Predicting Cancer Types Using Community factors. The Role of Deprivation, Race and Age

Objective: 25 different cancer types are examined in Scotland as function of demographic factors that include their deprivation level. Findings can help poli-cymakers to design recovery plans per case and one used pure open demographic data from NHSS to link race to cancer type. Methods: Cancer’s progression is factored on four basic prevalent factors that may be linked to all types. These are (a) age bands, (b) race, (c) deprivation index, and (d) gender. The distributions of cancers per such attribute are taken and then three 3 different statistics measures are taken to address two questions: (1) are there economic factors that can capture cancer occurrence differences among these types ? (2) can some type prevail using them ?. The methods are: (a) the Kullback Leibler divergence transformation, (b) the significance test (p.value) of the similarity using the Chi-squared test, (c) the Entropy. These are compared. The raw data link these economy factors to all cancer types and can support the study of their joint distributions. The attributes that cause the most diverging distributions (frequencies of cancer types) among its values are taken as the more informative ones. All values for those attributes are examined that give quite different frequencies of cancer. This is an information-theoretic approach to the problem that accounts for common economy variables and reveals that diversity of age and race may be driver for cancer. Brief statement of primary results: The results show that among the four attributes studied age is the prevalent cause for having quite different occurrences of cancer given the data at hand. Conclusion: Policymakers can use such findings to roughly understand how cancer progression (types) can link to major demographics.