TABLE 1—

Percentages of Significant SaTScan Clusters for Points Analyzed as Either Exact Coordinates or Centroids of Zip Code or Census Tracts

Radius of Simulated Cluster and No. of Regions	Exact Coordinates, %	Centroid, %	Odds Ratio (95% Confidence Interval)
Zip code
0.5 km
1	91.5	83.3	2.18 (1.82, 2.60)
2	89.4	27.7	21.90 (17.70, 27.10)
3	89.5	58.9	5.97 (4.97, 7.19)
4	100.0	29.1	∞
1 km
1	100.0	87.6	∞
2	71.8	36.8	4.37 (3.81, 5.02)
3	91.9	45.1	13.83 (11.09, 17.26)
4	85.5	28.3	14.90 (12.36, 17.98)
2 km
1	91.1	80.4	2.50 (2.08, 3.00)
2	59.8	31.4	3.26 (2.85, 3.72)
3	52.0	31.8	2.32 (2.05, 2.62)
4	37.2	11.5	4.58 (3.76, 5.56)
3 km
1	96.5	94.1	1.76 (1.27, 2.43)
2	41.1	31.3	1.53 (1.36, 1.73)
3	48.5	22.5	3.26 (2.81, 3.78)
4	23.0	16.4	1.53 (1.29, 1.80)
Census tract
0.5 km
1	100.0	99.8	∞
2	99.0	96.4	3.74 (2.13, 6.54)
3	93.3	70.0	5.94 (4.75, 7.43)
4	99.7	61.5	228.47 (73.69, 708.41)
1 km
1	100.0	97.7	∞
2	97.1	90.1	3.67 (2.55, 5.29)
3	71.2	64.5	1.36 (1.22, 1.51)
4	78.5	66.9	1.80 (1.57, 2.06)

Note. SaTScan was run 1100 times per cell to calculate percentages of clusters significant at P <.05. Number of regions indicates the number of administrative regions (zip code or census tract) into which the 10 simulated cluster points fell.