| Stage of process | Proportion of compliance adjusted mean (SE) | Least squares differences post-hoc testing | MANCOVA result | ||
| Compliance among all residents who fell (n = 205)* | (Hotelling trace 0.504) F(3,196) = 32.96, p<0.001 | ||||
| Intervention group (n = 104) | Control group (n = 101) | ||||
| Diagnosis | 0.71 (0.01) | 0.55 (0.02) | F(1,198) = 54.21, p = 0.001 | ||
| Management | 0.36 (0.02) | 0.31 (0.02) | F(1,198) = 4.77, p = 0.03 | ||
| Monitoring | 0.37 (0.02) | 0.18 (0.02) | F(1,198) = 53.34, p = 0.001 | ||
| Compliance among those with a fall history (n = 90)† | (Hotelling trace 0.537) F(3,76) = 13.60, p<0.001 | ||||
| Intervention group (n = 47) | Control group (n = 43) | ||||
| Diagnosis | 0.72 (0.02) | 0.55 (0.02) | F(1,78) = 31.46, p = 0.001 | ||
| Management | 0.39 (0.02) | 0.35 (0.02) | F(1,78) = 1.67, p = 0.20 | ||
| Monitoring | 0.39 (0.02) | 0.25 (0.03) | F(1,78) = 14.35, p = 0.001 | ||
Proportion scores were calculated as the number of compliant activities/(total possible items for the particular process stage minus items not applicable) for each resident.
*Covariates included number of falls during intervention period, age, race, fall history, marital status (eg, divorced or never married).
†Covariates included number of falls in preintervention and postintervention period, age, race, history of falls, marital status, and preintervention period recognition, diagnosis, management and monitoring audit scores.