Table 2. Intention-to-Treat Analysis of Health Care Utilization Outcomes, Except Medication Initiation, for All Intervention and Control Building Residents.
| Outcome | No. (%) of residentsa | AOR (95% CI)b | |||
|---|---|---|---|---|---|
| Baseline (N = 30 buildings) for all residents | Postintervention (N = 30 buildings) for all residents | ||||
| Control (n = 1846) | Intervention (n = 1849) | Control (n = 1846) | Intervention (n = 1849) | ||
| Binary outcomes | |||||
| ED visits | 689 (37.3) | 738 (39.9) | 764 (41.4) | 782 (42.3) | 0.99 (0.93-1.06) |
| ED visits by ambulance | 368 (19.9) | 368 (19.9) | 463 (25.1) | 445 (24.1) | 0.97 (0.89-1.05) |
| Hospital admissions | 264 (14.3) | 268 (14.5) | 338 (18.3) | 326 (17.6) | 0.98 (0.88-1.09) |
| Home care services | 430 (23.3) | 449 (24.3) | 499 (27.0) | 511 (27.6) | 0.99 (0.92-1.07) |
| Transfers to long-term care | NA | NA | 83 (4.5) | 75 (4.1) | 0.90 (0.68-1.20) |
| Kidney function laboratory test | 1229 (66.6) | 1302 (70.4) | 1277 (69.2) | 1324 (71.6) | 0.99 (0.95-1.04) |
| Electrolyte laboratory test | 1071 (58.0) | 1136 (61.4) | 1095 (59.3) | 1149 (62.1) | 1.01 (0.95-1.07) |
| Lipids laboratory test | 1011 (54.8) | 1047 (56.6) | 999 (54.1) | 1055 (57.1) | 1.02 (0.98-1.08) |
| Diabetes laboratory test | 946 (51.2) | 977 (52.8) | 1000 (54.2) | 1043 (56.4) | 1.01 (0.97-1.05) |
| Continuous outcomes | |||||
| No. of primary care visits, mean (SD) | 5.94 (6.59) | 6.04 (6.27) | 6.29 (6.77) | 6.22 (6.72) | AIRR, 1.00 (0.94-1.05) |
| Hospital length of stay, mean (SD), dc | 13.05 (21.89) | 12.16 (20.35) | 18.78 (37.20) | 17.93 (25.44) | AIRR, 0.94 (0.83-1.07) |
Abbreviations: AIRR, adjusted incidence rate ratio; AOR, adjusted odds ratio; ED, emergency department; NA, not applicable.
a
Unless otherwise indicated.
b
Except for the transfers to long-term care, all AORs and AIRRs were calculated using generalized estimated equation models adjusted for clustering of residents within buildings, building pairing (trial design), and the baseline value. For the transfers to long-term care outcome, the generalized estimated equation model was adjusted for clustering of residents within buildings and the building pairing only.
c
The hospital length of stay outcome was restricted to only those with a hospital admission.