From knowledge to action: a blended training to enhance shared decision making skills in primary care physicians – a pilot study Authors: Jasmien Jaeken, Pieter Van Bostraeten, Lien Mertens, Geertruida Bekkering, Bert Aertgeerts, Jacobus Jeroen Klaas Bossen, Hilde Vandenneucker, Birgitte Schoenmakers, David Spitaels, Gabriël Simon Beerten, Tiemen Peys, Cathoo Billiouw, Nicolas Delvaux Published: 2026 ABSTRACT BackgroundShared decision making (SDM) is a process to actively involve both patients and clinicians to weigh the benefits and risks of a healthcare decision, based on clinical guidelines and the patients’ preferences, needs and values. Despite the ethical foundation of SDM, its implementation remains limited. Possible physician-reported barriers for this limited uptake include insufficient level of SDM training. Training physicians in SDM could be a part of the puzzle. A recent systematic review showed that there is a shift towards blended training more than live or online learning. We therefore developed and pilot-tested a blended training program for general practitioners (GPs) in SDM in Belgium. Acquired skills were evaluated by three viewpoint – observer, patient and physician.MethodsIn a pre-post study, GPs participated in the blended training program consisting of an e-learning and a face-to-face session with simulation patients (SPs) GPs and SPs completed surveys before (T0) and after (T1) the blended training. Consultations were recorded for analysis by observer reported scales (OPTION12 and 4SDM scale). Secondary outcomes were SDM-Q9-patient, satisfaction with consultation, knowledge and intentions towards SDM.FindingsTen GPs were included. There was a significant increase in both OPTION12 (mean (SD) from 19·37 before to 37·70 after training, p = 0·0010, 95% CI [9·65 – 27·02]) and 4SDM scale (mean (SD) from 9·2 (4·66) before to 17·00 (5·08) after training, p = 0·0001, 95% CI [5·47 – 10·13]) with a moderate-large effect after training (Cohen’s D = 2·39, 95% CI [1·13 – 3·63].InterpretationThe SDM blended training for GPs improved their skills, knowledge and intentions in SDM in simulated consultations on the short term.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12875-026-03351-0. Introduction In medicine, shared decision making (SDM) is the preferred approach for decisions where there is more than one acceptable option. It is a process that actively involves both patients and clinicians to weigh the benefits and risks of a healthcare decision, based on clinical guidelines and the patients’ preferences, needs and values [1]. Approximately 60% of the physicians prefer SDM and think they already adequately involve patients, yet physicians tend to overestimate their own skills and their perceptions often do not match the patients’ ideas and preferences [2]. Although many physicians report that they already engage in shared decision making, observational studies consistently show low levels of SDM behaviour in clinical practice, indicating a gap between perceived and actual skills [2–4]. This mismatch can lead to unmet expectations of treatment outcomes and decreased patient satisfaction [5, 6]. Despite strong theoretical and ethical foundations, the implementation of SDM in primary care remains limited. Furthermore, lack of training and insufficient skills have consistently been identified as key barriers to SDM implementation across multiple studies [7–9]. In addition, previous qualitative work among GPs identified a clear need for more training in SDM and highlighted a lack of confidence in applying SDM in daily practice [10]. To tackle these barriers, numerous educational SDM interventions have been developed aimed at improving SDM knowledge, awareness and use among physicians [11–13]. They have been shown to increase the uptake of SDM by physicians with positive effects on patient and doctor satisfaction and increased patient participation. However, the quality of evidence is low due to a high heterogeneity in setting (primary vs secondary care), in teaching methods and outcome objectives which makes it difficult to identify the most effective and suitable teaching strategy [14–16]. More targeted, comparative studies with a combination of measurements of SDM skills and experiences from patient, physician and observer perspectives are needed [8, 9]. Moreover, a recent systematic review showed that there is a shift towards blended training more than live (use face to face OR in-person) or online learning [11, 17]. Blended SDM training programs have been developed and tested successfully but mainly in secondary care [18] even though previous research has shown that GPs are willing to be trained more in SDM [10, 19] but they lack confidence to engage in SDM [1, 20]. We therefore developed a blended training program for primary care physicians and evaluated its effect on the SDM skills measured from different viewpoints. The main objective of this study was (1)to evaluate whether a blended training program for GPs improves observable SDM performance in simulated consultations measured by GPs’, patients’ and observers’ perspectives, and (2) to examine whether the training improves GPs' attitudes, confidence and intentions toward SDM. Methods We performed a before-after study using a mixed-method approach (Fig. 1). We developed a blended