were 6 GPs, 6 CPs, and 4 expert patients. All health care professionals had their own activity in a practice or a community pharmacy in primary care, and most of them also had university activity. Three of the GPs were considered experts in SDM because they had conducted research on this topic.

The individual answers referred to a wide range of elements in terms of both content and form as well as the entire project from design to perspectives. They raised questions about the general method on which the design of the tool was based, such as the involvement of users in its construction, the notion of references for the data provided on drugs, and the implementation of the tool. The analysis of the answers was presented in the video meeting by 2 investigators to 9 participants: 5 GPs, 2 CPs, and 3 expert patients. The classification of the themes proposed in the synthesized document is presented with quotes in Table 1. After a presentation tour, the elements were clarified and commented upon by participants with recognition of their ideas and those who wished to react, enrich, or link their thoughts. At the end of the meeting, the investigators selected elements regarding feasibility and relevance to provide concrete guidelines for the graphic designer. The elements were rephrased according to the meeting’s comments and pooled into an online questionnaire of 16 questions that were sent individually by e-mail for voting. All participants in the nominal group voted.

Table 1

Classification of the Themes of the Individual Answers from the Nominal Group

Classification	Themes	Quote
Generalities	Validity and methods	“The editorial committee includes healthcare professionals and patients.”
Generalities	Scientific references	“The selection of information is unbiased and the result of the selection process is defined in advance.”
DECIdE homepage	Describes the objectives of DECIdE	“The decision aid must not replace the relationship. It must enable exchange.”
	Define shared decision	“Don’t confuse an informed decision with a shared decision.”
	Instructions for use	“Plan appropriate consultation time.”
Content	The symptom	“Explain simply, define the condition or symptom.”
	The medicine	“Illustrate the drug’s mode of action.”
	Different option	“Precise treatment options: reference medicine, nonpharmacological treatment, therapeutic abstention.”
	About benefits	“Mention the quality of supporting evidence (confidence in results), level of evidence, uncertainties.”
	About risks	“Mention the main (most frequent) side effects with NNH if available.”
	Patient decision	“The patient must be able to express these previous experiences.” “Propose these questions earlier as and when the information is provided.” “Conversational guide.”
Form	Access/context (Web site access)	“Ergonomics of the tool - Practical - Clear information for the patient.”
	Structure	“A different interface for GPs and pharmacists?”
	Representation of the effect size	“Numerical information is presented in several formats.”
	Representation of risks	“NNH if available.”
Diverse/out of topic?	Personalized data	“Attention paid to polymedicated patients”
Diverse/out of topic?	Connection to other medical devices	“Connected to patient personal data by professional healthcare software.”

GP, general practitioner; NNH, number needed to harm.

A total of 18 elements were selected by consensus of the nominal group. The social psychologist made some recommendations concerning the colors to be used or certain formulations. For example, to compare the “benefit” and “side effects” in an equivalent way, it was recommended to use fewer opposite words, such as “expected effects” and “undesirable effects.” All elements formed the specifications (Appendix V) given to the graphic designer to design the first 3 DECIdE prototypes based on REB data on phloroglucinol for pelvic or intestinal abdominal pain, sodium alginate–based antacids for symptoms associated with gastroesophageal reflux, and topical ketoprofen for ankle contusions or sprains (Appendix VI).

Step 2: User Tests

Study population

Two cycles of user tests were performed. The prototypes were first tested with 2 CPs, 2 GPs, and 8 patients and then improved according to the results of the first cycle. A second cycle tested the new prototypes with the same health care professionals and 8 new patients. The characteristics of the 20 involved users are presented in Table 2.

Table 2

Characteristics of Potential Users

Characteristic	n
Status, n
GP	2
CP	2
Patient	16
Sex, n
Male	10
Female	10
Mean age, y
GPs	53
CPs	51
Patients	63,7
Practice environment of health care professionals, n
Urban	3
Rural	1
Patients’ level of education, n
General certificate of secondary education^a	2
High school diploma	3
Graduate studies	11

CP, community pharmacist; GP, general practitioner.

French equivalent degree: Brevet des collèges.

Time of use

The mean consultation time across both user-test cycles was 7 min 20 s (Table 2). However, there was a decrease between the 2 cycles: the mean consultation time in cycle 1 was 9 min 12 s (±112 s), while in cycle 2 it was reduced to 5 min 27 s (±55 s).

