All projectsView source
BI
Medical Coding Accuracy Audit
Inaccurate medical coding leads to claim denials, compliance risks, and revenue leakage. This project audited coding accuracy across departments to identify training needs.
92% coding accuracy baseline established
2024PythonExcelPower BI
Dataset
Kaggle: Synthetic AR Medical Dataset
Key Questions
- What is the overall coding accuracy rate across the organisation?
- Which departments or specialities have the highest coding error rates?
- Are certain ICD/CPT code categories more prone to errors than others?
Methods
- Random sampling of 5,000 coded encounters for manual review comparison
- Error categorisation: upcoding, undercoding, incorrect modifier, missing code
- Department-level accuracy benchmarking
- Trend analysis of accuracy over time to measure training impact
Results
Overall coding accuracy was 92%, with the emergency department (86%) and outpatient surgery (88%) performing below the 95% target. The most common error type was missing secondary diagnoses (38% of all errors).
Coding Accuracy by Department
Recommendations
- Prioritise targeted coding training for ED and outpatient surgery coders
- Implement a secondary diagnosis prompt in the coding workflow
- Establish quarterly audit cycles to track improvement
- Consider computer-assisted coding tools for high-complexity specialities
Limitations
Manual review is itself subject to inter-rater variability. The sample may not capture rare code categories. This audit provides a snapshot; ongoing monitoring is essential.