Codifying healthcare – big data and the issue of misclassification

Abstract

The rise of electronic medical records has led to a proliferation of large observational studies that examine the perioperative period. In contrast to randomized controlled trials, these studies have the ability to provide quick, cheap and easily obtainable information on a variety of patients and are reflective of everyday clinical practice. However, it is important to note that the data used in these studies are often generated for billing or documentation purposes such as insurance claims or the electronic anesthetic record. The reliance on codes to define diagnoses in these studies may lead to false inferences or conclusions. Researchers should specify the code assignment process and be aware of potential error sources when undertaking studies using secondary data sources. While misclassification may be a short-coming of using large databases, it does not prevent their use in conducting meaningful effectiveness research that has direct consequences on medical decision making.