Author_Institution :
Dept. of Comput. Eng., Santa Clara Univ., Santa Clara, CA, USA
Abstract :
Workers´ compensation is a form of medical insurance provided by employers to their employees. Workers´ compensation databases have a complete picture of each case, including treatments, treatment costs, and treatment duration, and analyzing their data can be extremely helpful to identify patterns and possibly determine the correlation between diseases and treatments and also the correlation between different treatments. These correlations have several benefits, including the decrease in insurance costs, by making the process of authorization more accurate and less wasteful. Also, patterns enable prediction, which may help insurance companies to estimate costs and to detect irregularities and/or special cases, which may last longer or consume more resources. One problem in analyzing such data is the fact that medical information is not precise, e.g., different codes can be used for similar procedures. However, although not precise, the relationship between diseases and treatments seem to have some consistency, which might enable the correlation. In this paper, we present a method for analyzing workers´ compensation data to obtain information on correlation among diseases, treatments and expected costs. We present a study on Carpal Tunnel Syndrome to show that consistency enables prediction and the development of guidelines for treatment authorization.
Keywords :
data mining; insurance; medical information systems; patient treatment; Carpal tunnel syndrome; association rules; medical information; medical insurance; treatment authorization; treatment patterns; Anesthesia; Association rules; Correlation; Diseases; Surgery; Carpal Tunnel Syndrome; association rules; temporal patterns;