J Eval Clin Pract. 2026 Jun;32(4):e70459. doi: 10.1111/jep.70459.
ABSTRACT
BACKGROUND: Intensive Care Unit (ICU) resources are scarce in low- and middle-income countries (LMICs), with a median of 0.7 beds per 100,000 population. The Modified Early Warning Score (MEWS) aids early identification of clinical deterioration and supports standardized escalation and ICU triage decisions to optimize critical-care resource use. However, adherence to MEWS protocols remains inconsistent.
RESEARCH QUESTION: Does adherence to MEWS-based ICU admission criteria improve patient outcomes, and which admission categories/service lines contribute most to MEWS-threshold discordance?
METHODS: We conducted a descriptive cross-sectional baseline needs assessment of 120 adult ICU admissions over 3 months at a university teaching tertiary referral centre in Pakistan. MEWS at ICU admission was compared against the prespecified admission threshold (MEWS ≥ 7) to quantify threshold compliance and to characterise sub-threshold admissions by admission category/service line. Mortality and ICU length of stay were analysed as secondary, exploratory outcomes using univariable logistic regression and correlation analyses, respectively.
RESULTS: Only 34.2% of ICU admissions met the prespecified MEWS threshold, while 65.8% were admitted despite sub-threshold scores, indicating substantial MEWS-threshold discordance. Discordance was most frequent in trauma-related/neurosurgical and major surgical admissions. Meeting the MEWS threshold was associated with worse survival (OR = 0.247; 95% CI: 0.111-0.548), consistent with higher illness severity among threshold-concordant admissions. This association was not adjusted for confounders due to the limited events and should be interpreted cautiously. MEWS showed minimal correlation with ICU length of stay (r = 0.05; p = 0.576).
CONCLUSION: In this baseline needs assessment, ICU admissions frequently diverged from the prespecified MEWS threshold, suggesting that clinician discretion and service-line context commonly override score-based recommendations. These findings provide an institution-specific foundation for targeted quality improvement (protocol refinement, training, audit-feedback, decision support, and step-up/HDU pathways).
PMID:42070252 | DOI:10.1111/jep.70459
