[Garbage in - garbage out? Validity of coded diagnoses from GP claims records]

Area of Science:

• Medical Informatics
• Primary Care Research
• Health Services Research

Background:

• International Classification of Diseases, 10th Revision (ICD-10) coded diagnoses from claims data are widely used for research and risk adjustment.
• High validity of these diagnoses is crucial for accurate research and quality management.
• The reliability of claims-based diagnoses in general practitioner (GP) practices, reflecting actual patient health problems over a year, remains uncertain.

Purpose of the Study:

• To assess the validity of ICD-10 coded diagnoses from GP claims records by comparing them with diagnoses documented in patient medical records.
• To determine the extent of under-reporting and over-reporting of health problems in claims data.
• To evaluate the accuracy of common diagnoses in primary care settings.

Main Methods:

• A retrospective cross-sectional study design was employed.
• A random sample of 250 patients from 10 GP practices was analyzed.
• Patient medical records were compared with claims-based ICD-10 diagnoses for the year 2003.

Main Results:

• Despite a high average of 6.1 claims-based diagnoses per patient, 30% of treated health problems were under-reported, particularly less severe or chronic conditions.
• Over-reporting of diseases not treated during the period occurred in 19% of cases, often for permanent chronic conditions.
• Eleven percent of claims-based diagnoses had mismatched ICD-10 codes compared to medical records; correctness for common conditions ranged from 71-93%, while completeness ranged from 56-86%.

Conclusions:

• The findings indicate low validity of ICD-10 coded diagnoses derived from GP claims records.
• The reliability issues question the utility of these diagnoses as accurate morbidity indicators in primary care.
• Further research may be needed to improve the accuracy of diagnostic coding in claims data for research and quality assessment.