Validated models for pre-test probability of stable coronary artery disease: a systematic review suggesting how to improve validation procedures

An overuse of invasive and non-invasive anatomical testing for the diagnosis of coronary artery disease (CAD) affects patients and healthcare professionals safety, and the sustainability of Healthcare Systems. Pre-test probability (PTP) models can be routinely used as gatekeeper for initial patient management. Although with different positions, international organizations clearly underline the need for more information on the various risk factors acting as modifier of the PTP. This systematic review addresses validation of PTP models adopting variables available at the first-line assessment of a suspected stable CAD. A comprehensive search has been done in MEDLINE(R), HealthSTAR, and Global Health databases. Nearly all the models considered in the 27 analysed papers include age, sex, and chest pain symptoms. Other common risk factors are smoking, hypertension, diabetes mellitus and dyslipidaemia. Only one model considers genetic profile. Reported AUCs range from 0.51 to 0.81. Relevant heterogeneity sources have been highlighted, such as the sample size, the presence of a PTP cut-off and the adoption of different definitions of CAD which can prevent comparisons of results. Very few papers address a complete validation, making then impossible to understand the reasons why the model does not show a good discrimination capability on a different data set. We consequently recommend a more clear statement of endpoints, their consistent measurement both in the derivation and validation phases, more comprehensive validation analyses and the enhancement of threshold validations of PTP to assess the effects of PTP on clinical management.