Table 3.

Full multivariable hierarchical regression results for virtual treatment retention outcomes.

Variables			Percent no-showsa						Discharge with staff approvalb
			b		SE		P value		ORc (95% CI)		SE		P value
Clinician-level variables (Clinician, N=56; Patient, N=1018)
Prefer virtual format (vs other)			5.21		3.45		.13		0.64 (0.31-1.29)		0.23		.21
Patient load			0.05		0.06		.41		1.00 (0.99-1.01)		0.01		.91
Empathy scale
	Perspective taking	–0.60		2.74		.83		0.98 (0.56-1.74)		0.28		.96
	Empathic concern	–2.35		2.45		.34		1.04 (0.63-1.73)		0.27		.88
	Personal distress	–6.17d		2.43		.01		1.27 (0.77-2.09)		0.32		.35
Technology comfort scale			–1.48		1.69		.38		1.15 (0.81-1.62)		0.20		.44
Stress index			0.002		0.24		.99		1.00 (0.95-1.05)		0.02		.91
Resistance to change scale			1.25		1.87		.50		0.87 (0.59-1.28)		0.17		.47
Years with license			0.01		0.19		.95		1.03 (0.99-1.07)		0.02		.16
Agee			–1.02		0.95		.28		0.99 (0.81-1.20)		0.10		.90
Gender (female=1)			–0.86		2.24		.70		1.66 (1.04-2.63)		0.39		.03
White (vs another race or ethnicity)			4.13		4.09		.31		1.04 (0.44-2.42)		0.45		.94
Variance components (Clinician, N=56; Patient, N=1018)
	Clinician level	20.28				<.001		0.22 (0.08-0.60)				<.001
	Patient level	420.73				<.001

^aNo-shows are count measures and negative binomial regression was used.

^bDischarged or transferred with staff approval is binary and logistic regression was used.

^cOR: odds ratio.

^dItalic formatting indicates statistical significance at P<.05.

^eClinician age is measured categorically (1: 18-25, 2: 26-35, 3: 36-45, 4: 46-55, 5: 56-65, and 6: 65 years and older).