Table 2.

Logistic regression estimates after complete case analysis and multiple imputation

	Complete-case analysis (N = 9164)			Multiple imputation (N = 14,607)
	Coefficient	SE	P-value	Coefficient	SE	P-value
Cluster 2 versus Cluster 1	−1.26	0.19	< 0.0001	− 1.22	0.13	< 0.0001
Males versus females	0.38	0.12	0.0009	0.39	0.08	0.0202
Age, yrs	0.014	0.005	0.0026	0.009	0.004	0.0257
At least 1 comorbidities versus none	0.18	0.12	0.1205	0.12	0.08	0.1754
Professional field			0.0005			< 0.0001
(Cardio-respiratory as reference)	–	–	–	–	–	–
Geriatric	−0.43	0.32	0.3881	−0.03	0.23	0.9111
Neurologic	−0.26	0.3	0.0105	−0.29	0.23	0.2017
Orthopedic-Musculoskeletal	−0.72	0.28	0.1041	−0.75	0.22	0.0006
Urogynecologic and Oncologic	−1.68	1.04	0.7709	−0.29	0.43	0.5095
Pediatric	−0.11	0.38	0.9873	−0.25	0.29	0.3921
Mixed	−0.004	0.25	0.9158	−0.032	0.2	0.8694
Working Facilities			< 0.0001			< 0.0001
(Private/public hospitals as reference)	–	–	–	–	–	–
Residential Care Home	0.02	0.22	0.9158	0.22	0.15	0.1472
Private/public rehabilitation clinics	−0.41	0.17	0.0137	−0.35	0.12	0.0037
Home	−1.55	0.4	< 0.0001	−1.23	0.25	< 0.0001
Mixed	0.44	0.16	0.0047	0.17	0.12	0.1748
Private setting	−1.47	0.27	< 0.0001	−1.69	0.21	< 0.0001
Reallocation to a different unit (Yes versus No)	0.27	0.16	0.0862	0.26	0.16	0.1078
Changing job task (Yes versus No)	0.21	0.16	0.1794	0.29	0.16	0.0746
Constant	−3.3	0.32	< 0.0001	−2.7	0.25	< 0.0001

Estimates are expressed as log odds

Cluster 1 = Piedmont and Aosta Valley, Liguria, Lombardy, Veneto, Friuli-Venezia-Giulia, Trentino-Alto-Adige, Emilia-Romagna, Tuscany, Marche and Umbria; Cluster 2 = Abruzzo, Lazio, Molise, Campania, Apulia, Basilicata, Calabria, Sicily and Sardinia

SE Standard Error