Table 4. Multivariable model predicting LDL cholesterol^*.

Model covariates	Unstandardized beta estimate (95% CI)
(Constant)	127.9 (108.6, 147.3)¶
Female gender†	6.5 (2.1, 10.8) ‖
Age, yrs†	-0.5 (-0.7, -0.2)¶
Education level, yrs‡	-0.2 (-0.8, 0.4)
Insurance type (ref: commercial insurance)†
Uninsured	4.6 (-2.9, 12.2)
Medicaid	-4.5 (-10.4, 1.4)
Medicare	-1.5 (-7.0, 4.0)
Race/Ethnicity (ref: non-Hispanic white) ‡
Hispanic	-1.8 (-8.6, 4.9)
Vietnamese	-9.9 (-16.9, -2.8)‖
History of heart disease†	-8.3 (-14.2, -2.5) ‖
Other comorbidity (Total Illness Burden Index score)‡	0.2 (-0.8, 1.1)
Body mass index†	-0.1 (-0.2, 0.1)
Lipid regimen intensity (Number of classes of lipid lowering medications) †	-3.8 (-7.3, -0.3)§
Nonadherence related to cost‡	4.6 (-0.3, 9.6)
Nonadherence related to side effects‡	6.3 (2.0, 10.7) ‖

^*Results from a linear regression model predicting LDL cholesterol level (R² = .12). Unstandardized beta estimates can be interpreted as the mean difference in LDL cholesterol associated with a one unit change in a given model covariate, adjusted for all other model covariates.

^†Derived from medical record abstraction.

^‡From patient self-report in the baseline questionnaire.

^§p<0.05;

^‖p<0.01;

^¶p<0.001