Cycle 1 results

User tests from cycle 1 highlighted patients’ satisfaction with the readability of the international nonproprietary drug name, its pharmaceutical forms, and the symptoms it targets. Users also appreciated the source display. However, testers reported difficulties in visually distinguishing between the sections on expected effects and undesirable effects. They also questioned the relevance of the question, “Do you have enough information to form an opinion? If not, talk to a healthcare professional.” Since DECIdE was used during consultation, this question did not make sense. This first also identified formatting issues, such as misaligned text justification, and terminology errors, including the phrase “results to be confirmed” instead of “evidence to be confirmed.”

Development of prototype 2

These results were taken into consideration in the second version of the DECIdE prototype:

Enhanced readability: adjustments in font, colors of the boxes, and the use of framed sections to better separate expected effects from adverse effects.

Improved graphical representation of adverse effect frequencies (Figure 2), which involves the graphical expression of very small numbers: users had difficulty understanding the first pictograms using comparison from village, city, and country (V1), which were replaced with new ones representing gauges (V2).

Figure 2

Visual representation of the frequency of undesirable effects.

Cycle 2 results

The second cycle of user tests did not identify any major areas for further improvement beyond minor typographical corrections. The users expressed their satisfaction with the pictograms. The representation of the frequencies of side effects was meaningful to them. As the analysis did not highlight other issues, the data were sufficient to limit user tests at 2 cycles, leading to the elaboration of the third prototype of DECIdE.

Third prototype after user tests

For the third prototype (Appendix VII), typos were corrected and the representation of expected effects was replaced by the mention of absolute risks following advice from the social psychologist and an evidence-based medicine (EBM) expert from the scientific committee.

Step 3: Expert Validation by Consensus via the eDELPHI Method

Characteristics of the experts and flow chart

The characteristics of the 15 experts who participated are detailed in Table 3. The experts were from Belgium, Canada, France, and Switzerland (Figure 3).

Table 3

Characteristics of Experts in the eDELPHI Consensus Method

	Experts (n = 15)
Women	5
Men	10
Mean age, y	49
Profession
Pharmacist	3
General practitioner	8
Medical doctor from other specialty (public health)	3
Researcher	1
Country
France	8
Université Paris-Est Créteil (UPEC)	1
Université Paris Saclay	1
Université Claude Bernard Lyon 1 (UCBL1)	3
Aix Marseille Université (amU)	1
Université Clermont Auvergne (UCA)	1
Université de Bordeaux	1
Belgium (Université catholique de Louvain [UCLouvain])	1
Canada (Université Laval [ULaval])	4
Switzerland (Université de Lausanne [UNIL])	2

Figure 3

Flow chart of the eDELPHI experts.

First-round results

Analysis of the descriptive statistics on the level of agreement for each question (Table 4) revealed that questions 1 to 3, 5 to 7, and 10 to 13 were accepted by consensus, whereas questions 4, 8, and 9 led to disagreement between the experts. Following the experts’ detailed comments, DECIdE was improved after the first round:

harmonization of units and denominators for numerical data,

encouraging patients to express their values and preferences,

simplification of medical terms and clarification of their meaning,

numerical data about effect size were included only if the level of evidence was “solid” or “to be confirmed” according to the REB method, and

deletion of data on phloroglucinol’s effect on intestinal pain, as new data were published and DECIdE was designed for 1 molecule under a given condition.

