ICD-10 Coding Will Challenge Researchers: Caution and Collaboration may Reduce Measurement Error and Improve...

Area of Science:

• Health Services Research
• Medical Informatics
• Public Health

Background:

• The 2015 update from International Classification of Diseases, 9th Revision (ICD-9) to 10th Revision (ICD-10) significantly altered US health care administrative data.
• This transition risks introducing artificial changes in health and health services measures derived from claims data.

Purpose of the Study:

• To quantify measurement changes and discontinuities in common diagnoses resulting from the ICD-9 to ICD-10 code upgrade.
• To assess the impact of the ICD-10 transition on health care administrative data.

Main Methods:

• Utilized 100% Medicare inpatient data from 2012-2015 to calculate quarterly diagnosis frequencies.
• Compared 2015 fourth-quarter data (post-ICD-10 adoption) with baseline frequencies from 2012-2014.
• Employed Centers for Medicare and Medicaid Services Chronic Conditions Data Warehouse (CCW) definitions and translation software for analysis.

Main Results:

• Observed wide variations in recorded condition discontinuities post-ICD-10 adoption.
• Reported a 3.2% increase in chronic lung disease and a 1.8% decrease in depression frequency in Q4 2015 compared to Q4 2014.
• Translation software for conditions like Charlson-Deyo and Elixhauser showed discontinuities ranging from -8.9% to +10.9%.

Conclusions:

• ICD-9 to ICD-10 translations are not always aligned, leading to temporal discontinuities.
• These discontinuities may compromise ICD-based measurements and risk-adjustment methodologies.
• Proposed the creation of a public resource for researchers to share identified ICD-10 discontinuities and their solutions.