A survey of validity and utility of electronic patient records in a general practice

Objective: To develop methods of measuring the validity and utility of electronic patient records in general practice. Design: A survey of the main functional areas of a practice and use of independent criteria to measure the validity of the practice database. Setting: A fully computerised general practice in Skipton, north Yorkshire. Subjects: The records of all registered practice patients. Main outcome measures: Validity of the main functional areas of the practice clinical system. Measures of the completeness, accuracy, validity, and utility of the morbidity data for 15 clinical diagnoses using recognised diagnostic standards to confirm diagnoses and identify further cases. Development of a method and statistical toolkit to validate clinical databases in general practice. Results: The practice electronic patient records were valid, complete, and accurate for prescribed items (99.7%), consultations (98.1%), laboratory tests (100%), hospital episodes (100%), and childhood immunisations (97%). The morbidity data for 15 clinical diagnoses were complete (mean sensitivity=87%) and accurate (mean positive predictive value=96%). The presence of the Read codes for the 15 diagnoses was strongly indicative of the true presence of those conditions (mean likelihood ratio=3917). New interpretations of descriptive statistics are described that can be used to estimate both the number of true cases that are unrecorded and quantify the benefits of validating a clinical database for coded entries. Conclusion: This study has developed a method and toolkit for measuring the validity and utility of general practice electronic patient records