Table 4

Statistics from First- and Second-Round Answers

Proposal	Level of Agreement, Median	Level of Agreement, 1 to 3 (%)	Level of Agreement, 4 to 6 (%)	Level of Agreement, 7 to 9 (%)	Agreement between Experts	Consensus	Accepted or Rejected
P1	7	6.6	20	73.3	Agreement	Yes	Accepted
P2	7	0	6.6	93.3	Agreement	Yes	Accepted
P3	7	0	33.3	66.6	Agreement	Yes	Accepted
P4	6	13.3	40	46.6	Agreement	No	Doubts
P5	8	0	20	80	Agreement	Yes	Accepted
P6	8	6.6	6.6	86.6	Agreement	Yes	Accepted
P7	7	6.6	26.6	66.6	Agreement	Yes	Accepted
P8	5	33.3	53.3	6.66	Agreement	No	Doubts
P9	5	40	33.3	26.6	Agreement	No	Doubts
P10	8	0	33.3	66.6	Agreement	Yes	Accepted
P11	8	0	26.6	73.3	Agreement	Yes	Accepted
P12	8	0	0	100	Agreement	Yes	Accepted
P13	7	0	26.6	73.3	Agreement	Yes	Accepted
Statistics from second-round answers
P1	8	6.66	40	53.7	Agreement	Yes	Accepted
P2	6	26.6	26,7	46.7	Agreement	No	Doubts
P3	5	21.4	50	26,6	Agreement	No	Doubts

The results enabling consensus to be reached on the proposal are shown in bold.

The prototype on ketoprofene was replaced by one on diclofenac because in France, topical ketoprofene in gel is no longer classified as an NPD. A brochure explaining the levels of evidence for the public was created in response to the experts’ request and was attached to the invitation for the second round with the new prototypes.

Second round

Question 8 on the psychological implications of the option was slightly reworded to better correspond to DECIdE’s objective. Questions 4 and 9 were not modified, but the improvement of the prototypes could make the experts change their minds. The statistics concerning the experts’ answers for the second round are presented in Table 4. Only question 4 on the numerical expression of the data was accepted with consensus. At the end of the second round, there was still no consensus on questions 8 and 9.

The relevant experts’ comments enabled the DECIdE prototypes to be improved one last time to obtain the final prototype (Figure 4 and Appendix VIII):

Different representations of numerical data as percentages and ratios of the same denominator for side effects

Adaptation to the lowest possible level of numeracy: removal of the mathematical signs for “lower than” and replacement with short sentences, for example, “less than 1 patient in 1,000”

Removal of standard deviations, following expert recommendations to simplify data readability

Explicit display of the “do not take the drug” option, ensuring all alternatives were considered

Refinement of wording related to values and preferences to encourage active patient participation

An invitation was added to express the need for more information to make the decision

A distinction was made between diclofenac only and diclofenac epolamine

Figure 4

Final prototype of DECIdE for sodium alginate–based antacid in reflux.

Discussion

Main Results

DECIdE development followed a rigorous, iterative user-centered design process involving patients, health care professionals, and experts in SDM. Based on the consensus of 16 potential users in the nominal group, 18 key specifications were identified to produce DECIdE first prototypes. Two cycles of user testing with 16 patients and 4 health care professionals in simulated consultations allowed improvements in readability, graphical representation, and usability. Fifteen French-speaking SDM experts then validated 11 of 13 proposals inspired by IPDAS and HAS criteria through a 2-round eDELPHI consensus. The final version of DECIdE integrates optimized numerical data presentation, simplified terminology, and explicit decision-support elements, making it a well-balanced, evidence-informed PDA on NPDs, tailored for use in primary care.

Strengths and Limitations

The main strength of this research is the involvement of potential users at each step of design and the involvement of patients and health care professionals from the field from the first step. A total of 20 patients were involved in the first 2 steps. As described by Vaisson et al.,¹⁷ the early involvement of users, prior to prototype development and iterative methods, including observations of users interacting with prototypes and questioning about their experience (user tests), align with definitions of a user-centered design. Documenting these changes provides a structured record of design rationale, demonstrating the rigorous iterative approach adopted. This user-centered approach meets the IPDAS criteria¹⁶ and French HAS² criteria for the development of a PDA. Interprofessional collaboration is the originality of this project and occurs at each step to add even more value to the exchanges. The nominal group encouraged direct exchanges of views between GPs (prescribers) and pharmacists (dispensers). During the user tests, the clinical psychologists’ in-depth analysis of the users’ experiences was discussed with the GP investigators. Moreover, the particular attention paid to reducing cognitive biases induced by information presentation in DECIdE, with the involvement of a social psychologist specialized in cognitive biases at each stage of development, should be highlighted. This approach, combined with feedback from users (patients and health care professionals) and the expertise of SDM and EBM specialists during the eDELPHI consensus, aims to identify the desired challenge of a balanced presentation of information: objective, complete, salient, transparent, evidence informed, and unbiased.¹⁸ While we did not specifically test for cognitive biases as suggested by Ubel et al.,¹⁹ their findings highlight that even the order in which information is presented can influence patients’ knowledge and perceptions.

