Date of Award

January 2015

Document Type

Open Access Thesis

Degree Name

Medical Doctor (MD)

Department

Medicine

First Advisor

Jonathan Grauer

Subject Area(s)

Medicine, Surgery

Abstract

The use of national inpatient databases for orthopaedic surgery research has been increasing. However, large databases that rely on administrative data, such as International Classification of Diseases Ninth Revision (ICD-9) codes, may misrepresent patient information, thus affecting the results of studies using this data.

The present study uses easily quantified and objective variables of obesity and anemia as example comorbidities to assess the accuracy of ICD-9 codes in the setting of their continued use in orthopaedic surgery database studies.

For each study arm, a large inpatient population was obtained from the Yale-New Haven hospital. Each patient's medical record was reviewed, and the presence of ICD-9 discharge codes for obesity and anemia was directly compared to documented body mass index (BMI) and preoperative hematocrit, respectively.

ICD-9 discharge codes for both non-morbid obesity and anemia had a sensitivity of just 0.19. The sensitivity of the ICD-9 code for morbid obesity was 0.48.

Using obesity and anemia as examples, this study highlights the potential errors inherent to ICD-9 codes. This calls into serious question the utility of administrative databases for research purposes. Moreover, it is likely that these inaccuracies apply to additional variables as well. As database research continues to increase within orthopaedic surgery, it is important to realize that study outcomes can be skewed by data accuracy, and thus should not be blindly accepted simply by virtue of large sample sizes.

Share

COinS