Skip to main content

Table 2 Performance comparison of alternative estimators: mean error of prediction. 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 -0.006 0.004 -0.020 -0.221 0.141 -0.198
Population-weighted average 0.996 0.491 0.303 1.654 0.720 0.562
Global Empirical Bayes 1.245 1.127 0.162 3.094 1.921 0.581
Local Empirical Bayes 0.586 0.395 0.130 1.623 0.737 0.572
Poisson kriging (true γR(h)) 0.237 0.175 0.005 1.264 0.600 0.372
Poisson kriging 0.228 0.193 0.016 1.441 0.622 0.412
CERVIX CANCER       
Observed rates 0.000 0.001 -0.003 0.248 0.059 -0.010
Population-weighted average -0.362 -0.196 -0.139 -0.406 -0.261 -0.270
Global Empirical Bayes -0.386 -0.398 -0.092 -0.717 -0.697 -0.337
Local Empirical Bayes -0.184 -0.142 -0.066 -0.331 -0.215 -0.203
Poisson kriging (true γR(h)) -0.051 -0.024 -0.034 -0.299 -0.198 -0.197
Poisson kriging -0.049 -0.019 -0.028 -0.221 -0.118 -0.167