training program and tested this using pre-post recorded consultations between a simulation patient (SP) and a GP. SDM skills in the consultations were analyzed by three observers independently using the OPTION12 and 4SDM scale. The SPs also completed the SDM-Q9-patient questionnaire.Fig. 1Study Procedure. SP = simulation patient. GP = general practitioner Study Procedure. SP = simulation patient. GP = general practitioner Recruitment Study participants in the pilot study were a convenience sample of GPs recruited in the context of obligatory continuous medical education (CME) initiatives in Flanders, Belgium. In Belgium, GPs are required to earn CME accreditation points through locally organized trainings. The participants were invited to participate by phone and e-mail by the primary researcher who is a GP herself. All participants provided informed consent to participate in the study. Study procedure Development of the training The blended training consisted of an online e-learning module (2 h) and a face-to-face session (3 h) and was based on the learning principles described by Kolb and the SDM behavioral process elements developed by Stiggelbout et al. [21]. Content of the training can be found in Table 1. Participants had to complete the e-learning prior to the face-to-face session. The face-to-face session was organized as one group session on a location central to all participants.Table 1Training detailsTraining detailse-learning (2 h) 1 Self-reflection: PPOS scale 2 Theoretical base of SDM: definition of SDM, expectations of the patients, current pitfalls and misperceptions of SDM, practical tools and decision aids based on literature. Background information on SDM was provided using short videos. Self-developed knowledge tests: Most questions were forced-response, some questions were open-ended with free text response boxes. In different parts of the e-learning, there are self-test sections designed to test participants on knowledge they should have gained during the e-learning 3 Barriers and facilitators for implementing SDM in daily practice 4 Attitudes towards SDM 5 Role of the GP in the SDM process 6 Knowledge of SDM: theory on the 4-step model of SDM using videos and example questions to exercise with SDM support 7 Evaluation of the e-learning; feedbackFace-to-face session (3 h) 1. Refreshing theoretical background of SDM: a short PowerPoint presentation with a recap of the theoretical background of SDM 2. Introducing the osteoarthritis decision aid and the 4-steps SDM model: We taught 3. them how to use the 4 stage SDM model, which is based on the 3-talk model by Elwyn (adapted by Stiggelbout) being (1) clear out there is a decision that has to be made, (2) Option talk (3) Team talk and (4) Decision talk 4. Practicing SDM with simulation patients: trios of two GPs and one SP formed a team and were divided into the roles of GP and observer(s). They played a consult concerning a patient asking for treatment options for his/her osteoarthritis. Immediately after role-playing, the other GP and the SP evaluated the role-playing by personalized feedback and discussed ideas for improvement 5. Group discussion: training satisfaction, exploring barriers and facilitators for SDM in daily practice, possible solutions to improve the uptake of SDMPPOS Patient-Practitioner Orientation Scale (Krupat et al., 2000) which is an 18 item measure of the role orientations of patients and physicians. Responses are scored on a 6 point Likert Scale (strongly disagree to strongly agree). The average score is the sum of items divided by 18, which can range from 1 to 6. A higher score indicates a preference for patient-centered care, while a lower score indicates a preference for a doctor- or disease-centered approach Table 1Training detailsTraining detailse-learning (2 h) 1 Self-reflection: PPOS scale 2 Theoretical base of SDM: definition of SDM, expectations of the patients, current pitfalls and misperceptions of SDM, practical tools and decision aids based on literature. Background information on SDM was provided using short videos. Self-developed knowledge tests: Most questions were forced-response, some questions were open-ended with free text response boxes. In different parts of the e-learning, there are self-test sections designed to test participants on knowledge they should have gained during the e-learning 3 Barriers and facilitators for implementing SDM in daily practice 4 Attitudes towards SDM 5 Role of the GP in the SDM process 6 Knowledge of SDM: theory on the 4-step model of SDM using videos and example questions to exercise with SDM support 7 Evaluation of the e-learning; feedbackFace-to-face session (3 h) 1. Refreshing theoretical background of SDM: a short PowerPoint presentation with a recap of the theoretical background of SDM 2. Introducing the osteoarthritis decision aid and the 4-steps SDM model: We taught 3. them how to use the 4 stage SDM model, which is based on the 3-talk model by Elwyn (adapted by Stiggelbout) being (1) clear out there is a decision that has to be made, (2) Option talk (3) Team talk and (4) Decision talk 4. Practicing SDM with simulation patients: trios of two GPs and one SP formed a team and were divided into the roles of GP and observer(s). They played a consult concerning a patient asking for treatment options for