From a methodological point of view, the need for empirical data regarding needs in real life led to the choice of the nominal group technique. This consensus method enabled the participation of each participant, both patients and health care professionals, particularly in the phase of individual answers. Furthermore, proposed elements arising from discussions and debates between professionals and patients were explored. This structured group technique encouraged and enhanced the participation while avoiding problems of group dynamics that could arise in focus groups.²⁰ During the video meeting feedback, the recruitment of patients from patients’ associations, who were accustomed to exchanging views with health care professionals, encouraged all participants to express themselves. Then, user tests seem to be the best way to observe users’ reactions and experiences with DECIdE. The aim of user tests is to optimize ease of use, ergonomics, legibility, and user experience.²¹ By simulating the tasks, the user’s behavior and reactions are observed to identify any interaction difficulties.²² The recruitment of patients from among those usually treated by health care professionals made it possible to diversify the profile of the patients included (compared with representative patients from patient associations in the nominal group). It also allowed for observations in conditions closer to real-life practice, considering the therapeutic alliance and existing patient–provider trust. Possible biases in interpreting and classifying the proposed elements were limited by the parallel individual analysis by 2 researchers, the adjustments made during the video meeting, and the comments during the votes.

However, each step of the project has its limits. There was an imbalanced nominal group with fewer patients than health care professionals, which was offset by the larger number of patients in the user tests. The main limits of the user tests were related to simulated consultations. The health care professionals asked patients to participate, whereas in real life, the request would come from the patient. This could have modified their behavior and motivation. The professional’s selection of patients from his or her usual patients favored behaviors most faithful to a real-life consultation and resulting from a real relationship. This resulted in an obvious selection bias but allowed the tests to stay close to the real conditions in which the PDA would be used. Most patients had completed superior studies and may have had a better level of literacy; this could have been the result of this selection bias. The limited inclusion of patients with low literacy levels, along with the exclusion of those suffering from severe or unstable pathologies—motivated by ethical approval requirements—also limits the generalizability of the tool. Moving forward, testing and adapting DECIdE for these specific groups will be essential for its implementation in clinical practice and will be made possible by its evaluation in the ongoing real-world RCT that follows on from the project. The 2 cycles did not enable the health care professionals to be familiar with the prototype, and they needed time to adjust. However, this situation allowed them to test DECIdE for potential first users. Only 3 prototypes for 3 drugs were developed. In the pharmacy, the absence of alternative drugs in the DECIdE leaflets disrupted the usual process of pharmacists, who usually suggest alternatives to patients. Two proposals of the e-DELPHI questionnaire did not reach a consensus for acceptance, which raises the issue of different results in a third round. Following analysis of the experts’ comments, the investigators chose not to propose a third round because these questions did not seem to be relevant for DECIdE. Question 8 about psychological impact did not correspond to any data available in DECIdE. Question 9, which was about space saved to express patient values and preferences, should be considered to develop a preconsultation DECIdE version. For this initial within-consultation version, final prototypes proposed a simple invitation for patients to express and highlight what is important for them in the form of a conversation guide. The focus on French context in the development and initial testing of DECIdE was essential to address the specific needs arising from the use NPDs in primary care settings in France. During the development process, collaboration with potential users and experts from other French-speaking countries suggests that users could foresee DECIdE’s usefulness across different health care contexts. The flexibility of DECIdE—designed as an adaptable framework for various symptoms and medications—supports its realistic adaptation to other languages and health systems for broader international use.

Comparison with the Literature

This user-centered development does not guarantee that professionals will incorporate this tool into their practice. The time of use was accurately cited by potential users at each step of development. As reported by Légaré et al.²³ in 2008, the lengthening of consultation times is one of the main obstacles identified by health care professionals. The average time of use of DECIdE did not exceed 8 min during the user tests. In a review by Stacey et al.³ in 2024, the length of consultation was 1.5 min longer in 8 studies representing 2,702 participants, with within-consultation PDAs. The absence of the difference with PDAs used in preparation for consultation could encourage the adaptation of DECIdE in a preconsultation version.

