Multivariable Logistic Regression of Family Physician Adoption of PCV7 Recommendations
| Model/Variables | Odds Ratio | 95% Confidence Interval |
|---|---|---|
| Model 1—Practice characteristics* | ||
| Number of newborns annually (>40)† | 3.64 | (1.20, 11.01) |
| Prop. patients on Medicaid (≥0%)† | 2.26 | (1.27, 4.04) |
| Number of physicians (≥4)† | 2.19 | (1.21, 3.98) |
| Prop. patients African American (≥10%)† | 2.13 | (1.10, 4.15) |
| Female physician | 1.62 | (0.82, 3.22) |
| Model 2—Clinical experiences & attitudes‡ | ||
| Considers giving ≥5 injections at 1 visit† | 17.29 | (6.35, 47.05) |
| Considers giving 4 injections at 1 visit† | 8.69 | (4.21, 17.94) |
| Thinks PCV7 will effectively prevent pneumococcal meningitis | 1.86 | (0.93, 3.73) |
| Has experienced severe patient morbidity or mortality from pneumococcal disease | 1.45 | (0.66, 3.17) |
| PCV7 approach influenced by rotavirus vaccine | 1.09 | (0.55, 2.17) |
| Reports ≥3 problems with multiple injections† | 0.43 | (0.22, 0.86) |
* Reference group is male family physician with <4 physicians in practice, <10% of patients on Medicaid, <10% African American patients, who enrolls ≤40 newborns annually in her/his panel.
† Statistically significant association with PCV7 adoption.
‡ Reference group is family physician who will consider giving <4 vaccine injections at 1 visit, does not think PCV7 will be effective in preventing pneumococcal meningitis, has not experienced patient morbidity and/or mortality from pneumococcal disease, was not influenced by rotavirus vaccine, and reports <3 problems with multiple injections.