Download presentation
Presentation is loading. Please wait.
Published bySteven Burke Modified over 9 years ago
1
Brain Cancer Mortality in the United States Joint work with: Zixing Fang, UCLA David Gregorio, Univ Connecticut
2
U.S. Brain Cancer Mortality 1986-1995 deathsrate* (95% CI) Children (age <20): 5,0620.75(0.66-0.83) Adults (age 20+): 106,7106.0(5.8-6.2) Adult Women: 48,6504.9(4.7-5.0) Adult Men: 58,0607.2(7.0-7.5) * annual deaths / 100,000
3
Brain Cancer Known risk factors: High dose ionizing radiation Selected congenital and genetic disorders Explains only a small percent of cases. Potential risk factors: N-nitroso compounds?, phenols?, pesticides?, polycyclic aromatic hydrocarbons?, organic solvents?
4
Adjustments Age Gender Ethnicity (African-American, White, Other) All subsequent analyses where adjusted for:
5
Brain Cancer Mortality, Children 1986-1995
6
Cuzick-Edward’s Test: Children k p-value 2000.04 5000.13
7
Tango’s Excess Events Test: Children p-value 10000.005 20000.06 50000.21 100000.29
8
Spatial Scan Statistic, Children
9
Children: Seven Most Likely Clusters Cluster Obs Exp RR p= 1. Carolinas86511.70.24 2. California164.93.30.74 3. Michigan 318 2501.30.74 4. S Carolina24102.50.79 5. Kentucky-Tenn 127881.40.79 6. Wisconsin102.44.10.98 7. Nebraska123.63.30.99
10
Conclusions: Children Some evidence of global spatial clustering, but rather weak. No statistically significant clusters detected. Any part of the pattern seen on the original map may be due to chance.
11
How About Adults?
12
Brain Cancer Mortality, Adults 1986-1995
13
Cuzick-Edward’s k-NN: All Adults k p-value 40000.0001 100000.0001
14
Tango’s EET: All Adults p-value 10000.0001 20000.0001 50000.0001 100000.0001
15
Spatial Scan Statistic: Adults
16
Brain Cancer Mortality, Adults 1986-1995
17
Cuzick-Edward’s: Women k p-value 15000.0001 30000.0001
18
Tango’s EET: Women p-value 10000.0001 20000.0001 50000.0001 100000.0001
19
Spatial Scan Statistic, Women
20
Women: Most Likely Clusters Cluster Obs Exp RR p= 1. Arkansas et al. 283023281.220.0001 2. Carolinas178315181.170.0001 3. Oklahoma et al.170914961.140.003 4. Minnesota et al.261623691.100.01 10. N.J. / N.Y.180923000.790.0001 11. S Texas 127 2140.590.0001 12. New Mexico et al. 84910490.810.0001
21
Cuzick-Edward’s: Men k p-value 20000.0001 40000.0001
22
Tango’s EET: Men p-value 10000.0001 20000.0001 50000.0001 100000.0001
23
Spatial Scan Statistic: Men
24
Men: Most Likely Clusters Cluster Obs Exp RR p= 1. Kentucky et al. 329528601.150.0001 2. Carolinas192516581.160.0001 3. Arkansas et al. 1143 9641.190.001 4. Washington et al.166414551.140.003 5. Michigan125110741.170.005 11. N.J. / N.Y. 208426150.800.0001 12. S Texas 157 2620.600.0001 13. New Mexico et al.141816800.840.0001 14. Upstate N.Y. et al.164218950.870.0001
25
Conclusions: Adults Strong evidence of global spatial clustering. It is possible to pinpoint specific areas with higher and lower rates that are statistically significant, and unlikely to be due to chance. The exact borders of detected clusters are uncertain. Similar patterns for men and women.
26
Conclusion: General Tests for spatial randomness are very useful additions to cancer maps, in order to determine if the observed patterns are likely due to chance or not. Different tests provide complementary information.
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.