Distance dependency of childhood leukaemia risk near the Mol-Dessel site
The results for rate ratios (RRs) – that is, the ratio of the rates in four distance zones to the rates in all of Belgium – are presented in table 1 in their study. Unfortunately, the distance zones are defined as 0–5, 0–10, 0–15 and 0–20 km, and hence each distance zone includes all former zones. To find out whether there is a distance dependency of leukaemia risk, it would have been more informative to report RRs for the individual distance zones.
Figure 1 shows the RRs for distance zones 0–5, 5–10, 10–15 and 15–20 km around the Mol-Dessel site. The error bars are 1 SD. Marked increases are observed in the 0–5 and 10–15 km zones, but not in the 5–10 km zone. On using a Poisson test, the increases in the RRs in the individual distance zones are not statistically significant, but the increase in the first three distance zones combined (0–15 km) is statistically significant (observed=15, expected=7.9, RR=1.90, P=0.016, one-sided test). Interestingly, there is a marked decrease in leukaemia risk in the 15–20 km zone where the rate is about half the Belgian average (RR=0.52).
No increased leukaemia incidence is found in any of the distance zones near the remaining three NPP sites combined (Doel, Tihange, Fleurus); the overall RR in the 0–20 km zone is equal to 0.93.