Improved Accuracy of Co-morbidity Coding Over Time After ICD-10 Introduction
Author Information
Author(s): Januel Jean-Marie, Luthi Jean-Christophe, Quan Hude, Borst François, Taffé Patrick, Ghali William A, Burnand Bernard
Primary Institution: Institute of Social and Preventive Medicine, CHUV and University of Lausanne, Switzerland
Hypothesis
Is there evidence of improved data accuracy in co-morbidity coding after the introduction of ICD-10?
Conclusion
The accuracy of administrative data in recording clinical conditions improved slightly between 1999 and 2003.
Supporting Evidence
- Sensitivity for Charlson co-morbidities increased from 36.5% in 1999 to 42.8% in 2003.
- Sensitivity for Elixhauser co-morbidities increased from 34.2% in 1999 to 41.6% in 2003.
- Statistically significant improvements were observed for six conditions.
- Kappa values increased for 29 co-morbidities from 1999 to 2003.
- Professional coders were increasingly employed during the study period.
Takeaway
This study shows that hospitals got better at coding patient conditions over time after switching to a new coding system.
Methodology
Cross-sectional time trend evaluation study using hospital chart data from 3,499 randomly selected patients discharged in 1999, 2001, and 2003.
Potential Biases
Inter-coder variation and differences in coding practices across hospitals may introduce bias.
Limitations
The study relied on chart data as the reference standard, which may have missing clinical information, and the sample size was limited for some rare conditions.
Participant Demographics
Mean age was 58.1 years in 1999, 57.3 years in 2001, and 55.3 years in 2003; sex ratio was approximately 1.1 female to male.
Statistical Information
P-Value
p<0.05
Confidence Interval
95% CI for Kappa values ranged from 0.05 to 0.78 across years.
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website