Table 3.
Multiple linear regression models for delayed recall score.a
Independent variable |
R2 | Coefficient (SE) |
P | R2 | Coefficient (SE) |
P | R2 | Coefficient (SE) |
P |
---|---|---|---|---|---|---|---|---|---|
Total B12 | <0.001 | 0.0003 (0.001) | NS | 0.316 | −0.001 (0.001) | NS | 0.276 | −0.001 (0.001) | NS |
HoloTC | 0.003 | 0.005 (0.003) | NS | 0.315 | 0.0001 (0.002) | NS | 0.276 | 0.001 (0.002) | NS |
HoloTC/B12 | 0.003 | 7.97 (0.29) | NS | 0.316 | 1.24 (0.88) | NS | 0.276 | 0.921 (0.85) | NS |
Model 1, univariate analysis; model 2, model 1 plus age, sex, years of education, diabetes, stroke, homocysteine, and creatinine; model 3, model 2 plus CES-D score (< or ≥16). R2 values are for the entire model including the B12 status measure and all covariates. Coefficients (SE) and P values are for each B12 status measure within the model. Sample sizes are smaller in models 2 and 3 due to missing values among the covariates. NS, not significant (P > 0.05).