Skip to main content

Table 7 Performance comparison of alternative estimators: spread of the model of uncertainty. Results obtained on average over 100 realizations generated under two different population size scenarios and 3 types of risk map (1 = observed, 2 = smooth, 3 = random). Poisson kriging was conducted with the semivariogram estimated from the underlying risk values (true γR(h)) or the simulated mortality rates. Bold numbers refer to best performances outside the ideal case where the true semivariogram of risk is known.

From: Geostatistical analysis of disease data: estimation of cancer mortality risk from empirical frequencies using Poisson kriging

Estimators WF population BF population
BREAST CANCER Scenario 1 Scenario 2 Scenario 3 Scenario 1 Scenario 2 Scenario 3
Observed rates 5.147 5.002 4.784 22482 21577 20641
Population-weighted average 4.057 1.603 7.712 17.60 16.28 20.97
Global Empirical Bayes 2.438 1.681 2.372 31.74 30.42 27.65
Local Empirical Bayes 2.050 1.143 2.422 16.98 16.53 18.88
Poisson kriging (true γR(h)) 0.940 0.603 2.076 2.456 1.890 6.621
Poisson kriging 1.042 0.470 2.074 8.775 6.663 5.987
CERVIX CANCER       
Observed rates 0.586 0.608 0.684 2359 2372 2625
Population-weighted average 0.494 0.242 0.883 2.870 2.631 3.233
Global Empirical Bayes 0.293 0.259 0.317 4.242 4.327 4.148
Local Empirical Bayes 0.229 0.170 0.307 2.843 2.722 3.130
Poisson kriging (true γR(h)) 0.155 0.043 0.307 0.763 0.711 1.201
Poisson kriging 0.169 0.069 0.269 1.552 1.825 1.529