International Severity of Illness Scoring System

• In ICU
• Wed, 26 Nov 2025

Developing a globally applicable critical care severity score is an urgent priority for improving research, benchmarking, quality improvement, and public health response across diverse health systems. Severity scores have long been central to ICU practice, supporting benchmarking, risk adjustment, resource allocation, and disease characterisation. Yet, most existing scores were designed within specific national or regional contexts, limiting their effectiveness when applied internationally. 

The COVID-19 pandemic highlighted the need for standardised global approaches: severity scores were instrumental for describing trial populations, assessing treatments, and comparing outcomes across heterogeneous systems. As similar global challenges continue, such as trauma care improvement and expansion of complex medical, surgical, and obstetric services, the need for a unified international scoring system becomes increasingly evident.

Several barriers prevent existing scores from functioning well worldwide. Many scoring systems incorporate comorbidities and disease profiles common in high- and middle-income countries but less representative of low-income settings. Age categories also assume life expectancies typical of higher-income regions, limiting score calibration across countries with different demographic profiles. As a result, predictive performance is inconsistent when applied beyond their original development context.

Also, many severity scores rely on laboratory markers or diagnostics that are not universally accessible. This results in large amounts of missing data when the scores are used in resource-limited ICUs. The problem is compounded by limited reporting in validation studies from lower- and middle-income countries. Differences in local clinical decision-making also influence which tests are ordered, and these contextual variations are not captured in existing models.

ICUs operate differently across countries, due to variability in case mix, available treatments, pre-ICU care, and post-ICU support. Such differences significantly influence score performance but are not accounted for in most severity models. Despite these limitations, regionally calibrated scores are routinely used to stratify participants in multinational clinical trials, creating challenges for global data synthesis and limiting the applicability of findings across diverse populations.

A truly global severity score would unlock substantial benefits across research, clinical care, and public health. It would enable large-scale epidemiological studies spanning continents and income levels. These globally representative analyses would enhance understanding of critical illness patterns and improve the characterization of diseases with worldwide impact.

Countries without established benchmarking systems, particularly those with limited ICU capacity, would gain a tool to compare performance meaningfully across health systems. This comparative approach could help identify gaps, guide quality improvement, and promote equitable resource development.

A global score would improve stratification in clinical trials and ensure treatment effect estimates are more generalisable. It would also facilitate multinational recruitment, increasing diversity in study populations and enhancing external validity. Consistent severity classification would support better synthesis of evidence across trials, improving systematic reviews and meta-analyses.

A standardised severity score could enhance surveillance of critical illness due to communicable diseases, allowing earlier detection of shifts in disease phenotype or severity. It could aid in forecasting ICU demand, particularly during global crises such as pandemics, natural disasters, and conflicts.

Efforts to adapt SAPS 3 and APACHE IV to specific regions represent early steps toward broader relevance. However, these remain limited because they do not address data scarcity in low-resource settings and have not been fully validated across socioeconomic contexts.

Scores such as the Global Open Source Severity of Illness Score (GOSSIS) have been constructed using international datasets. While promising, challenges remain regarding applicability in the lowest-income contexts and in settings lacking certain diagnostic capabilities. Other scores such as SEVERITAS, SMS-ICU, UVA, and TropICS address data unavailability by relying on small sets of universally accessible variables. Although useful, they still require large-scale validation across multiple continents and diverse ICUs. Growing international critical care registries, like the LOGIC initiative, offer an unprecedented opportunity for such validation.

An international score should not replace locally calibrated models, which remain vital for national benchmarking. Instead, it should serve as a common global language, enabling collaboration and more meaningful integration of international research.

The global critical care community should prioritise developing an international severity score using a small set of universally available variables, validated through real-world data from registries across diverse regions and economic contexts. Such a model may be newly constructed or derived through adaptation of existing tools. Performance may be strengthened by incorporating socioeconomic, geographic, and system-level variables. With expanding international datasets and harmonised data definitions, the moment is ripe to build a scoring system that can support global research, quality improvement, and coordinated public health response.

Source: Critical Care Science
Image Credit: iStock 

References:

Tracy A, Salluh JIF, Buanes EA et al. (2025) The case for an international severity of illness scoring system. Critical Care Science. 37:e20250293.