Table 2. Logistic regression analysis of West Nile virus seroprevalence in wild mammals*.
| Predictor | Coefficient | Odds ratio (95% CI) | p value |
|---|---|---|---|
| Constant | –7.52 ± 2.75 | 0.006 | |
| Age (adult) | 7.83 ± 2.97 | 2,508.54 (7.48–8.4 x 105) | 0.008 |
| Juvenile date of capture† | 0.024 ± 0.01 | 1.02 (1–1.05) | 0.025 |
| Adult date of capture | –0.004 ± 0.005 | 1 (0.99–1.01) | 0.503 |
| UI |
0.015 ± 0.008 |
1.02 (1.0–1.03) |
0.045 |
| Species‡ | 0.007 | ||
| Tamias striatus | –1.7 ± 0.68 | 0.18 (0.05–0.69) | 0.012 |
| Didelphis virginiana | 0.46 ± 0.46 | 1.59 (0.64–3.94) | 0.32 |
| Peromyscus leucopus | –1.52 ± 0.68 | 0.22 (0.06–0.84) | 0.026 |
| Procyon lotor | 0.88 ± 0.77 | 2.41 (0.53–11) | 0.26 |
| Rattus norvegicus | –0.7 ± 0.75 | 0.5 (0.11–2.19) | 0.36 |
*Analysis used an 80% neutralization cutoff in plaque-reduction neutralization tests (PRNTs). Date refers to Julian date (January 1 = 1) and ranged from 165 (June 14) to 285 (October 15). All effects were significant when using a 90% cutoff in PRNTs at p<0.05 except urbanization index (UI) (p = 0.10). CI, confidence interval. †Squirrel samples collected in September and April at Takoma Park, MD, and Foggy Bottom, Washington DC, were significantly different (logistic regression with age, site, and month as categorical factors; September coefficient −2.22 ± 0.85; p = 0.009). ‡Species effect χ2 16.3; df 5; p = 0.007. Coefficients and odds ratios used eastern gray squirrels (Sciurus carolinensis) for the reference level.