Considering the representations of data on drugs in DECIdE, the choice of communicating effect sizes in the form of absolute risk rather than relative risk is widely supported in the literature.^24,25 Particular attention has been given to the graphic representations proposed in the DECIdE prototypes with the expert for graphic facilitation. Zipkin et al.²⁴ reported that visual aids can improve patients’ understanding of probabilistic information. Icon arrays are usually used to express size effects, but Bonner et al.²⁶ highlighted that the effectiveness of specific visual displays crucially depends on how they are designed. Icon arrays were not favored by all potential users of the nominal group of DECIdE, and the number of data to be presented for the different effects of an NPD had to be limited to favor an ergonomic tool.

Regarding patients’ involvement in the DECIdE project, they were invited to participate in 2 ways: in the nominal group for patients from patients’ associations and in the user tests for patients selected directly by professionals. Fifty percent of the PDAs’ projects reported involving users in at least 1 step for understanding users, 35% in at least 1 step for developing/refining the prototype, and 84% in at least 1 step for observing users’ interaction with the prototype.¹⁷

As dispensers, CPs play an important role in this project. The strengthening of the therapeutic alliance with CPs and their role in supporting therapeutic choices is in line with developments in the profession. In 2012, Bedhomme et al.²⁷ studied the role of CPs in therapeutic education and stated that due to proximity to the patient, CPs clearly play a role in improving compliance with treatment and ensuring that the messages are consistent throughout the patient’s care pathway. This role in patient support was strengthened in 2018 by the introduction of the shared medication assessment in the French national agreement for CPs.²⁸ The provision of support tools, such as DECIdE, meets one of the expectations for developing the clinical pharmacy.²⁹ Because patients’ decisions in self-care and self-medication are a factor that contributes to health outcomes, the effective interaction of CPs with patients in terms of medication adherence and quality use of drugs provides better health outcomes among patients, according to Rajiah et al.³⁰

Implications for Practice

As the first PDA in French about usual drugs in primary care, DECIdE responds to the need to inform patients about what they can realistically expect from drugs, based on the latest scientific data in line with an EBM approach for GPs and CPs. The user-centered design and free availability of DECIdE in a dedicated Web site should enable GPs and CPs to make widespread use of this tool. First developed for NPD, DECIdE could contribute to more reasonable use of drugs for symptomatic issues, particularly for self-medication. DECIdE must be evaluated in real-life interactions by focusing on its effects on decisional conflict and the consumption of NPDs. Although initially tested in a French context, DECIdE’s flexible design suggests it could be adapted for other health care systems through actual and future international collaborations.

Conclusion and Perspectives

DECIdE, DECIsion in hEalth, is a PDA for NPDs in primary care, based on scientific data from the REB project. It was designed following 3 user-centered steps by combining consensus methods and user tests involving patients, GPs, CPs, and experts in SDM. Interprofessional collaboration at each step was the main strength of this project. DECIdE will be evaluated in a randomized controlled trial for its effects on decision conflict in 48 practices and pharmacies with 288 patients in real-life conditions.

Supplemental Material

sj-docx-1-mpp-10.1177_23814683251382260 – Supplemental material for Designing DECIdE Together: An Interprofessional and Patient-Centered Approach to Develop a Patient Decision Aid for Drugs in Primary Care (DECIsion in hEalth)

Supplemental material, sj-docx-1-mpp-10.1177_23814683251382260 for Designing DECIdE Together: An Interprofessional and Patient-Centered Approach to Develop a Patient Decision Aid for Drugs in Primary Care (DECIsion in hEalth) by Elodie Charuel, Thibault Menini, Sarah Chateauneuf, Léa Mathieu, Mélody Mailliez, Céline Deveuve-Murol, Marielle Duchassaing, Sabrina Bedhomme, Philippe Vorilhon and Hélène Vaillant-Roussel in MDM Policy & Practice

Footnotes

Acknowledgements

We especially thank Pr Jean-Pierre Lebeau and Dr Nora Moumjid-Ferdjaoui for their invaluable methodological advice. We thank all the members of the scientific committee of the DECIdE project. We thank all the participants of the nominal group and the user tests and the experts of the e-DELPHI method. We thank the graphic designer Facillistrus.

