Table 3. Comparison of 5 different non-spatial and spatial models for S. japonicum infections based on Kato-Katz examination showing the importance of including spatial correlation in the analyses as well as the inclusion of the different demographic, reservoir and environmental covariates.
| Covariate group | Covariate | Covariate category | Bayesian non-spatial | Bayesian spatial | ||||||||
| Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | ||||||||
| ORa | BCIb | ORa | BCIb | ORa | BCIb | ORa | BCIb | ORa | BCIb | |||
| Demography | ||||||||||||
| Age group (years) | ||||||||||||
| 0–10 | 1 | 1 | 1 | |||||||||
| 11–20 | 1.36 | 0.69, 2.65 | 1.43 | 0.70, 2.61 | 1.47 | 0.75, 2.68 | ||||||
| 21–30 | 2.44 | 0.89, 5.84 | 2.64 | 0.95, 5.90 | 2.80 | 1.08, 6.13 | ||||||
| 31–40 | 3.06 | 1.16, 7.29 | 3.31 | 1.22, 7.31 | 3.51 | 1.40, 7.60 | ||||||
| 41–50 | 4.22 | 1.58, 10.05 | 4.56 | 1.69, 10.07 | 4.85 | 1.94, 10.55 | ||||||
| >50 | 4.47 | 1.67, 10.69 | 4.82 | 1.79, 10.50 | 5.13 | 2.09, 11.05 | ||||||
| Sex | ||||||||||||
| Female | 1 | 1 | 1 | |||||||||
| Male | 3.18 | 2.70, 3.75 | 3.18 | 2.70, 3.74 | 3.20 | 2.71, 3.77 | ||||||
| Occupation | ||||||||||||
| Herdsman, farmer, fisherman or boatman | 1 | 1 | 1 | |||||||||
| Preschool or student | 0.48 | 0.23, 0.91 | 0.50 | 0.24, 0.93 | 0.52 | 0.26, 0.96 | ||||||
| Civil servant or businessman | 0.40 | 0.11, 0.93 | 0.40 | 0.12, 0.92 | 0.41 | 0.12, 0.94 | ||||||
| Other | 0.62 | 0.25, 0.91 | 0.63 | 0.27, 1.17 | 0.63 | 0.26, 1.17 | ||||||
| Reservoir | 1 | 1 | ||||||||||
| Presence of infected buffaloes in the village | 5.34 | 1.14, 16.03 | 7.04 | 1.56, 24.23 | ||||||||
| Environment | ||||||||||||
| Mean NDVI during winter | 0.74 | 0.31, 1.69 | ||||||||||
| Distance to Dongting Lake (km) | 1.22 | 0.55, 2.33 | ||||||||||
| Endemic types | ||||||||||||
| Lake embankment | 1 | |||||||||||
| Inside embankment | 2.15 | 0.15, 9.76 | ||||||||||
| Lake fork | 0.51 | 0.03, 2.20 | ||||||||||
| Grassy lake beach/marshland | 13.49 | 0.43, 75.17 | ||||||||||
| Hills/hilly area | 16.07 | 0.01, 57.75 | ||||||||||
| u c | 15.21 | 2.51, 137.50 | 19.96 | 2.64, 185.50 | 19.47 | 3.11, 175.40 | 11.26 | 1.58, 87.00 | ||||
| σ2 d | 6.35 | 3.08, 13.32 | 6.50 | 3.26, 13.22 | 5.92 | 2.88, 11.43 | 7.29 | 2.98, 18.89 | ||||
| DICe | 8705 | 7149 | 6652 | 6652 | 6649 |
a
OR: odds ratio.
b
BCI: Bayesian credible interval.
a
u is scalar parameter representing the rate of decline of correlation with distance between points.
b
σ2 is the estimate of the geographic variability.
c
DIC is the measure for the model fit. A smaller DIC indicates a better performance of the model.