his/her osteoarthritis. Immediately after role-playing, the other GP and the SP evaluated the role-playing by personalized feedback and discussed ideas for improvement 5. Group discussion: training satisfaction, exploring barriers and facilitators for SDM in daily practice, possible solutions to improve the uptake of SDMPPOS Patient-Practitioner Orientation Scale (Krupat et al., 2000) which is an 18 item measure of the role orientations of patients and physicians. Responses are scored on a 6 point Likert Scale (strongly disagree to strongly agree). The average score is the sum of items divided by 18, which can range from 1 to 6. A higher score indicates a preference for patient-centered care, while a lower score indicates a preference for a doctor- or disease-centered approach Training details PPOS Patient-Practitioner Orientation Scale (Krupat et al., 2000) which is an 18 item measure of the role orientations of patients and physicians. Responses are scored on a 6 point Likert Scale (strongly disagree to strongly agree). The average score is the sum of items divided by 18, which can range from 1 to 6. A higher score indicates a preference for patient-centered care, while a lower score indicates a preference for a doctor- or disease-centered approach Based on the feedback of our previous qualitative study [10], an ‘ideal’ SDM training should be case-based because it incorporates principles of the adult learning theory and it allows direct connection of the theory with clinical practice. This allows newly acquired knowledge and skills to be applied to real-life situations that GPs encounter in daily practice. The training adopted techniques from the theory of planned behavior and used teaching methods such as instruction (e-learning and PowerPoint presentation), modelling (videos demonstrating SDM in a primary care setting) and practice (role-play with simulation patients) [22]. We decided to develop a live workshop about osteoarthritis and its evidence-based practice to teach the participants the process of SDM using a self-developed decision aid (Appendix A1). We chose osteoarthritis because it is a very common health problem in primary care, with many different treatment options. This workshop was developed in cooperation with different experts in the field: two orthopedic surgeons from the orthopedics department of UZ Leuven (JB, HV) and four GPs with an educational background (BS, ND, DS and JJ). We also asked the SPs to provide their view and feedback on the workshop. This way, we could combine the. Simulated consultation SPs were sent to the participants to present with a health problem common in primary care. An SP is a trained healthy individual who presents a specific personal, physical, social, and psychological history. SPs are extensively used as a teaching and training resource in various medical training programs. Human simulation with SPs can create a realistic patient encounter for learners to practice and enhance skills and techniques in a risk-free environment. SPs were trained in advance to teach them about the clinical case osteo-arthritis and the principles of SDM. They were also asked to provide feedback from a patient’s view about the training program and the case scenario used. We furthermore discussed what they think is important in involving patients during a clinical encounter. This first consultation was used as baseline measurement (= T0) and was held before the participants gained access to the e-learning and the face-to-face session. The second consultation was held 4–5 weeks after completing the training (= T1). The participants were aware an SP would visit them since in Belgium it is not possible to book a consult with a GP as an unknown patient. SPs were recruited from a database available at the university of Leuven. Qualitative assessment A few weeks after the second GP-SP consultation, we performed face-to-face interviews with the participants and held a focus group session with the SPs to elucidate their experiences and pitfalls with SDM in daily practice during and after the blended program. This qualitative assessment will be reported elsewhere. Data collection Our primary outcome was an observer-reported outcome using the OPTION12 and 4SDM scale. Primary outcome: observer reported outcomes Before and after the face-to-face session, SPs visited the participants to conduct a consultation about knee osteoarthritis. These consultations were video- and audio recorded for analysis using the OPTION12 and the 4SDM scale. The OPTION12 scale is a validated observer scale and consists of 12 items scored on a five-point scale from (0) the behavior was not observed to (4) the behavior is executed to a high standard. A validated Dutch version of the scale has been developed previously. Observed SDM was also assessed with the 4SDM scale, a new instrument based on the 4-stage SDM model [23]. The 4SDM (1) assesses essential elements of SDM, leaving out all non-specific communication skills, (2) assesses SDM, irrespective of who shows the behavior, thereby giving credit to both clinicians’ and patients’ initiatives, and (3) allows for the distinction between the four SDM steps. The 4SDM has eight items (two for each step) which are coded on a 4 point scale (0 “not observed” to 3 “observed and of high quality”). Although the 4SDM scale