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Financial support for this study was provided in part by a grant from the call for interregional primary care research projects (Recherche en Soins Primaires-InterRégional; ReSP-Ir) launched by the Ministry of Health and Prevention and the General Directorate for Healthcare Provision (Direction Générale de l’Offre des Soins; DGOS). The funding agreement ensured the authors’ independence in designing the study, interpreting the data, and writing and publishing the report.

Authors’ Note

This work was presented at the following:

• “Journées de l’école doctorale 2024”: the days of the “Life Sciences, Health, Agronomy, Environment doctoral school” (ED SVASAE) in 2024 in Clermont-Ferrand, France

• 12th International Shared Decision Making (ISDM) Conference of the International Shared Decision Making Society in 2024 in Lausanne, Switzerland

Ethical Considerations

For the user tests, this study obtained approval from the ethical committee of the GP National College (CNGE) (approval No. 101122415). The data collection was reported to the French National Commission Informatic and Liberties (CNIL) following the MR004 reference method via Clermont Auvergne University.

Consent to Participate

For the user tests, before the simulated consultation, one of the investigators explained the study information leaflet to the patients and obtained the patients’ written consent regarding audio recording and participating in the study.

Consent for Publication

Not applicable. There is no identifying information, including patients’ names, initials, or numbers, in the written descriptions. For the user tests, this study was declared to the French National Commission Information and Liberties (CNIL) following the MR004 reference method via Clermont Auvergne University.

ORCID iDs

Elodie Charuel

Mélody Mailliez

Philippe Vorilhon

Data Availability

All data will be available in the trusted digital repository HAL (Hyper Article en Ligne, hyper online article). HAL is a multidisciplinary open archive for the French scientific and academic community: .

References

Coulter

Parsons

Askham

. Where are the patients in decision-making about their own care?World Health Organization. 2008. Available from: http://apps.who.int/iris/bitstream/handle/10665/107980/Policy-brief-3-1997-8073-eng.pdf?sequence=15&isAllowed=y

Haute Autorité de Santé. Éléments pour élaborer une aide à la prise de décision partagée entre patient et professionnels de santé - fiche méthologique. 2018. Available from: https://www.has-sante.fr/upload/docs/application/pdf/2018-03/elaborer_une_aide_a_la_prise_de_decision_partagee_mel.pdf. [Accessed 31 August, 2021].

Stacey

Lewis

Smith

, et al. Decision aids for people facing health treatment or screening decisions. Cochrane Database Syst Rev. 2024;1(1):CD001431. DOI: 10.1002/14651858.CD001431.pub6

Haute Autorité deSanté. Etat deslieux. Patients etprofessionnels de santé: décider ensemble. Concepts aides destinéesaux patients et impact de la‘décision médicale partagée’. 2013. Available from: https://www.has-sante.fr/upload/docs/application/pdf/2013-10/12iex04_decision_medicale_partagee_mel_vd.pdf

NereS Association. Baromètre 2023 des produits de santé et de prévention de premier recours en pharmacie. 2024. Available from: https://neres.fr/wp-content/uploads/2024/01/Barometre-2023-NereS.pdf

Maynié-François

Tudrej

Tawil

, et al. Quality of evidence of the efficacy of therapeutic interventions on patient-important outcomes in Cochrane’s systematic reviews’ abstracts: a survey. Therapies. 2021;76:403–8.

Vanderkam

Blanchard

Naudet

, et al. Efficacy of acetylleucine in vertigo and dizziness: a systematic review of randomised controlled trials. Eur J Clin Pharmacol. 2019;75:603–7.

Clara

Paul

Denis

Stéphanie

Hélène

Rémy

Efficacy of phloroglucinol for the treatment of pain of gynaecologic or obstetrical origin: a systematic review of literature of randomised controlled trials. Eur J Clin Pharmacol. 2020;76:1–6.

Haute Autorité de Santé. Patient et professionnels de santé: décider ensemble. 2013. Available from: https://www.has-sante.fr/upload/docs/application/pdf/2013-10/synthese_avec_schema.pdf. [Accessed 14 October, 2021].

10.

Boussageon

Blanchard

Charuel

, et al. Project rebuild the evidence base (REB): a method to interpret randomised clinical trials and their meta-analysis to present solid benefit-risk assessments to patients. Therapies. 2023;78(4):353–65.

