Table 3. Participant-Reported Quality of Shared Decision-making.
| Outcome | No./total No. (%) | Effect (95% CI) | Intracluster correlation | ||
|---|---|---|---|---|---|
| Intervention arm (n = 463) | Standard care arm (n = 459) | Clinic | Clinician/clinic | ||
| Quality of communication | |||||
| Easy to understand | 431/432 (99.8) | 422/425 (99.3) | NA | NA | NA |
| Listens carefully | 428/430 (99.5) | 427/427 (100) | NA | NA | NA |
| Shows respect | 426/428 (99.5) | 427/427 (100) | NA | NA | NA |
| Knowledge transfer scorea | |||||
| ≤3 | 24/445 (5.4) | 30/433 (6.9) | 1.01 (1.0 to 1.02)b | 0 | 0.003 |
| 4 | 76/445 (17.1) | 88/433 (20.3) | |||
| 5 | 207/445 (46.5) | 191/433 (44.1) | |||
| 6 | 138/445 (31.0) | 124/433 (28.6) | |||
| Knowledge of riskc | |||||
| Strict threshold | 30/445 (6.7) | 22/434 (5.1) | 1.4 (0.8 to 2.2)b | 0 | 0.06 |
| Liberal threshold | 49/445 (11.0) | 40/434 (9.2) | 1.3 (0.8 to 1.8)b | 0.05 | 0.10 |
| Patient-clinician decision concordanced | |||||
| Overall | 381/465 (81.9) | 369/461 (80.0) | 1.0 (0.9 to 1.1)b | 0.13 | 0.15 |
| Start or continue warfarin | 149/382 (39.0) | 139/366 (38.0) | NA | NA | NA |
| Start or continue DOAC | 196/384 (51.0) | 190/373 (50.9) | NA | NA | NA |
| Do not receive anticoagulant medication | 7/350 (2.0) | 9/450 (2.0) | NA | NA | NA |
| Start or continue aspirin | 1/500 (0.2) | 6/300 (2.0) | NA | NA | NA |
| Delay decision | 28/400 (7.0) | 24/343 (7.0) | NA | NA | NA |
| Other | 0 | 1/333 (0.3) | NA | NA | NA |
| Patient-clinician decision discordanced | 54/450 (12.0) | 56/431 (13.0) | NA | NA | NA |
| Decisional Conflict Scale score, unadjusted mean (SD)e | |||||
| Overall | 16.6 (14.4) | 17.9 (14.9) | −1.2 (−3.2 to 0.6)f | 0.06 | 0.07 |
| Informed subscale | 18.0 (16.2) | 20.7 (17.8) | −2.7 (−6.1 to 0.7)f | NA | NA |
| Values subscale | 16.6 (16.1) | 18.8 (17.1) | −2.2 (−5.2 to 0.9)f | NA | NA |
| Support subscale | 14.2 (14.9) | 14.3 (14.7) | −0.3 (−2.2 to 1.6)f | NA | NA |
| Uncertainty subscale | 18.6 (18.6) | 19.6 (19.0) | −1.1 (−3.5 to 1.4)f | NA | NA |
| Effective subscale | 15.9 (16.0) | 16.3 (16.2) | −0.7 (−2.8 to 1.4)f | NA | NA |
| Patient recommends information-sharing approach to others | 390/429 (90.9) | 378/425 (88.9) | 1.0 (0.97 to 1.1)b | 0.14 | 0.24 |
| Clinician recommends information-sharing approach to others | 396/453 (87.4) | 199/448 (44.4) | 2.1 (2.0 to 2.2)b | 0.22 | 0.52 |
| Clinician satisfied with discussion | 400/453 (88.3) | 277/448 (61.8) | 1.49 (1.42 to 1.53)b | 0.16 | 0.43 |
Abbreviations: CHA2DS2-VASc, congestive heart failure, hypertension, age ≥75 years, diabetes, previous stroke or transient ischemic attack or thromboembolism, vascular disease, age 65-74 years, and sex category; DOAC, direct oral anticoagulant; NA, not applicable.
Six questions about atrial fibrillation and anticoagulant treatment were used to assess knowledge transfer (score range, 0 to 6, with higher scores indicating greater knowledge).
Adjusted relative risk. Adjusted by treatment arm, cohort (start vs review), and stroke risk (CHA2DS2-VASc score of 1 vs ≥2 for men and 1-2 vs ≥3 for women), with the random effect of clinic and clinician.
Patients were asked to provide the number of people like them (of 100 people) whom they expected to experience a stroke within the next year. Patients’ estimates were compared with their actual CHA2DS2-VASc risk score. A correct response was considered any answer that was within either 10% (strict threshold) or 30% (liberal threshold) of the respondent’s actual CHA2DS2-VASc risk score.
Data were missing for 28 participants in the intervention arm and 34 patients in the standard care arm. Clinician and patient responses were paired; therefore, the total numbers in this category varied, as not all patients who were missing a response aligned with clinicians who were missing a response.
The Decisional Conflict Scale was used to measure decisional satisfaction. Data were missing for 31 participants in the intervention arm and 31 patients in the standard care arm.
Adjusted mean difference between the intervention and standard care arms. Adjusted by study arm, cohort (start vs review), and stroke risk (CHA2DS2-VASc score of 1 vs ≥2 for men and 1-2 vs ≥3 for women), with the random effect of clinic and clinician.