has not yet been validated within primary care settings, its underlying structure corresponds closely to the design of our training intervention, suggesting strong alignment with our outcome measures. Data collection of the video-recordings pre-training and post-training were carried out respectively in February and May 2025. Secondary outcomes Patient reported outcome SPs completed the Shared Decision-Making Questionnaire (SDM-Q9) immediately after leaving the GP. The SDM-Q9 patient consisted of a 9-item questionnaire on a six-point rating scale (0: completely disagree – 5 completely agree). Responses were given on a visual analog scale and rescaled to a score 0–100 as recommended by the authors. All other secondary outcomes are listed in Table 2.Table 2Secondary outcome measuresOutcomesMeasureSourceTheoretical knowledge of SDMSelf-developed knowledge-test about SDME-learningSDM confidenceGPs’ confidence in performing SDM scale of 0–4After the e-learning andafter the face-to-face sessionSDM intentionsGPs’ Intention to engage in SDM scale of 0–4After the e-learning andafter the face-to-face sessionSDM perceptionsICAN questionnaireGPs completed the scale after both consultations T0 and T1GPs attitudes towards patient-centered carePPOS scale: 6 point Likert ScaleQuestionnaire at the start and at the end of the e-learningSatisfaction with consultationPatient satisfaction with consultationDoctor satisfaction with consultationRespectively doctor and patient reported after consultation T0 and T1Consultation durationTime of consultationRecordings of consultationsPPOS Patient-Practitioner Orientation Scale (Krupat et al., 2000) which is an 18 item measure of the role orientations of patients and physicians. Responses are scored on a 6 point Likert Scale (strongly disagree to strongly agree). The average score is the sum of items divided by 18, which can range from 1 to 6. A higher score indicates a preference for patient-centered care, while a lower score indicates a preference for a doctor- or disease-centered approach. ICAN Table 2Secondary outcome measuresOutcomesMeasureSourceTheoretical knowledge of SDMSelf-developed knowledge-test about SDME-learningSDM confidenceGPs’ confidence in performing SDM scale of 0–4After the e-learning andafter the face-to-face sessionSDM intentionsGPs’ Intention to engage in SDM scale of 0–4After the e-learning andafter the face-to-face sessionSDM perceptionsICAN questionnaireGPs completed the scale after both consultations T0 and T1GPs attitudes towards patient-centered carePPOS scale: 6 point Likert ScaleQuestionnaire at the start and at the end of the e-learningSatisfaction with consultationPatient satisfaction with consultationDoctor satisfaction with consultationRespectively doctor and patient reported after consultation T0 and T1Consultation durationTime of consultationRecordings of consultationsPPOS Patient-Practitioner Orientation Scale (Krupat et al., 2000) which is an 18 item measure of the role orientations of patients and physicians. Responses are scored on a 6 point Likert Scale (strongly disagree to strongly agree). The average score is the sum of items divided by 18, which can range from 1 to 6. A higher score indicates a preference for patient-centered care, while a lower score indicates a preference for a doctor- or disease-centered approach. ICAN Secondary outcome measures After the e-learning and after the face-to-face session After the e-learning and after the face-to-face session Patient satisfaction with consultation Doctor satisfaction with consultation PPOS Patient-Practitioner Orientation Scale (Krupat et al., 2000) which is an 18 item measure of the role orientations of patients and physicians. Responses are scored on a 6 point Likert Scale (strongly disagree to strongly agree). The average score is the sum of items divided by 18, which can range from 1 to 6. A higher score indicates a preference for patient-centered care, while a lower score indicates a preference for a doctor- or disease-centered approach. ICAN Data analysis For measuring the level of SDM skills, the total OPTION score is generated by converting the scores to a 0 to 100 scale and then calculating the average, as recommended by the authors of the OPTION [24]. A higher score indicates a higher level of SDM. As for the 4SDM scale, the eight items were coded on a 4-point likert scale leading to a total score of 24 where again, a higher score indicates a higher level of SDM. Each item is reported separately to more clearly identify areas for improvement and, if necessary, adjust the training accordingly. All three observers were trained by using the OPTION package and 4SDM scale, that included the manuals and a pre-recorded consultation showing an ‘ideal’ and ‘less ideal’ SDM consult. They compared their ratings and resolved discrepancies. If consensus was not reached, a fourth researcher (PVB) reviewed the recordings. Researchers were blinded for T0 or T1. Consistency was determined by Cronbach’s alpha and was considered sufficient if the ICC’s and the Fleiss kappa across items were higher than 0.60 and the raters do not differ more than one point in their scores across items and consultations. The intraclass correlation was > 0.80 for both OPTION12 (0.912) and 4SDM (0.94) indicating excellent agreement between raters in their overall