| Summary: | Background: Treatment data from general practitioners (GP) in private practice have great potential for research into diagnostics, therapy and quality of care. The quality of these data influences not only the primary research results, but also the results of later secondary data analyses, but then for other research questions. Completeness and availability of data at the time of data collection are important quality characteristics that depend on medical documentation and technical data export. Using GP treatment data as an example, we examine these quality characteristics and propose a procedure for continuous quality assurance.Methods: A fictitious patient was defined with 22 treatment characteristics of routine general practitioners treatment (e.g. age, gender, medication, diagnoses, risk factors, inability to work, therapies, hospitalization, etc.). Without prior training, the data of the fictitious patient were electronically documented and exported by general practitioners. Comparisons between fictitious patient data, data documented electronically by general practitioners and exported data enabled the assessment of data completeness and availability.Results: Approximately 80% of the tasks could be completed without medical training. Exemplary studies with six practice software systems show that without software adaptation and medical training, only approx. 37% of the treatment data entered is available for secondary research in data export. An increase to 83.4% availability would be possible through software adaptations. If, in addition, medical training were carried out, a data availability of approx. 95% could be achieved.Discussion: Missing or incorrectly mapped data must be reckoned with in the data export prior to secondary research. The completeness and availability of the data must be tested individually for each practice software during the primary data collection and documented in the metadata of the research data. Fictitious patient data with defined characteristics can also be used for continuous quality assurance in quarterly software updates.Conclusions: Independent, continuous quality assurance of the data interfaces, medical training and investigations into the completeness of routine and treatment data can considerably improve the quality of subsequent secondary data research.
|
|---|
