OBJECTIVES: To evaluate a semi-automatic data extraction from the electronic medical record (EMR) of general practitioners (GPs) through a comparison with a paper sheets data collection simultaneously used in a primary care research project on the quality of prescribing for osteoarthritis in the elderly. SUBJECTS: One hundred and fifty-two GPs using five different EMR-software systems participated with the semi-automatic data extraction from the EMR and 233 GPs collected data with paper registration sheets. METHODS: The proportion of patients with respectively a drug prescription, paracetamol, a non-steroidal anti-inflammatory drug (NSAID) and ibuprofen were compared between the semi-automatic extraction and the paper data collection and among the EMR-software systems. RESULTS: Using the semi-automatic data extraction, a significantly lower proportion of patients on drugs was obtained compared to the paper data collection (adjusted OR: 0.31; 95% CI 0.25-0.39). However, the proportion of patients on a specific type of drug was comparable. Within the results from the semi-automatic extraction, the results were heterogeneous among the different EMR-software systems. CONCLUSIONS: The semi-automatic data extraction with multiple EMR-software systems proposed in this study seems suitable for quality of prescribing assessment in primary care. However, it may be less reliable when only a single EMR-software is used.
OBJECTIVES: To evaluate a semi-automatic data extraction from the electronic medical record (EMR) of general practitioners (GPs) through a comparison with a paper sheets data collection simultaneously used in a primary care research project on the quality of prescribing for osteoarthritis in the elderly. SUBJECTS: One hundred and fifty-two GPs using five different EMR-software systems participated with the semi-automatic data extraction from the EMR and 233 GPs collected data with paper registration sheets. METHODS: The proportion of patients with respectively a drug prescription, paracetamol, a non-steroidal anti-inflammatory drug (NSAID) and ibuprofen were compared between the semi-automatic extraction and the paper data collection and among the EMR-software systems. RESULTS: Using the semi-automatic data extraction, a significantly lower proportion of patients on drugs was obtained compared to the paper data collection (adjusted OR: 0.31; 95% CI 0.25-0.39). However, the proportion of patients on a specific type of drug was comparable. Within the results from the semi-automatic extraction, the results were heterogeneous among the different EMR-software systems. CONCLUSIONS: The semi-automatic data extraction with multiple EMR-software systems proposed in this study seems suitable for quality of prescribing assessment in primary care. However, it may be less reliable when only a single EMR-software is used.