Development and evaluation of the accuracy of an indicator of the appropriateness of interventional cardiology generated from a French registry

Development of appropriateness indicators of medical interventions has become a major quality-of-care issue, especially in the domain of interventional cardiology (IC). The objective of this study was to develop and evaluate the accuracy of an indicator of the appropriateness of interventional cardiology acts (invasive coronary angiographies (ICA) and percutaneous coronary interventions (PCI)) in patients with coronary stable disease and silent ischemia, automated from a French registry. All ICA and PCI recorded in a Regional IC Registry (ACIRA) and operated for a stable coronary artery disease or silent ischemia from January 1st to December 31th 2013 in eight IC hospitals of Aquitaine, southwestern France, were included. The indicator was developed to reflect European guidelines. Classification of appropriateness by the indicator, measured on the registry database, was compared to the classification of a reference standard (expert judgment applied through complete record review) on a random sample of 300 interventions. Accuracy parameters were estimated. A second version of the indicator was defined, based on the analysis of false negative and positive results, and its accuracy estimated. The second indicator accuracy was: sensitivity 63.5% (95% confidence interval CI [51.7-75.3]), specificity 76.0% (95%CI [70.4-81.6]), PPV 43.0% (95% CI [33.0-53.0]) and NPV 88.0% (95% CI [83.4-92.6]). When stratified on the type of act, parameters were better for ICA alone than for PCI. Accuracy of the indicator should raise with improvement of database quality. Despite its average accuracy, it is already used as a benchmark indicator for cardiologists. It is sent annually to each IC center with value of the indicator at the region level to allow a comparison.