Table 1.
Bayesian Multilevel Models of Adherence to Medication and Psychotherapy
| Model 1 | Model 2 | |||
|---|---|---|---|---|
| Estimate | 95% Cl | Estimate | 95% Cl | |
| Antidepressant | ||||
| Fixed Effects | ||||
| Intercept (b00) | 1.08 | [1.06, 1.10] | 1.26 | [1.23, 1.30] |
| Racial-Ethnic status (b10) | −0.67 | [−0.69, −0.66] | −0.44 | [−0.48, −0.39] |
| Prior Mental Health (b20) | 0.17 | [0.16, 0.19] | 0.18 | [0.16, 0.20] |
| Prior Medication treatment(b30) | 0.24 | [0.22, 0.26] | 0.24 | [0.22, 0.26] |
| Neighborhood Income (b40) | −0.13 | [−0.15, −0.12] | −0.12 | [−0.13, −0.10] |
| Proportion Minority Patients (b50) | - | - | −0.42 | [−0.50, −0.34] |
| Prop Minority X Racial-Ethnic Status (b60) | - | - | −0.39 | [−0.48, −0.30] |
| Provider-Level Random Effects | ||||
| Intercept (S0j) | 0.36 | [0.34, 0.38] | 0.35 | [0.33, 0.37] |
| R-E status (S1j) | 0.10 | [0.03, 0.16] | 0.07 | [0.00, 0.13] |
| Correlation (rj) | 0.05 | [−0.20, 0.51] | −0.25 | [−0.69, 0.26] |
| Psychotherapy | ||||
| Fixed Effects | ||||
| Intercept (b00) | −0.04 | [−0.07, 0.01] | 0.42 | [0.36, 0.48] |
| Racial-Ethnic status (b10) | −0.14 | [−0.16, −0.11] | −0.26 | [−0.31, −0.20] |
| Prior Mental Health (b20) | 0.08 | [0.06, 0.09] | 0.08 | [0.06, 0.09] |
| Prior Medication treatment (b30) | −0.14 | [−0.16, −0.12] | −0.14 | [−0.16, −0.12] |
| Neighborhood Income (b40) | −0.06 | [−.0.07, −0.04] | −0.05 | [−0.06, −0.03] |
| Proportion Minority Patients (b50) | - | - | −1.06 | [−1.18, −0.94] |
| Prop Minority X Racial-Ethnic Status (b60) | - | - | 0.29 | [0.18, 0.39] |
| Provider-Level Random Effects | ||||
| Intercept (S0j) | 0.72 | [0.69, 0.75] | 0.68 | [0.66, 0.71] |
| Racial-ethnic status (S1j) | 0.21 | [0.18, 0.24] | 0.20 | [0.17, 0.23] |
| Correlation (rj) | −0.80 | [−0.87, −0.72] | −0.78 | [−0.86, −0.70] |
95% CI 95% credible interval. Model 1 is a Bayesian mixed effects model that estimates if the effect of racial-ethnic minority status on adherence varied across providers. Model 2 is identical except that effect for the proportion of racial-ethnic minority patients in a providers caseload are added as co-variates.