Journal of Biomedical Fuzzy Systems Association
Online ISSN : 2424-2578
Print ISSN : 1345-1537
ISSN-L : 1345-1537
Decision Tree Analysis of Factors Relating to Lung Cancers Using DPC Data
Chieko SAKAMOTOHinako TOYAMAKeiichi SAITO
Author information
JOURNAL OPEN ACCESS

2014 Volume 16 Issue 1 Pages 1-6

Details
Abstract
A decision tree was used to find factors relating to lung cancers and 14,383 cases including 363 lung-cancer cases collected during two years from April 2010 by the diagnosis procedure combination / per-diem payment system (DPC/PDPS) were examined. The candidate factors of the decision tree were a smoking index (Brinkman Index: B.I.), gender, age, and comorbidities i.e. a high blood pressure and a type-2 diabetes, of which many lung-cancer cases have onsets. As results, B.I. was extracted as the most significant factor relating to lung cancers. The rates of the lung cancer cases in 3 groups were 1.5% in the low group (B.I.<441), 5.2% in the middle group (440<B.I.<856) and 8.2% in the high group (856<B.I.), and the largest difference of the percentages among the groups was more than 5 times. The second factors were the age and the high blood pressure. These factors were presented as the factors relating to lung cancers in previous clinical reports. The decision tree is useful to detect the factors of lung cancers from DPC data.
Content from these authors
© 2014 Biomedical Fuzzy Systems Association
Previous article Next article
feedback
Top