Improving discharge data fidelity for use in large administrative databases.

Area of Science:

• Neurosurgery
• Health Services Research
• Medical Informatics

Background:

• Large administrative databases are crucial for population-based health care studies.
• Lumbar fusion surgery rates are rising, with indications like stenosis and spondylosis often ill-defined.
• Accuracy of International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes in reflecting surgeon's intent is unclear.

Purpose of the Study:

• To compare ICD-9-CM codes assigned by medical coders with surgeon's indications from medical charts for lumbar fusion.
• To identify barriers to data fidelity in administrative databases.

Main Methods:

• Retrospective review of lumbar fusions performed between August 2011 and August 2013.
• Categorization of surgeon's indications (e.g., spondylolisthesis, stenosis, degenerative disc disease).
• Comparison of surgeon diagnoses with primary and secondary ICD-9-CM codes; follow-up interviews with coders.

Main Results:

• For non-degenerative conditions (tumor, infection, fracture), 98% of primary diagnoses matched surgical indications.
• For degenerative diseases (126 cases), only 48% of primary ICD-9-CM diagnoses matched surgeon's indications.
• Considering all ICD-9-CM codes, 79% of degenerative disease cases had the indication identified, but 21% of hospitalizations had missed surgical diagnoses.

Conclusions:

• Limitations in ICD-9-CM coding accuracy affect studies using administrative claims data.
• Factors like physician documentation quality, coder training, and ICD code ambiguity contribute to errors.
• Enhanced coder training and surgeon-physician communication can improve data fidelity for more reliable research and decision-making.