Authors: Xiaoyi Wang, Suraj M. Alexander
Addresses: Tax Refund Solution, Republic Bank & Trust Company, 601 West Market Street, Louisville, KY 40202, USA. ' Department of Industrial Engineering, University of Louisville, Louisville, KY 40292, USA
Abstract: Linking medical records across multiple databases is a very important task, since medical errors, uncompensated care and medical costs are rising at a rapid rate. However, inconsistencies in data records, caused primarily by errors in data entry, make matching of records and satisfactory data linkage difficult. This paper presents and assesses the performance of an approximate matching methodology developed utilising an algorithm for machine learning. The results indicate that this approach to record linkage is promising.
Keywords: multiple databases; approximate matching; records linkage; machine learning; predictive modelling; medical records; record linking; medical errors; uncompensated care; medical costs; data inconsistencies; algorithms; collaborative enterprises; collaboration; healthcare systems; systems engineering.
International Journal of Collaborative Enterprise, 2010 Vol.1 No.3/4, pp.394 - 406
Published online: 01 Feb 2011 *Full-text access for editors Access for subscribers Purchase this article Comment on this article