Logistic Regression of Factors (including flu shot status in 1998) Significantly Associated with Flu Shot in the Subsequent Year (1999)*†
| Variable | Beta | Wald Statistic | df | P Value |
|---|---|---|---|---|
| Flu shot in 1998 | 4.073 | 2101 | 1 | <.01 |
| No. of primary care provider claims in 1998 | −0.022 | 24 | 1 | <.01 |
| HMO member in 1998 | 0.470 | 16 | 1 | <.01 |
| Level of education | −0.075 | 12 | 1 | <.01 |
| African American | −0.382 | 6 | 1 | <.05 |
| Current smoker | −0.321 | 6 | 1 | <.05 |
| Disease burden | −0.059 | 4 | 1 | <.05 |
| Constant | −4.661 | 124 | 1 | <.01 |
* Flu shot: yes = 0, no = 1; for all categorical variables (African American, Hispanic ancestry, smoking, married?, HMO membership, living in urban area) yes = 0, no = 1.
† Variables not statistically significant in the model: total number of physician claims, marital status, age, Hispanic ancestry, household income, self-reported health status, level of disability, living in an urban area, and number of specialist physician claims. R square of model = 0.64.