Table 7 Comparison of results considering the association between network and physician-level characteristics on a physician’s likelihood to participate using both a fixed effects and a random effects framework
Fixed effects model | Random effects model | |||
|---|---|---|---|---|
Term | OR (95% CI) | p-value | OR (95% CI) | p-value |
Age\(^\dagger\) | 0.97 (0.94, 0.99) | 0.04 | 0.96 (0.93, 0.99) | 0.01 |
AgeObserved\(^\ddagger\) | 5.16 (1.06, 25.17) | 0.04 | 6.16 (1.37, 27.81) | 0.02 |
Years with physician organization | 1.63 (1.20, 2.20) | <0.01 | 1.57 (1.18, 2.11) | 0.02 |
\(H_{i}\) | 5.00 (1.27, 19.73) | 0.02 | 7.28 (1.97, 26.96) | 0.03 |
\(WY_{wi,i}\) | 1.05 (1.03, 1.06) | <0.01 | - | - |
\(WY_{ac,i}\)
| 0.99 (0.97, 1.02) | 0.54 | 1.00 (0.98, 1.02) | 0.99 |
\(H_i\times WY_{ac,i}\) | 1.09 (0.97, 1.22) | 0.13 | 1.12 (1.00, 1.25) | 0.05 |
-
Across participation rate is ln (natural log) transformed and taken with respect to physician step - \(\dagger\) The effect of age is estimated for the subgroup of physicians whose age is observed; this is enabled by representing Age in the model as Age \(\times\) AgeObserved
- \(\ddagger\) AgeObserved is the indicator variable equal to 1 if age is observed and 0 otherwise
Across participation rate is ln (natural log) transformed and taken with respect to physician step