Table 3.

Racial and Ethnic Differences in Total Medical Expenditure as Measured by Quantile Regression Coefficients

	Coefficient	SE
Quantile regression
0.50
Black	−0.3025	0.0197***
Hispanic	−0.3078	0.0199***
0.75
Black	−0.2259	0.0207***
Hispanic	−0.2249	0.0212***
0.90
Black	−0.1128	0.0280***
Hispanic	−0.1509	0.0288***
0.95
Black	−0.0400	0.0308
Hispanic	−0.1329	0.0316***
SQreg—test of 25th versus 75th quantile
0.25
Black	−0.3718	0.0224
Hispanic	−0.3455	0.0205
0.75
Black	−0.2258	0.0204†
Hispanic	−0.2483	0.0232†
SQreg—test of 50th versus 95th quantile
0.50
Black	−0.3272	0.0147
Hispanic	−0.3217	0.0140
0.95
Black	−0.0135	0.0388‡
Hispanic	−0.1140	0.0220‡

Notes. On log scale—proportional effects (no risk ratios).

In addition to race/ethnicity indicator variables, quantile regression models adjust for health status variables (self-reported, list of conditions, BMI, PCS-12 and MCS-12 scores, age, gender), socioeconomic status variables (poverty status, education, and marital status), and interactions between race/ethnicity and education and income. Main effects involved in interaction terms were centered by subtracting the mean value so that main effects are readily interpretable (Kraemer and Blasey 2004).

^***Significantly different at α=0.01 level.

^†Significantly different from 25th quantile at α=0.01 level.

^‡Significantly different from 50th quantile at α=0.01 level.