Table 2 |
||||
|
Posterior estimates of regression coefficients and the variances related to random effects obtained by the different models |
||||
| Coefficients |
Non-Spatial Model |
Spatial Model |
Spatial Model + Non-structured random effects |
Spatial-seasonal model |
|
|
||||
| Age |
||||
| < = 1 year old |
0.49 (0.22,0.73) |
0.49 (0.22, 0.76) |
0.49 (0.24, 0.77) |
0.49 (0.24, 0.75) |
| (1 and 5 years) |
0.59 (0.35, 0.75) |
0.59 (0.35, 0.85) |
0.56 (0.35, 0.86) |
0.60 (0.34, 0.86) |
| Over 5 years |
baseline |
baseline |
baseline |
baseline |
|
|
||||
| Season |
||||
| Wet |
0.10 (-0.26, 0.45) |
0.07 (-0.30, 0.42) |
0.06 (-0.31, 0.42) |
0.29 (-0.13, 0.73) |
| Dry |
baseline |
baseline |
baseline |
baseline |
|
|
||||
| Period |
||||
| First |
-0.39 (-0.71, -0.07) |
-0.37 (-0.69, -0.05) |
-0.37 (-0.70, -0.04) |
-0.35 (-0.66, -0.02) |
| Second |
-0.03 (-0.40, 0.33) |
-0.04 (-0.42, 0.33) |
-0.04 (-0.41, 0.32) |
-0.03 (-0.42, 0.35) |
| Third |
baseline |
baseline |
baseline |
baseline |
|
|
||||
| Season × Period |
||||
| Wet & Second |
0.49 (0.02, 0.98) |
0.49 (0.03,0.96) |
0.49 (0.03, 0.98) |
0.462 (-0.02, 0.94) |
| Wet & Second |
0.76 (0.30, 1.25) |
0.82 (0.33, 1.32) |
0.82 (0.35, 1.30) |
0.792 (0.32, 1.28) |
|
|
||||
| Random effects |
||||
|
|
||||
| Spatial variation |
- |
1.29 (0.68, 0.93) |
1.26 (0.97, 1.59) |
1.67 (1.18, 2.28)1 |
|
|
||||
| 1.22 (0.93, 1.57)2 |
||||
|
|
||||
| Non-structured variation |
- |
- |
0.18 (0.06,0.39) |
- |
|
|
||||
| DIC3 |
2333.65 |
2060.16 |
2061.59 |
2070.090 |
|
|
||||
|
1spatial variance for the dry season, 2 spatial variance for the wet season, 3 DIC: Deviance Information Criterion. The smaller DIC value indicates a better fitting model. Credibility intervals at 95% in brackets |
||||
|
Abellana et al. Malaria Journal 2008 7:228 doi:10.1186/1475-2875-7-228 |
||||
