Quality of coding within clinical datasets: A case-study using burn-related hospitalizations

Area of Science:

• Health Services Research
• Medical Informatics
• Hospital Administration

Background:

• Accurate clinical data in administrative databases is vital for healthcare funding.
• Diagnosis-Related Groups (DRGs) are key to hospital payment mechanisms, making proper coding essential.
• Burn hospitalization data quality and coding practices require scrutiny.

Purpose of the Study:

• To characterize and assess coding patterns in Portuguese burn-related hospitalizations.
• To identify potential miscoding practices affecting All-Patient Refined Diagnosis-Related Groups (APR-DRG) classification.
• To compare APR-DRG and Severity of Illness (SOI) frequencies between hospitals with and without burn units.

Main Methods:

• Analysis of a nationwide Portuguese inpatient administrative database (2011-2015).
• Inclusion of 4,182 burn-related admissions.
• Comparison of APR-DRG and SOI frequencies, and individual diagnosis/procedure codes across hospitals.

Main Results:

• Significant differences in APR-DRG frequencies (842, 844) and SOI levels were observed between hospitals with burn units.
• Potential miscoding of extensive third-degree burns and debridement procedures noted.
• Discrepancies in reporting comorbidities and non-operating room procedures may influence SOI levels.

Conclusions:

• Suspected miscoding practices in burn data could impact APR-DRG classification and healthcare funding.
• Findings highlight the need for improved medical coding standards and audit processes.
• Data quality is crucial for DRG grouping, clinical research, and healthcare management.