Table 2.

Multiple Logistic Regression Analysis: Final Models for Treatment Response at Post-assessment (i.e., Reliable Improvement on the UCLA-9 Loneliness Scale).

	Post-assessment
	WL vs. GU			AM vs. GU
	OR	95%CI	pOR	OR	95%CI	pOR
Intercept	0.78	0.28–2.14	0.62	0.99	0.60–1.65	0.98
UCLA-9 prea	0.99	0.86–1.15	0.93	1.08	0.97–1.21	0.18
Conditionb	1.40	0.37-5.47	0.62	1.55	0.77–3.15	0.23
Agea	0.97	0.94-1.00	0.06	0.98	0.95–1.00	0.08
Lsuffer a	1.89	0.80–4.79	0.16
Education	0.30	0.07–1.16	0.08
Lsuffer * conditionb	0.33	0.10–0.96	0.046
Education * conditionb	5.56	0.99–31.51	0.05
N of Obs.	115			134
Nagelkerkers R2	0.22			0.06
Hosmer-Lemeshow-Test for goodness of fit	χ2(8) = 4.46, p = 0.81			χ2(8) = 7.73, p = 0.46

Note. WL = Waitlist-control group, AM = SOLUS-D + automated messages, GU = SOLUS-D + guidance. UCLA-9 = University of California Loneliness Scale–9 item version; Lsuffer = Subjective burden of feelings of loneliness.

^aContinuous variables were centered at the grand mean.

^bReference group is WL when compared to intervention groups and AM when intervention groups are compared against each other. Condition, therefore, describes the estimate for the non-reference category.