Why the study?
National medical quality registries are used to improve treatment, in research, and as management tools, but data quality in the Norwegian Myocardial Infarction Registry needed investigation.
Observational (n=617)
Yes
Registry variables that can be extracted without interpretation have high data quality, whereas those requiring interpretation of free-text clinical notes have significantly lower accuracy.
Registry data quality is high for non-interpretive variables but low for interpretive ones; leaves open reliability of complication rates in cardiovascular research.
Background: The main purpose of national medical quality registries is to contribute to better patient treatment. The registries are also used in research and as management tools. The aim of the study was to investigate data quality in the Norwegian Myocardial Infarction Registry by comparing the proportion of correctly registered variables with a reference standard. Material and method: A total of 641 myocardial infarction cases registered in the Norwegian Myocardial Infarction Registry in 2020 were randomly sampled from seven hospitals. Seven doctors - specialists in cardiovascular diseases or specialty registrars in cardiovascular diseases or internal medicine - reviewed the patient records and re-registered 23 registry variables for the sample. These new registrations constituted the reference standard. The variables were divided into three categories: categorical variables with and without interpretation of the text in the records and continuous variables for the time indications. We then calculated the proportion of correctly registered variables and response alternatives. Results: For variables retrieved from the patient record text without interpretation, 87 % to 100 % were correctly registered. For variables that were based on interpretations of the patient record text, the proportion that were correctly registered was lower: 20 % to 70 % for 'yes' for variables on clinical instability in non-ST-elevation myocardial infarction (NSTEMI), and 60 % for 'yes' for the variable 'myocardial infarction as complication'. The proportion of correctly registered time variables was 48 % to 100 %. Interpretation: The proportion of correctly registered variables was higher for categorical variables that could be retrieved without interpretation from the patient record text than for continuous time variables and categorical variables based on interpretations of the patient record text.
No takes yet. Share an insight, caveat, or question.
Kiel et al. (2024) conducted an observational in Acute myocardial infarction (n=617). Registry data vs. Reference standard (physician re-registration) was evaluated on Proportion of correctly registered variables compared to a reference standard.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: