Table 2. Spatial autocorrelation in the incidence of childhood acute leukaemia in France (1990–2000) – Moran's statistic I (Elliott et al, 2000).

	Whole period 1990–2000				1990–1994				1995–2000
	No. cases	I	Stata	P-valueb	No. cases	I	Stata	P-value	No. cases	I	Stata	P-value
Age at diagnosis – d=40 km
0–14 years	4873	0.006	1.78	0.04	2226	0.006	1.72	0.04	2647	0.002	0.69	0.26
0–4 years	2458	−0.004	−0.97	0.84	1136	−0.001	−0.29	0.63	1322	−0.002	−0.62	0.76
5–9 years	1427	−0.001	−0.10	0.56	662	−0.001	−0.14	0.56	765	0.001	0.24	0.38
10–14 years	988	0.001	0.50	0.28	428	0.004	1.31	0.11	560	−0.002	−0.39	0.65
Spatial limit of proximity d – Analyses in the whole age group (0–14 years)
20 km	4873	0.006	0.95	0.15	2226	0.009	1.47	0.05	2647	−0.001	−0.09	0.54
30 km	4873	0.002	0.47	0.31	2226	0.005	1.10	0.11	2647	−0.002	−0.46	0.68
40 km	4873	0.006	1.78	0.04	2226	0.006	1.72	0.04	2647	0.002	0.69	0.26
50 km	4873	0.003	1.04	0.16	2226	0.002	0.67	0.26	2647	−0.002	−0.42	0.65

^aNormalized statistic [I−E(I)]/σ(I) with E(I)=−1/(n−1)=−0.0003 (n=3644 cantons).

^bThe statistical thresholds were evaluated with the one-sided tail probability of the distribution of I under the null hypothesis (based on 999 Monte Carlo simulations).

Bold highlights the main results, those associated to a P-value <5%.