Authors

  1. Roberts, Eric M. MD, PhD
  2. Kumar, Bahar MPH
  3. Collins, Natalie MSW
  4. Guo, Liang MS
  5. King, Galatea MPH
  6. Wong, Michelle MPH
  7. Barlow, Janice BSN
  8. Bichler, Joyce MSW
  9. Cady, Linda
  10. Engel, Connie PhD
  11. Garrett, Debbie BS, MHA
  12. Harrison, Marie
  13. Morieko, Adriana MBA
  14. Murgai, Neena PhD, MPH
  15. Tanjasiri, Sora Park DrPH, MPH
  16. Pierce, Karen JD

Abstract

Context: Small area (eg, subcounty) cancer mapping is one of the analytic services most commonly requested of cancer registries and local public health agencies, and difficulties in providing it have been noted to undermine public confidence. Although a great many statistical protocols have been published to enable this practice, none of them are in common use to generate information for the general public.

 

Objectives: To evaluate the utility of subcounty breast cancer mapping and articulate guidelines and a possible protocol for its implementation by cancer registries and local public health agencies.

 

Methods: We convened an Expert Advisory Group of breast cancer stakeholders from around California to elicit values, priorities, and preferred characteristics of protocols for proactive subcounty breast cancer mapping. Upon formulating a protocol, we applied it to 9 years of data (2000-2008) describing invasive breast cancer in California for evaluation by the Expert Advisory Group.

 

Results: Maps with subcounty resolution were seen to provide important information with a wide range of applications. Priorities included the avoidance of false-positive findings, scientific credibility, and the provision of information elucidating social and environmental characteristics. A protocol using Kulldorff's Scan Statistic along with postanalytic steps for refining results was elaborated; when applied to the data, 4 discrete regions with elevated rates of invasive breast cancer were identified and described.

 

Conclusions: Expert Advisory Group priorities were readily translatable into a scientifically rigorous protocol that protected confidentiality and avoided statistically unstable rate estimates. The resulting maps enabled participants to visualize geographically defined populations falling within and crossing county boundaries. These findings support the enactment of policies for the routine and proactive analysis of breast cancer surveillance data to provide subcounty information.