scoring. The Fleiss κ was 0.035 for OPTION12 and 0.024 for the 4SDM scale. Since the data met the assumption of normality, results are presented as mean ± standard deviation. We calculated means and SDs for each scale and paired t-test was used to compare the results before and after training. Cohen’s d was presented as a measure of effect size (d = 0.20 small, d = 0.50 medium, d = 0.80 large effects). Furthermore, we used a Pearson correlation to correlate the patient scores with the observed SDM scores (OPTION12 and 4SDM) and the observer scores among themselves. Data was analyzed using the statistical software R. Results A total of 10 GPs participated in our pilot in April 2025 in Flanders, Belgium. Mean age of the participants was 44 years, three were female and nine participants work in a group practice. The mean number practicing as a GP was 17 years. The participants completed two consultations in total (one before and one after the training). Primary outcome: observer reported outcome The OPTION12 scale increased from 19·37 (SD 11·78) to 37.70 (SD 16·15) and the 4SDM also significantly increased from 9·20 (SD 4·66) to 17·00 (SD 5·08). Both instruments measure highly similar constructs related to SDM behavior as there was a high correlation between the OPTION12 and 4SDM scales both before and after training (Table 3). As for the OPTION12, the highest scored items after training where item 2 and 4, the lowest scored items where item 3 and 10. As for the 4SDM, the highest scored item was item 3 and the lowest scored item 5.Table 3Observer reported outcomes (OPTION12 and 4SDM scale) pre and post training with means and standard deviations, Cohen’s D and Pearsons’ correlationOPTION12Pre-testMean (SD)Each item ranging 0–5Post-testMean (SD)Each item ranging 0–5p-valueItem 1: the clinician draws attention to an identified problem as one that requires a decision making process0·90 (0·99)2·10 (0·57)0·0010Item 2: The clinician states that there is more than one way to deal with the identified problem (‘equipoise’)1·50(0·97)2·60 (0·70)0·0032Item 3: The clinician assesses the patient’s preferred approach to receiving information to assist decision making (e.g. discussion, reading printed material, assessing graphical data, using videotapes or other media)0·00 (0·00)0·20 (0·42)0·1679Item 4: The clinician lists ‘options’, which can include the choice of ‘no action’1·80 (1·23)2·60 (0·84)0·0868Item 5: The clinician explains the pros and cons of options to the patient (taking ‘no action’ is an option)1·30 (1·06)2·70 (1·06)0·0001Item 6: The clinician explores the patient’s expectations (or ideas) about how the problem(s) are to be managed1·00 (0·94)1·70 (1·06)0·1323Item 7: The clinician explores the patient’s concerns (fears) about how problem(s) are to be managed0·30 (0·48)1·30 (1·06)0·0085Item 8: The clinician checks that the patient has understood the information0·20 (0·42)0·60 (0·97)0·2229Item 9: The clinician offers the patient explicit opportunities to ask questions during the decision making process1·20 (0·42)1·20 (0·79)1·0000Item 10: The clinician elicits the patient’s preferred level of involvement in decision-making0·10 (0·32)0·10 (0·32)1·0000Item 11: The clinician indicates the need for a decision making (or deferring) stage0·10 (0·32)1·10 (0·74)0·0038Item 12: The clinician indicates the need to review the decision (or deferment)0·90 (1·20)1·90 (1·66)0·2126TOTAL mean score/100 (SD)19·37 (11·78)37·70 (16·15)0·001095% CI [9·65–27·02]Cohen’s d = 1·5195% CI [0·57–2·42]4SDMPre-testMean (SD)Each itemranging 0–3Post-testMean (SD)Each item ranging 0–3p-valueSetting the agendaItem 1: It is stated (or re-affirmed) that a decision about management or treatment needs to be made1·0 (0·94)2·1 (0·57)0·0011Item 2: It is stated (or re-affirmed) that the decisiondepends on the values and preferences of thepatient0·9 (0.57)1·8 (0.63)0·0039Informing about optionsItem 3: The available management or treatment options are stated (or re-affirmed)1·7 (0·82)2·5 (0·71)0·0031Item 4: The pros and cons of each option are stated or re-affirmed1·2 (1·03)2·2 (0·79)0·0011Exploring values (each item range 0–3)Item 5: The patient states the outcomes that are important to him/her (values)1·1 (0·32)1·7 (0·67)0·0239Item 6: The patient states how s(h)e appraises the(characteristics of) the management or treatmentoptions1·4 (0·52)2·2 (0·63)0·0107Making a decision (each item range 0–3)Item 7: The patient expresses or confirms his/herpreference or the (provisional) lack of a preference1·10 (0·74)2·20 (1·03)0·0011Item 8: The moment of making (or deferring) the decision is explicit and decision making occurs in agreement0·89 (0·78)2·33 (1·00)0·0003TOTAL mean score/24 (SD)9·2 (4·66)17·00 (5·08)0·000195% CI [5·47–10·13]Cohen’s d = 2·39 95% CI [1·13–3·63]Pearson correlation OPTION12 versus 4SDMr(8) =.971,p < 0·00195% CI: [0·877–0·993]r(8) =.858,p = 0·00195% CI [0·497–0·966] Table 3Observer reported outcomes (OPTION12 and 4SDM scale) pre and post training with means and standard deviations, Cohen’s D and Pearsons’ correlationOPTION12Pre-testMean (SD)Each item ranging 0–5Post-testMean (SD)Each item ranging 0–5p-valueItem 1: the clinician draws attention to an identified problem as one that requires a decision making process0·90 (0·99)2·10 (0·57)0·0010Item 2: The clinician states that there is more than