DocumentCode :
3200157
Title :
Predicting the onset of bullous pemphigoid with co-morbidities: A survey based on a nationwide medical database
Author :
Chu-Cheng Kuo ; Fang-Chi Yang ; Meng-Han Yang ; Ding-Dar Lee
Author_Institution :
Dept. of Comput. Sci. & Inf. Eng., Nat. Kaohsiung Univ. of Appl. Sci., Kaohsiung, Taiwan
fYear :
2013
fDate :
18-21 Dec. 2013
Firstpage :
30
Lastpage :
37
Abstract :
Bullous pemphigoid (BP) is an acute or chronic autoimmune skin disease. It has been reported that poor conditions of general health and increased mortality rates were observed in BP patients, while pathophysiologic reasons of these phenomena merit further investigation yet. In this paper we tried to investigate the co-morbid associations between bullous pemphigoid and various diseases, using a nationwide population-based medical database, NHIRD. 5,933 adult BP patients of ages >= 18 years were extracted from the NHIRD, accompanied with 17,712 age- and sex-matched control subjects. Thirty-one disorders were selected from the Charlson or Elixhauser index for which we tried to evaluate their respective associations with BP. The results obtained revealed that there existed co-morbid relationships between BP and various illnesses, especially for hemiplegia/paraplegia (odds ratio [OR] 8.71), dementia (OR 8.04), fluid electrolyte disorders (OR 6.6), cerebrovascular disease (OR 5.97), moderate/severe liver disease (OR 5.22), weight loss (OR 4.9), and psychoses (OR 4.64). Furthermore, using the study cohort as input data and the retrospective diagnosis records of the above-mentioned co-morbidities occurring within 24 months before onset of BP as features, we employed the algorithm of support vector machine (SVM) to generate the prediction model for onset of BP afterward. Subjects labeled as positive by the prediction model were about 9 times more likely than those as negative to actually develop BP. Accordingly, the prediction model created by SVM was valuable for physicians to screen patients with higher risk of BP.
Keywords :
bioinformatics; diseases; support vector machines; Charlson index; Elixhauser index; NHIRD database; acute autoimmune skin disease; bullous pemphigoid; cerebrovascular disease; chronic autoimmune skin disease; comorbidities; fluid electrolyte disorders; mortality rates; nationwide medical database; psychosis; support vector machine; weight loss; Diseases; Fluids; Indexes; Medical diagnostic imaging; Predictive models; Support vector machines; NHIRD; bullous pemphigoid; co-morbidity; support vector machine;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Bioinformatics and Biomedicine (BIBM), 2013 IEEE International Conference on
Conference_Location :
Shanghai
Type :
conf
DOI :
10.1109/BIBM.2013.6732731
Filename :
6732731
Link To Document :
بازگشت