TABLE 3.

Adjusted Odds Ratio for Receipt of New Prescription Medication Advising Among Spanish-Speaking Latinos Adjusting for Interpreter Use, Sociodemographic Characteristics, Health Status, and Clinical Site

	Adjusted Odds Ratio (95% Confidence Intervals)
Patient Characteristics	Explained Purpose of Medication	Explained Medication Side Effects	Explained Medication Directions	Written Medication Information From Pharmacy in Spanish	Medicine Bottle Label Written in Spanish
Interpreter not  needed	1.00	1.00	1.00	1.00	1.00
Interpreter needed,  used	3.55* (2.28, 5.51)	1.81† (1.20, 2.73)	2.50* (1.62, 3.87)	2.12† (1.37, 3.29)	2.73* (1.74, 4.28)
Interpreter needed,  not used	1.72 (0.96, 3.06)	1.37 (0.79, 2.40)	1.83‡ (1.02, 3.30)	1.66 (0.92, 3.00)	1.81 (0.99, 3.33)

^*P < 0.001.

^†P < 0.01.

^‡P < 0.05.

Controlling for age, gender, marital status, education, income, insurance status, self-reported health status, survey year, and site of care using random-effects models; Likelihood ratio test of rho = 0 is significant for all models (P < 0.05). A comparison with fixed-effects (using indicator variables for 7 of the 8 sites in regression equation) yielded similar results. Both fixed effects and random effects coefficients yielded consistent results under Hausman-Wu test. We only report results from more efficient random effects models.