Table 3.

Multivariable logistic regression model for parental accurate versus inaccurate understanding of prognosis, less favorable prognosis subset^*

	Odds Ratio	95% CI	P value
Information quality (high vs low)	0.45	0.19, 1.10	0.08
Communication quality (high vs low)	1.06	0.47, 2.36	0.89
Prognostic disclosure (per element)	0.89	0.66, 1.20	0.45
Site (Philadelphia vs Boston)	0.76	0.33, 1.74	0.76
Child gender (male vs female)	1.54	0.70, 3.40	0.29
Child age			0.47
0–2	Ref	--
3–6	0.42	0.13, 1.33
7 to 12	0.50	0.15, 1.72
13 to 18	0.44	0.12, 1.64
Child diagnosis			0.88
Hematologic malignancies	Ref	--
Solid tumor	0.81	0.31, 2.10
Brain tumor	0.96	0.33, 2.81
Child’s prognosis (assessed by oncologist)			0.02
Moderately likely (50–74% chance)	Ref	--
Less than moderately likely	0.39	0.17, 0.88
Parent age			0.12
<30	Ref	--
30 to 39	3.00	0.69, 13.1
40 to 49	1.27	0.23, 6.94
50+	0.78	0.09, 7.08
Parent gender (male vs female)	0.92	0.35, 2.41	0.86
Parent ethnicity			0.06
White non-Hispanic	Ref	--
Non-white or Hispanic	0.35	0.12, 1.04
Parent education			0.89
High school graduate or less	Ref	--
College graduate	0.84	0.33, 2.18
Graduate/professional school	1.07	0.36, 3.17
Parent marital status			0.16
Married/living as married	Ref	--
Other	0.42	0.13, 1.42

^*OR >1 represent higher likelihood parental report of accurate prognosis.

ORs adjusted for parent-reported information quality, communication quality, prognostic disclosure; site; child age/gender/diagnosis; parent age/gender/race/ethnicity/education/marital status; oncologist-reported prognosis. Within-physician clustering accounted for by GEE