Table 2. Two-part multivariate model disentangling covariates associated with the presence and increased incidence of Lyme borreliosis, France, 2016–2019.
| Variable | Unit/category | Risk coefficients | 95% Crla |
|---|---|---|---|
| Logistic model (LB presence vs absence) b | |||
| NDVI (Q) | < 0.6 | Ref. | |
| ≥ 0.6 | 1.37 | 1.25–1.51 | |
| Rodent species richness | > 1 to ≤ 3.27 | Ref. | |
| > 3.27 to ≤ 5 | 0.83 | 0.70–0.99 | |
| Gamma model (Increase of LB incidence conditional of LB presence) c | |||
| Index of deer presence | ≤ 40% | Ref. | |
| > 40 to ≤ 60% | 1.04 | 0.97–1.12 | |
| > 60 to ≤ 80% | 1.13 | 1.03–1.24 | |
| > 80% | 1.25 | 1.11–1.41 | |
| ST (Q) | ≤ 7 °C | Ref. | |
| > 7 to ≤ 15 °C | 1.04 | 0.97–1.10 | |
| > 15 to ≤ 22 °C | 1.18 | 1.11–1.25 | |
| > 22 °C | 0.76 | 0.69–0.82 | |
| SD (Q) | ≤ 1.5 mmHg | Ref. | |
| > 1.5 to ≤ 3 mmHg | 1.08 | 1.02–1.15 | |
| > 3 to ≤ 5 mmHg | 1.18 | 1.05–1.31 | |
| > 5 mmHg | 1.06 | 0.94–1.20 | |
| Frequency of tick bite reports (Q) | ≤ 0.05% | Ref. | |
| > 0.05 to ≤ 0.25% | 1.20 | 1.14–1.27 | |
| > 0.25% | 1.35 | 1.24–1.46 | |
Crl: credible interval; LB: Lyme borreliosis; NDVI: normalised difference vegetation index; OR: odds ratios; Q: quarter; RR: relative risk; SD: saturation deficit; ST: soil temperature.
a The risk coefficients (odds ratio and relative risk, respectively for the logistic and gamma model) and their 95% Crl were obtained by exponentiating the coefficients’ mean value of the posterior marginal distribution and their 95% Crl.
b The risk coefficient refers to the odds ratio in the logistic model, representing the ratio of the odds of an LB case occurring in the current category compared to the odds of an LB case occurring in the reference category.
c The risk coefficient refers to the relative risk in the gamma model, representing the ratio of the probability of an LB case occurring in the current category compared to the reference category.