Measuring and governing potentially low-value invasive coronary angiography in Europe: a systematic review, construct-specific meta-analysis, and policy framework.

BACKGROUND

Potentially low-value invasive coronary angiography is defined inconsistently across European health systems. It remains unclear which indicators can validly support measurement, comparison and governance.

OBJECTIVE

To determine a measurement framework for identifying and governing distinct forms of potentially low-value invasive coronary angiography without encouraging underuse, using construct-specific synthesis to assess existing indicators.

METHODS

We conducted a PRISMA 2020 systematic review and construct-specific random-effects meta-analysis. PubMed/MEDLINE, Web of Science, Cochrane CENTRAL and Europe PMC were searched through December 2024, with citation tracking and contextual searches updated through June 2026. Estimates were grouped as formal appropriateness adjudication, guideline-discordant referral pathways or low diagnostic yield and pooled using binomial-normal generalized linear mixed models. An author-derived framework translated the findings using wider policy literature; it was not evaluated by the review.

RESULTS

Thirty reports entered the evidence map; 13 studies contributed 14 estimates. Primary analyses included ten estimates covering 30,952 procedures in eight countries. Construct-specific pooled values were 8.4% (95% CI 2.1-28.3) for formal appropriateness adjudication, 58.2% (24.9-85.4) for guideline-discordant pathways and 42.6% (33.1-52.7) for low diagnostic yield. Heterogeneity was extreme. Two constructs included only two studies, and Knapp-Hartung intervals spanned essentially the full proportion scale. These values are construct-dependent signals, not estimates of the European prevalence of unnecessary angiography.

CONCLUSIONS

Existing indicators capture different governance problems and cannot establish how much angiography in Europe is genuinely unnecessary. A harmonised minimum dataset, construct-specific indicators, clinical-record validation and monitoring for unintended underuse should precede benchmarking, public reporting or payment-based use.