one way to deal with the identified problem (‘equipoise’)1·50(0·97)2·60 (0·70)0·0032Item 3: The clinician assesses the patient’s preferred approach to receiving information to assist decision making (e.g. discussion, reading printed material, assessing graphical data, using videotapes or other media)0·00 (0·00)0·20 (0·42)0·1679Item 4: The clinician lists ‘options’, which can include the choice of ‘no action’1·80 (1·23)2·60 (0·84)0·0868Item 5: The clinician explains the pros and cons of options to the patient (taking ‘no action’ is an option)1·30 (1·06)2·70 (1·06)0·0001Item 6: The clinician explores the patient’s expectations (or ideas) about how the problem(s) are to be managed1·00 (0·94)1·70 (1·06)0·1323Item 7: The clinician explores the patient’s concerns (fears) about how problem(s) are to be managed0·30 (0·48)1·30 (1·06)0·0085Item 8: The clinician checks that the patient has understood the information0·20 (0·42)0·60 (0·97)0·2229Item 9: The clinician offers the patient explicit opportunities to ask questions during the decision making process1·20 (0·42)1·20 (0·79)1·0000Item 10: The clinician elicits the patient’s preferred level of involvement in decision-making0·10 (0·32)0·10 (0·32)1·0000Item 11: The clinician indicates the need for a decision making (or deferring) stage0·10 (0·32)1·10 (0·74)0·0038Item 12: The clinician indicates the need to review the decision (or deferment)0·90 (1·20)1·90 (1·66)0·2126TOTAL mean score/100 (SD)19·37 (11·78)37·70 (16·15)0·001095% CI [9·65–27·02]Cohen’s d = 1·5195% CI [0·57–2·42]4SDMPre-testMean (SD)Each itemranging 0–3Post-testMean (SD)Each item ranging 0–3p-valueSetting the agendaItem 1: It is stated (or re-affirmed) that a decision about management or treatment needs to be made1·0 (0·94)2·1 (0·57)0·0011Item 2: It is stated (or re-affirmed) that the decisiondepends on the values and preferences of thepatient0·9 (0.57)1·8 (0.63)0·0039Informing about optionsItem 3: The available management or treatment options are stated (or re-affirmed)1·7 (0·82)2·5 (0·71)0·0031Item 4: The pros and cons of each option are stated or re-affirmed1·2 (1·03)2·2 (0·79)0·0011Exploring values (each item range 0–3)Item 5: The patient states the outcomes that are important to him/her (values)1·1 (0·32)1·7 (0·67)0·0239Item 6: The patient states how s(h)e appraises the(characteristics of) the management or treatmentoptions1·4 (0·52)2·2 (0·63)0·0107Making a decision (each item range 0–3)Item 7: The patient expresses or confirms his/herpreference or the (provisional) lack of a preference1·10 (0·74)2·20 (1·03)0·0011Item 8: The moment of making (or deferring) the decision is explicit and decision making occurs in agreement0·89 (0·78)2·33 (1·00)0·0003TOTAL mean score/24 (SD)9·2 (4·66)17·00 (5·08)0·000195% CI [5·47–10·13]Cohen’s d = 2·39 95% CI [1·13–3·63]Pearson correlation OPTION12 versus 4SDMr(8) =.971,p < 0·00195% CI: [0·877–0·993]r(8) =.858,p = 0·00195% CI [0·497–0·966] Observer reported outcomes (OPTION12 and 4SDM scale) pre and post training with means and standard deviations, Cohen’s D and Pearsons’ correlation 0·0010 95% CI [9·65–27·02] Cohen’s d = 1·51 95% CI [0·57–2·42] Item 2: It is stated (or re-affirmed) that the decision depends on the values and preferences of the patient Item 6: The patient states how s(h)e appraises the (characteristics of) the management or treatment options Item 7: The patient expresses or confirms his/her preference or the (provisional) lack of a preference 0·0001 95% CI [5·47–10·13] Cohen’s d = 2·39 95% CI [1·13–3·63] r(8) =.971, p < 0·001 95% CI: [0·877–0·993] r(8) =.858, p = 0·001 95% CI [0·497–0·966] Secondary outcomes An overview of the secondary outcomes can be found in Table 4.Table 4Secondary outcomesSecondary outcomesPre-testMean (SD)Post-testMean (SD)p-valuePatient-reported: SDM-Q-9 patient*(appendix A2)75·99 (15·04)73·33 (14·61)0·5995% CI [−13·338–8·006]Confidence in performing SDM+2·86 (0·69)3·57 (0·49)0·04795% CI [0·02–1·41]Intention to engage in SDM+3·43 (0·53)3·57 (0·53)0·3695% CI [−0·49–0·21]ICAN scale+(appendix A4)5·20 (1.04)5·18 (1.05)0.9395% CI [−0·6751–0·6250]PPOS scale*2·91 (0·12)3·15 (0·18)0·000195% CI [0·180–0·300]Satisfaction with consultation – patient+3·6 (0·52)3·3 (0·67)0·3495% CI [−0·979–0·379]Satisfaction with consultation – doctor+2·8 (0·42)3·3 (0·48)0·01595% CI [0·123–0·877]Consultation duration (in minutes)8 minRange 6–12 min11·6 minRange 5–22 minp = 0·00695% CI [1·28–5·92]+Score on a scale of 0–4, (0 = totally disagree to 4 = totally agree)* Scores on a scale 0–5 (0 = strongly disagree to 5 = strongly agree)GP general practitioner, SP simulation patient Table 4Secondary outcomesSecondary outcomesPre-testMean (SD)Post-testMean (SD)p-valuePatient-reported: SDM-Q-9 patient*(appendix A2)75·99 (15·04)73·33 (14·61)0·5995% CI [−13·338–8·006]Confidence in performing SDM+2·86 (0·69)3·57 (0·49)0·04795% CI [0·02–1·41]Intention to engage in SDM+3·43 (0·53)3·57 (0·53)0·3695% CI [−0·49–0·21]ICAN scale+(appendix A4)5·20 (1.04)5·18 (1.05)0.9395% CI [−0·6751–0·6250]PPOS scale*2·91 (0·12)3·15 (0·18)0·000195% CI [0·180–0·300]Satisfaction with consultation – patient+3·6 (0·52)3·3 (0·67)0·3495% CI [−0·979–0·379]Satisfaction with consultation – doctor+2·8 (0·42)3·3 (0·48)0·01595% CI [0·123–0·877]Consultation duration (in minutes)8 minRange 6–12 min11·6 minRange 5–22 minp = 0·00695% CI [1·28–5·92]+Score on