Skip to main content

Table 2 Multivariate stationary and non-stationary spatial analyses results for P. falciparum infection prevalence for the Man region, western Côte d'Ivoire.

From: Spatially-explicit risk profiling of Plasmodium falciparum infections at a small scale: a geostatistical modelling approach

Indicator

Bayesian logistic regression models

 

Stationary

Non-stationary

 

ORa

95% BCIb

ORa

95% BCIb

Age (years)

    

   6–10

1

 

1

 

   11–16

0.75

0.65, 0.87

0.75

0.65, 0.87

Socioeconomic status

    

   Most poor

1

 

1

 

   Very poor

0.90

0.71, 1.13

0.90

0.71, 1.13

   Poor

1.21

0.95, 1.51

1.21

0.95, 1.51

   Less poor

0.91

0.90, 1.15

0.90

0.71, 1.14

   Least poor

0.85

0.66, 1.08

0.84

0.65, 1.08

Sleeping under a bed net

0.92

0.72, 1.15

0.92

0.73, 1.15

Distance to health care facility

1.07

0.87, 1.29

1.04

0.82, 1.27

Annual mean NDVI

1.16

0.98, 1.38

1.17

0.98, 1.40

Mean RFE during transmission season

1.06

0.87, 1.27

1.06

0.87, 1.27

Distance to rivers

    

   <500 m

1

 

1

 

   500–999 m

1.32

0.87, 1.94

1.27

0.81, 1.89

   = 1000 m

0.75

0.48, 1.14

0.72

0.47, 1.09

ρ 1 c

0.0014

0.0003, 0.002

0.0015

0.0003, 0.002

ρ 2

  

0.0014

0.0004, 0.002

σ 1 2d

0.30

0.17, 0.49

0.23

0.10, 0.48

σ 2 2

  

0.40

0.18, 0.79

DIC e

4899.8

 

4900.1

 
  1. aOR: odds ratio
  2. bBCI: Bayesian credible interval
  3. cρ: scalar parameter representing the rate of decline of correlation with distance between points
  4. dσ2: estimate of geographic variability
  5. eDIC: deviance information criterion; a composite measure of how well the model does, i.e. a compromise between fit and complexity, with smaller DICs indicating better performance of the model