11.

Vaillant-Roussel

Blanchard

Menini

, et al. Projet rebuild the evidence. Exercer. 2023;34:81–8.

12.

Witteman

Maki

Vaisson

, et al. Systematic development of patient decision aids: an update from the IPDAS collaboration. Med Decis Making. 2021;41:736–54.

13.

van den Haak

de Jong

MDT

Schellens

PJ.

Retrospective

vs.

concurrent think-aloud protocols: testing the usability of an online library catalogue. Behav Inf Technol. 2003;22:339–51.

14.

Lebeau

J-P.

Initiation à la recherche qualitative en santé: le guide pour réussir sa thèse ou son mémoire. Puteaux, Paris: Global Média Santé CNGE Productions; 2021.

15.

Blais

Martineau

L’analyse inductive générale: description d’une démarche visant à donner un sens à des données brutes. Rech Qual. 2006;26:1.

16.

Elwyn

O’Connor

Bennett

, et al. Assessing the quality of decision support technologies using the International Patient Decision Aid Standards Instrument (IPDASi). PLoS One. 2009;4:e4705.

17.

Vaisson

Provencher

Dugas

, et al. User involvement in the design and development of patient decision aids and other personal health tools: a systematic review. Med Decis Making. 2021;41:261–74.

18.

Martin

Brogård Andersen

O’Brien

, et al. Providing balanced information about options in patient decision aids: an update from the international patient decision aid standards. Med Decis Making. 2021;41:780–800.

19.

Ubel

Smith

Zikmund-Fisher

, et al. Testing whether decision aids introduce cognitive biases: results of a randomized trial. Patient Educ Couns. 2010;80:158–63.

20.

Gallagher

Hares

Spencer

Bradshaw

Webb

The nominal group technique: a research tool for general practice?

Fam Pract. 1993;10:76–81.

21.

Nielsen

Landauer

TK.

A mathematical model of the finding of usability problems. In: Proceedings of the SIGCHI Conference on Human Factors in Computing Systems - CHI ’93. Amsterdam (the Netherlands): ACM Press, 1993. p 206–213.

22.

Raynor

Knapp

Silcock

Parkinson

Feeney

“User-testing” as a method for testing the fitness-for-purpose of written medicine information. Patient Educ Couns. 2011;83:404–10.

23.

Légaré

Ratté

Gravel

Graham

ID.

Barriers and facilitators to implementing shared decision-making in clinical practice: update of a systematic review of health professionals’ perceptions. Patient Educ Couns. 2008;73:526–35.

24.

Zipkin

Umscheid

Keating

, et al. Evidence-based risk communication: a systematic review. Ann Intern Med. 2014;161:270.

25.

Trevena

Bonner

Okan

, et al. Current challenges when using numbers in patient decision aids: advanced concepts. Med Decis Making. 2021;41:834–47.

26.

Bonner

Trevena

Gaissmaier

, et al. Current best practice for presenting probabilities in patient decision aids: fundamental principles. Med Decis Making. 2021;41:821–33.

27.

Bedhomme

Roche

Ramin

Tauveron

Vennat

Place du pharmacien d’officine dans l’éducation thérapeutique du patient diabétique. Médecine Mal Métaboliques. 2012;6:435–40.

28.

Arrêté du 9 mars 2018 portant approbation de l’avenant 12 à la convention nationale du 4 mai 2012, organisant les rapports entres les pharmaciens titulaires d’officine et l’assurance maladie. Journal officiel de la République française[Internet]. 2018. Available from: https://www.legifrance.gouv.fr/jorf/id/JORFTEXT000036711358 [Accessed 22 June, 2024].

29.

Ordre national des pharmaciens. La pharmacie clinique. Etat des lieux et perspectives d’une discipline en développement. LES CAHIERS de l’Odre national des pharmaciens, 2018. Available from: https://www.calameo.com/read/00244939565ba5a24c060

30.

Rajiah

Sivarasa

Maharajan

MK.

Impact of pharmacists’ interventions and patients’ decision on health outcomes in terms of medication adherence and quality use of medicines among patients attending community pharmacies: a systematic review. Int J Environ Res Public Health. 2021;18:4392.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

6.81 MB