Table 2. Minimally adjusted and fully adjusted results on primary, secondary, and exploratory outcomes.
| Intervention Arms | Minimally Adjusted Model* | Fully Adjusted Model** | ||||
|---|---|---|---|---|---|---|
| Outcomes | Intervention (n = 611) | Control (n = 615) | Estimate (95% CI) † | p-value | Estimate (95% CI)† | p-value |
| Primary Outcome | ||||||
| Change in systolic blood pressure, mean (SD), mm Hg | −7.1 (18.5) | −4.3 (18.9) | −2.8 (−4.8, −0.9)‡ | 0.005 | −3.3 (−5.2, −1.4) | 0.001 |
| Secondary Outcomes | ||||||
| Change in diastolic blood pressure, mean (SD), mm Hg | −3.9 (9.6) | −2.3 (9.6) | −2.2 (−3.2, −1.3) | <0.001 | −2.34 (−3.3, −1.4) | <0.001 |
| Change in health-related quality of life score, mean (SD)§ | 0.01 (0.15) | −0.03 (0.14) | 0.04 (0.01, 0.06) | 0.006 | 0.04 (0.01, 0.06) | 0.008 |
| Change in physical activity, mean (SD), MET min/wk | 1,203.9 (2,243.7) | 750.9 (2,097.2) | 528.2 (286.3, 770.1) | <0.001 | 490.2 (244.1, 736.3) | <0.001 |
| Timed up and go (time of completion ≥14 s)¶ | 256 (43.7%) | 298 (50.9%) | 0.87 (0.77, 0.98) | 0.023 | 0.87 (0.77, 0.98) | 0.022 |
| Medication adherence in Antiplatelets, n (%)†† | 308 (69.7%) | 244 (66.7%) | 1.03 (0.93, 1.14) | 0.614 | 1.02 (0.92, 1.14) | 0.658 |
| Medication adherence Statins, n (%)†† | 133 (77.3%) | 112 (62.9%) | 1.21 (1.06, 1.38) | 0.005 | 1.23 (1.07, 1.40) | 0.003 |
| Medication adherence Antihypertensives, n (%)†† | 383 (73.7%) | 315 (66.5%) | 1.10 (1.00, 1.22) | 0.051 | 1.11 (1.00, 1.22) | 0.039 |
| Exploratory Outcomes | ||||||
| Stroke Recurrence, n (%) | 27 (4.4%) | 57 (9.3%) | 0.46 (0.32, 0.66) | <0.001 | 0.45 (0.31, 0.66) | <0.001 |
| Stroke hospitalization in the past year, n (%) | 27 (4.4%) | 57 (9.3%) | 0.45 (0.32, 0.64) | <0.001 | 0.44 (0.31, 0.64) | <0.001 |
| Moderate to severe disability, n (%)‡‡ | 128 (20.9%) | 186 (30.2%) | 0.65 (0.53, 0.79) | <0.001 | 0.67 (0.55, 0.81) | <0.001 |
| Death, n (%)§§ | 11 (1.8%) | 19 (3.1%) | 0.52 (0.28, 0.96) | 0.036 | NA§§ | NA |
CI, confidence interval; MET, metabolic equivalents; NA, not applicable; SD, standard deviation.
*Prespecified main analysis (minimally adjusted model): Adjusted for baseline outcome, township, sex, and age; removing outliers in the outcome variable (based on a priori decision to remove those that are more than 2 interquartile range above the third quartile or below the first quartile).
**Sensitivity analysis (fully adjusted model): Adjusted for baseline outcome, township, sex, age, variables noted to be differential by treatment arm at baseline (baseline diastolic blood pressure, having hypertension, having none of the assets asked about, taking antihypertensive medications), and loss to follow-up (baseline systolic blood pressure, annual household income, type of phone owned and smoking status); removing outliers in the outcome variable (based on a priori decision to remove those that are more than 2 interquartile ranges above the third quartile or below the first quartile).
†For continuous outcomes (systolic blood pressure, diastolic blood pressure, EQ5D-5L, physical activity), “estimate” refers to the differences between the arms in mean 1-year change in the outcome (control arm is the reference); for binary outcomes (timed up and go, medication adherence, stroke recurrence, stroke hospitalization, disability, and death), “estimate” refers to the risk ratio (control arm is the reference).
‡The intercluster coefficient is less than 0.001 for the model.
§Health-related quality of life was measured by using EQ5D-5L and was converted into a utility score based on the Chinese value set.
¶“Timed up and go test” results were recorded in seconds during measurement and dichotomized into binary as ≥14 (indicating lower limb mobility) versus <14 s (higher limb mobility) based on previous literature.
††Medication adherence refers to a perfect adherence with score of 0 based on the 4-item Morisky Green Levine Scale. Medication adherence was only measured among participants who were taking medicines. Medication adherence outcomes were not adjusted for baseline outcome, since the set of participants taking a given medication at baseline was not the same set taking the medicines at follow-up.
‡‡Disability was measured by the modified Rankin Scale, and people who received a score above 3 were grouped into the "moderate to severe disability" group.
§§The statistical model with death as the outcome was not adjusted for variables differential by the loss to follow-up, since those who died during the study were a subset of the group lost to follow-up.