a scale of 0–4, (0 = totally disagree to 4 = totally agree)* Scores on a scale 0–5 (0 = strongly disagree to 5 = strongly agree)GP general practitioner, SP simulation patient Secondary outcomes Patient-reported: SDM-Q-9 patient* (appendix A2) 0·59 95% CI [−13·338–8·006] 0·047 95% CI [0·02–1·41] 0·36 95% CI [−0·49–0·21] ICAN scale+ (appendix A4) 0.93 95% CI [−0·6751–0·6250] 0·0001 95% CI [0·180–0·300] 0·34 95% CI [−0·979–0·379] 0·015 95% CI [0·123–0·877] 8 min Range 6–12 min 11·6 min Range 5–22 min p = 0·006 95% CI [1·28–5·92] +Score on a scale of 0–4, (0 = totally disagree to 4 = totally agree) * Scores on a scale 0–5 (0 = strongly disagree to 5 = strongly agree) GP general practitioner, SP simulation patient Theoretical knowledge of SDM Nine of the 10 participants (90%) correctly answered that SDM is not simply informing the patient of their options and letting them choose compared to one out of ten participants (10%) before the e-learning. In the pre-test, all participants thought that in the process of SDM, the patient needs to make the decision while in the post-test, nine of the 10 participants answered correctly. In the pre-test, three of the 10 participants answered correctly to the statement “If a patient stated “you can decide for me doc”, it means you do not have to further explore the wishes and preferences of the patient” compared to all of the participants (100%) in the post-test. A full overview of all questions and answers can be found in appendix A3. Discussion This study primarily aimed to evaluate SDM from multiple perspectives (observer, patient, and physician) and to explore the extent to which these perspectives align following a blended training intervention. We noted a significant and moderate-large effect of our blended SDM training program for GPs on SDM performance in simulated conversations, measured by both the OPTION12 and 4SDM scale indicating the potential for blended training to further support the uptake of SDM in GPs. The participants felt more confident in performing SDM after the training and planned to perform more SDM in future daily practice. There was a high satisfaction with consultation for both the participants as for the SPs. Although we found a significant improvement for the OPTION12 scale, the overall scores remained low. This is in line with previous studies where untrained GPs showed a mean score of 26 compared to trained GPs with a mean score of 36 on the OPTION scale [3, 4, 25–27]. GPs performed poorly on eliciting patients’ level of involvement and assessing patients’ preferred approach to receiving information, similar to the study of Bakhit et al. [4]. On the contrary, GPs scored high on equipoise, listening the options and discussing the pros and cons, which is in line with Driever and Dillon et al. [28, 29]. Physicians tended to limit SDM to discussing the treatment options rather than paying attention to other key steps in the 4step SDM model. It has been reported that physicians think these items are inappropriate to ask a patient, nevertheless current clinical practice should go beyond solely correctly diagnosing a patient and explaining the pros and cons of a certain treatment option [30, 31]. Besides explaining the options, patient preferences should also be a key element in a consult to avoid misdiagnosed preferences and expectations. A possible explanation for these low scores could be this was one single face-to-face session and some GPs never heard of SDM. Furthermore, this could indicate that GPs should be more specifically trained in the other steps of the SDM model. Literature also suggests multiple training sessions are needed for a better and sustained SDM behavior over time [32, 33]. Another important note is that there is no clear cut-off concerning the OPTION12 scale. There is no consensus what score to obtain to consider the consultation as being a ‘good SDM’ consult. Currently, the higher the OPTION12 score, the better however, despite the significant increase, in our study we can conclude there is room for improvement in the SDM skills. As for the 4SDM scale, we also found a significant improvement with a doubled mean from 9.2/24 at T0 to a score of 17/24 at T1 which is in line with a study in an oncology setting [34]. The 4SDM scale performs comparably to the well-established OPTION12 in capturing variations in SDM performance, supporting its suitability as an outcome measure in this context. Blended training in SDM is a relatively new teaching method for GPs. To our knowledge, blended trainings in SDM for GPs exist only for disease-specific cases like antibiotics prescription, palliative care or asthma and are limited to reaction to training and self-reported learning [35–37]. One study assessed SDM using a self-developed scale showing a moderate effect of the blended training for GPs in SDM skills in a palliative care setting [8]. To the best of our knowledge, this is the first SDM blended training program for GPs where SDM behaviors are evaluated in the context of clinical practice consultations and where feedback and reflection on performance were included according to the learning circle of Kolb [21]. The SP’s perceived level of SDM was measured using the SQM-Q-9 patient. We found no significant difference after training which is in line with other studies [38–41]. The