This study included two steps: (1) item formulation, and (2) validation. It is a secondary study of a larger study to design and evaluate a professional e-learning program that is reported elsewhere.20
Item Formulation
An expert panel (AMCG, EF-B, DC) developed a preliminary list of 11 items, in French, based on a review of the literature about the barriers perceived by clinicians to adopt SDM.17,18,21−28 The panel selected the barriers most often reported and wrote them as affirmative statements (Table 1). Respondents rated the degree to which they agreed with each statement on a visual analogue scale ranging from 0 (strongly disagree) to 10 (strongly agree). A total score on the IcanSDM scale can be calculated from the mean of its items, ranging from 0 to 10 with higher scores indicating higher barrier perception and thus lower ability to adopt SDM.
Table 1
IcanSDM items. The items were formulated in French. The English version is a translation that was not culturally adapted.
Retained items |
#1 | Shared decision-making results in longer clinical encounters. |
#2 | Patients often prefer that the clinician make the decision. |
#4 | Shared decision-making does not apply to all patients, nor does it apply to all clinical situations. |
#6 | Communicating scientific data to patients is too complex. |
#7 | Shared decision-making takes up too many resources. |
#8 | Shared decision-making is inconsistent with clinical practice guidelines. |
#9 | Shared decision-making is just a passing trend. |
#11 | Initial: The shared decision-making process highlights the uncertainty associated with interventions. This could affect treatment adherence. Final: During shared decision-making, the patient becomes aware of the uncertainty associated with interventions and might become confused. |
Discarded items |
#3 | Often, patients have already made their decision. |
#5 | My team and I already use shared decision-making. |
#10 | With shared decision-making, I find that many of the interventions I recommend are less effective than I thought. I prefer to continue with my usual practice. |
Validation
Study Participants
We recruited a sample of clinicians from various professions (e.g., physicians, nurses, social workers) who worked in family medicine clinics in rural regions of the Province of Québec (Canada). We presented the project during one of the scheduled clinical team meetings. Clinicians who were interested completed the informed consent document and a study entry questionnaire to assess their demographic and professional characteristics (age, profession, year of licensure).
Data Collection
Respondents answered electronic surveys before (t0) and after (t1) completion of the web-based training program on SDM. As this was done during a think-aloud session to assess the training program, we recorded their comments as they completed IcanSDM, but did not prompt them to get their impressions of the items. We transcribed these comments verbatim.
The surveys included the IcanSDM and the Belief about capabilities subscale of the CPD-Reaction.
The CPD-Reaction questionnaire is meant to measure the determinants influencing adoption of a behaviour, namely intention, social influence, beliefs about capabilities, moral norm, and attitude/beliefs about consequence.29–31 The present study reports exclusively the results of the Belief about Capabilities subscale of the CPD-Reaction, which reflect clinicians’ general beliefs about their ability to adopt SDM, and comprises three items.
Similarly to the CPD-reaction, IcanSDM is also meant to measures beliefs about capabilities. However, it is more precise than the CPD-reaction as it allows measuring a set of salient beliefs underlying this determinant, as they have been extracted from the literature.
Analyses
Content validity and item analyses
We analyzed respondents’ comments about each item as they completed the survey, looking for any mention of incomprehension and evaluating acceptance.
For each item, we also visually inspected the distribution of respondents’ responses before and after training to explore each item’s instructional sensitivity.
Internal consistency
We evaluated the scale’s internal consistency using Cronbach’s alpha coefficient for measurements made before and after exposure to e-TUDE.
We also checked item-wise consistency using partial correlation coefficients at both t0 and t1.
Sensitivity to change
We hypothesized that training primary care professionals in SDM would increase their perceived ability to adopt SDM, which should result in a lower score on the IcanSDM scale (i.e., fewer perceived barriers). To verify this hypothesis, we compared the means of participants’ total scores before and after training using the paired Student’s t test. We also visually compared the frequency distribution of answers to the pooled items before and after training.
Convergent validity
To evaluate the convergent validity, we calculated the Pearson’s correlation coefficient (r) between the total score on the IcanSDM scale and the total score on the Belief about Capabilities subscale of the CPD-Reaction. We expected a negative correlation between the two scales.
We conducted all statistical analyses with the SAS package version 9.4 (SAS Institute Inc., Cary, NC, USA). We set the statistical significance of all analyses at 0.05.