Toward Standardized Performance Metrics in the Cardiovascular ICU: A Systematic Review of Quality Indicators

TL;DR

This paper reviews quality indicators in cardiovascular ICU settings to identify gaps and suggest ways to standardize performance metrics for better patient outcomes.

Contribution

The study proposes a framework for standardizing cardiovascular ICU quality indicators and highlights the need for better validation and reproducibility.

Findings

Most studies focused on outcome quality indicators, particularly mortality prediction using machine learning.

Structural and process quality indicators lacked robust measurement frameworks and reproducibility.

Future work should validate predictive models and develop standardized process and structural metrics.

Abstract

The cardiovascular intensive care unit (CVICU) requires robust quality indicators (QIs) to standardize performance measurement and improve patient outcomes. However, heterogeneity in QI definitions, measurement tools, and implementation practices persists. This systematic review synthesizes evidence on CVICU QIs, evaluates their methodological rigor, and proposes a framework for standardization. Following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines, we searched PubMed, Embase, Scopus, Web of Science, and CINAHL for relevant studies. Eight studies met the inclusion criteria, encompassing retrospective cohorts, predictive models, and mixed-methods designs. Quality assessment employed the Newcastle-Ottawa Scale (NOS) for cohort studies and the Mixed Methods Appraisal Tool (MMAT) for non-randomized studies. Narrative synthesis categorized QIs by Donabedian domains (structure, process, outcome). Included studies (n=8) predominantly focused on outcome QIs (5/8 studies), particularly mortality prediction using machine learning. Risk of bias was moderate to high, with most studies lacking prospective validation or objective measurements. Structural QIs were especially underrepresented, and although Delphi methods were employed, they lacked external validation and reproducibility, limiting generalizability. Process QIs relied on subjective surveys, while structural QIs lacked robust measurement frameworks. Alignment with Donabedian and Institute of Medicine (IOM) frameworks was reported in 6/8 studies, yet consistency in application was limited. CVICU QIs prioritize outcome measurement through artificial intelligence (AI)-driven tools but lack standardization in the development, validation, and operationalization of process and structural indicators. Future work should (1) validate predictive models in multicenter, prospective settings, (2) develop objective and reproducible process metrics, and (3) expand structural QIs for global applicability, accounting for resource constraints, variability in infrastructure, and cultural differences in care delivery. Given the limited number of studies, findings should be interpreted cautiously and considered hypothesis-generating rather than definitive. This review informs efforts to harmonize CVICU performance measurement.

Figures