SDM-Q-9 relies entirely on patients’ self-reported perceptions, which can be influenced by memory or social desirability bias [42]. Patients do generally know what they feel about the encounter with their physician and whether they are satisfied. Consequently, patients might not be able to distinguish between the constructs of involvement and satisfaction leading to an overestimation or underestimation of the extent of SDM. It can be questioned whether patients really correctly interpret what they are asked about when using the SDM-Q-9 and whether the SDM-Q-9 patient truly elicits the patients’ level of involvement and not mainly discussing the options [43]. This is also in line with our findings of the OPTION12 where GPs scored low in eliciting patients’ level of involvement and assessing patients’ preferred approach to receiving information. Furthermore, patients’ understanding of SDM can vary across cultures, health literacy levels, and healthcare settings [44, 45]. Therefore assessment of patients’ perceptions of SDM support would help understand whether the training program indeed benefit patients in making treatment decisions. Furthermore, there is a recall bias. The SPs completed the SDM-Q-9 directly after the consultation where the observers could rewatch the recorded consultations for multiple times when completing the observer scales. At last, although we chose for SPs fitting the profile of an osteo-arthritis patient, it was still a simulated consult. Strengths and limitations Our study has several strengths. First, the development of the training was in co-creation with SPs to include a patients’ view. Second, despite the low number of participants, we found a significant improvement of SDM skills based on the OPTION12 and 4SDM scale. Next, we had three independent and blinded observer raters to objectively score both scales. The intraclass correlation coefficient (ICC) among the three independent observers was high, demonstrating a strong level of inter-rater reliability and suggesting that the scoring procedure was consistent and reproducible. The observers were also trained in using both scales before analyzing the recordings. Furthermore, SPs were trained in SDM consults on forehand and we trained the clinical scenarios with possible SDM-minded questions and answers. The use of SPs is also a great strength of this study as it allowed investigation of GP’s SDM skills under equivalent circumstances increasing the internal validity of our study [46, 47]. In addition, SPs were also asked to participate in co-creating the training. They provided their view and feedback on the training content. The GPs were also asked for their qualitative evaluation of the blended training program to obtain information regarding effective training components and any needs for modification or change (results reported elsewhere). This way, we could combine the feedback of different stakeholders to develop and optimize the training program. Furthermore, in terms of learning we both evaluated the participants’ ‘knows how’ (knowledge test) and ‘shows how’ in terms of Miller’s model of clinical competence [48]. Finally, this study combined observer, patient and doctor reported outcome measures which followed the recommendation of a triadic approach for measuring SDM [49, 50]. Our study has several limitations. First, an easily accessible sample of physicians were recruited who were willing to participate in the study. This means that it was likely that the study sample largely consisted of physicians with a particularly positive attitude towards SDM. These GPs were colleagues from the primary researcher (JJ) from different GP practices. Next, we were unable to blind the GPs and SPs. GPs knew they would be recorded and evaluated by an SP. In Belgium it is not possible to make an appointment with a GP if you are not subscribed to that particular practice. This means that the SPs had to identify themselves to make an appointment with the GPs. This could lead to measurement bias due to the Hawthorne effect and a possible overestimation of our results. Future research Muller et al. also suggested that a qualitative assessment would provide valuable input for adapting and improving a training combined with quantitative measures [51]. We organized focus groups and face-to-face interviews with the participants which we are currently analyzing. Future research should also demonstrate if the acquired skills remain over time. One year after the training session, the SPs will visit the GPs again for another consultation which will be audio-and video recorded for further analysis. At last, this pilot study showed a significant improvement of the SDM skills so implementation of this training in CME for GPs would be the next step. Conclusion This study showed that a blended SDM training for GPs significantly improved SDM behaviors, knowledge and confidence, with a high satisfaction of consultation for both GPs and SPs. Although significant, the overall OPTION scores remain low after training. It is conceivable that physicians need opportunities to train SDM multiple times and to reflect more on their daily practice for learning basic patient-centered communication skills before being able to implement SDM. Supplementary Information Supplementary Material 1. Supplementary Material 1.