Table 4.

Effect of Restorative Care on Change in Activities of Daily Living (ADL) Dependency Over 18 Months Controlling for Resident and Nursing Home Confounders

Characteristic	Coefficient	SE	p value
Intercept	−1.6	.42	N/A
Resident Traits
Age in years	.01	.003	.15
Length of stay in years	<.0001	<.0001	.89
Cognitive Performance Scale score (range 0–6)	.38	.03	<.001
Frailty (range 0–15)	.63	.01	<.001
Number of disabling diseases (range 0–10)	.10	.02	<.001
Mood scale (range 0–14)	−.16	.02	<.001
Social engagement (range 0–6)	−.39	.02	<.001
Pain: mild (reference: severe)	−.57	.20	.004
Pain: moderate (reference: severe)	−.37	.20	.07
Pain: none (reference: severe)	−.55	.20	.006
Number of physical impairments (range 0–4)	.97	.03	<.001
Activities of daily living baseline score (range 0–28)	.68	.01	<.001
Nurse indicated resident did not have the ability to improve ADL dependency	−.13	.07	.06
Received any restorative care over time	.09	.06	.12
Nursing home traits
Percentage of residents with Medicare reimbursement	−.07	.22	.76
Percentage of residents with Medicaid reimbursement	−.02	.08	.75
Hours of patient contact with nursing staff	−.0001	.0003	.68
Medical Director has no certification	.19	.09	.04
Director of Nursing has no certification	.07	.07	.31
Facility has no accreditation	−.11	.10	.29

Notes: The model included the resident ID as an indication of which records were repeated on the same subject. State of nursing home location was included as a random effect. If the sign of the coefficient is positive for continuous variables, then as the value of the predictor increases, so does the level of ADL dependency over time. For categorical variables a positive coefficient is associated with increasing ADL dependency in comparison to the reference category for the categorical variable.