. 2014 Apr 1;17(4):435–447. doi: 10.1089/jpm.2013.0214

Table 3.

Race and Socioeconomic Status: Associations with Ratings of End-of-Life Care and Communication^a

		Patient ratings		Family ratings
Outcome	Predictors	b	p	b	p
QEOLC	Evaluator race/ethnicity^b		0.500		0.006
	African American	−0.180		−0.646
	Other minority	−0.132		−0.948
	Evaluator education^c	−0.019	0.731	−0.162	0.098
	Evaluator income^d	0.042	0.320
QOCgen	Evaluator race/ethnicity^e		>0.500		0.009
	African American	−0.136		−0.567
	Other minority	−0.006		−0.935
	Evaluator education^f	0.010	0.861	−0.086	0.370
	Evaluator income^g	0.031	0.436
QOCeol	Evaluator race/ethnicity^h		0.026		0.110
	African American	0.831		1.088
	Other minority	0.516		−0.041
	Evaluator educationⁱ	−0.273	0.006	−0.687	<0.001
	Evaluator incomeⁱ	−0.390	<0.001

^{^a}

Associations were based on cross-classified (evaluator by trainee) hierarchical models, using full maximum likelihood estimation, with continuous linear outcomes and with level-1 slopes fixed over cross-classified level-2 units.

^{^b}

Patient model adjusted for recruitment site; trainee age, postgraduate year, and race; patient gender, SF1 health rating, inpatient vs. outpatient site of patient-trainee interaction, and four diagnoses (cancer, congestive heart failure, end-stage liver disease, and AIDS). Family model adjusted for recruitment site.

^{^c}

Patient model adjusted for recruitment site and inpatient/outpatient. Family model adjusted for patient race and recruitment site.

^{^d}

Model adjusted for two patient diagnosis indicators (cancer and AIDS), education, SF1 health rating, inpatient/outpatient, recruitment site, and the trainee's postgraduate year.

^{^e}

Patient model adjusted for recruitment site; trainee's postgraduate year and race; one patient diagnosis (AIDS), inpatient/outpatient, and SF1. Family model adjusted for recruitment site.

^{^f}

Patient model adjusted for recruitment site and inpatient/outpatient. Family model adjusted for recruitment site, trainee age, and family race.

^{^g}

Model adjusted for recruitment site, trainee's post-graduate year, patient education, SF1, and inpatient/outpatient.

^{^h}

Patient model adjusted for patient age and inpatient/outpatient. Family model adjusted for recruitment site; trainee age, postgraduate year, and race; family gender, age, and inpatient/outpatient.

^ⁱ

Patient model for education adjusted for patient race. There were no confounders of the family education or patient income model.

QEOLC, Quality of End-of-Life Care Questionnaire; QOCgen, general communication score from the Quality of Communication questionnaire; QOCeol, end-of-life care communication score from the Quality of Communication questionnaire.