Socioeconomic inequality of cancer mortality in the United States: a spatial data mining approach

Srinivas Vinnakota,Nina Sn Lam

doi:10.1186/1476-072x-5-9

Abstract

BackgroundThe objective of this study was to demonstrate the use of an association rule mining approach to discover associations between selected socioeconomic variables and the four most leading causes of cancer mortality in the United States. An association rule mining algorithm was applied to extract associations between the 1988–1992 cancer mortality rates for colorectal, lung, breast, and prostate cancers defined at the Health Service Area level and selected socioeconomic variables from the 1990 United States census. Geographic information system technology was used to integrate these data which were defined at different spatial resolutions, and to visualize and analyze the results from the association rule mining process.ResultsHealth Service Areas with high rates of low education, high unemployment, and low paying jobs were found to associate with higher rates of cancer mortality.ConclusionAssociation rule mining with geographic information technology helps reveal the spatial patterns of socioeconomic inequality in cancer mortality in the United States and identify regions that need further attention.

Highlights

The objective of this study was to demonstrate the use of an association rule mining approach to discover associations between selected socioeconomic variables and the four most leading causes of cancer mortality in the United States
Shown in red is the spatial distribution of areas having a medium-high rate of prostate cancer mortality among black men and high density of households with black female householder with no husband present and with no children under the age of 18 years
This pattern has the highest support among the association rules involving prostate cancer mortality rate

Summary

Introduction

From the application of a paper based point-map by John Snow to investigate the outbreak of cholera epidemic in 1854 to the current day practices of using digital maps, medical and public health profession has greatly benefited from the use of spatial information as manifested by maps In this regard contemporary Geographic Information Systems (GIS) provide an excellent platform in which digital maps and data can be manipulated to extract valuable information, and serve as an excellent medium for analyzing and visualizing spatial patterns. GIS and its associated spatial analysis have previously been used successfully to detect disease clusters [1,2], predict disease outbreaks [3,4], evaluate accessibility to health care facilities [5], determine health-environment interactions [6,7,8], and analyze spatial distribution of disease [9,10] One such area has been the study of spatial distribution of cancer incidence and mortality [11]. The purpose of this study is to introduce the use of an association rule mining approach to (page number not for citation purposes)

Objectives

Methods

Results

Discussion

Conclusion