A systematic review of modeling approaches to evaluate diagnostics programs for stable coronary heart disease: suggestions for future health economic models

To critically review health economic models used to evaluate diagnostic strategies for stable coronary artery disease (CAD), with a focus on methodological characteristics in model development. A systematic literature review (SLR) was conducted across major electronic databases and health technology assessment (HTA) agency websites from inception to January 2025. Studies were independently screened using predefined criteria. Data extraction focused on general characteristics, model structure, approaches to uncertainty, model validation, and transparency. These aspects were then systematically described to identify common methodological characteristics and potential areas for improvement. Seventy articles were included, comprising 67 peer-reviewed publications and three HTA reports. Most studies were model-based (n = 46), and models were structured as decision tree models (n = 27), Markov models (n = 6), or hybrid approaches (n = 13). While only 18 studies explicitly justified their choice of model structure. Substantial heterogeneity was observed in key modeling assumptions, including time horizon, cycle length and data sources. Among studies using long-term Markov models (n = 19), none explicitly reported the use of half-cycle correction. Although most studies conducted sensitivity analyses (n = 56), approaches to uncertainty varied widely, and only seven models incorporated deterministic, probabilistic, and scenario analyses. Model validation was infrequently reported, with only two studies describing any formal validation activities. Overall, transparency was limited, as none of the studies provided sufficient technical documentation and reporting of model implementation tools was also incomplete (n = 28). Methodological heterogeneity and incomplete reporting exit in economic evaluation of diagnostic strategies for stable CAD, which may affect their transparency and comparability. Future research may benefit from clearer model justification, more consistent reporting of key technical assumptions, and systematic approaches to uncertainty and validation.