Referral Decision Models for Febrile Children in Resource-Limited Asian Communities
Overview
This study developed and validated clinical prediction models integrating vital signs, danger signs, pulse oximetry, and host biomarkers to guide referral decisions for febrile pediatric patients in South and Southeast Asia. Among 3,405 children studied, 3.9% progressed to severe illness, highlighting the need for improved risk stratification tools beyond current WHO guidelines.
Background
Infectious diseases cause the majority of deaths in children aged 1−59 months globally, with many fatalities occurring due to delayed or absent healthcare access. Existing WHO IMCI guidelines rely on clinical danger signs for hospital referral but have suboptimal accuracy and interobserver variability. Pulse oximetry and host biomarker testing have emerged as promising adjuncts to improve early identification of children at risk of severe infection, especially in resource-constrained community settings where referral decisions are critical.
Data Highlights
Parameter
Value
Total children screened
11,962
Eligible children
3,998 (33.4%)
Recruited children
3,423 (85.6% of eligible)
Final analyzed cohort
3,405
Median age
16.8 months (IQR 8.7−31.0)
Male participants
59.6%
Wasted children
17.2%
Stunted children
19.5%
Children with severe acute malnutrition
42.4% of wasted
Median symptom duration
3 days (IQR 2−4)
Microbiological cause identified
26.4%
Children progressing to severe illness (death or organ support)
3.9%
Key Findings
Among 3,405 febrile children, 133 (3.9%) progressed to severe illness requiring organ support or resulting in death within 2 days.
Current WHO IMCI danger signs have limited sensitivity and specificity for early identification of severe febrile illness.
Pulse oximetry identified hypoxemia, a strong predictor of poor outcomes, but remains underutilized due to practical barriers in community settings.
Host biomarker testing showed promising prognostic performance and feasibility when integrated with clinical assessment by community healthcare workers.
Multivariable clinical prediction models combining vital signs, danger signs, pulse oximetry, and biomarkers improved risk stratification for referral decisions.
Most children lived within 1 hour of hospital, but referral barriers and early recognition challenges persist in resource-limited communities.
Clinical Implications
Incorporating pulse oximetry and host biomarker testing into community-level assessments can enhance early identification of febrile children at risk of severe illness, enabling timely referral and treatment. Training and resource allocation to overcome barriers to pulse oximetry use and biomarker testing are essential to improve outcomes in resource-constrained settings. These integrated models may optimize referral decisions, reducing missed critical illness and unnecessary hospital transfers.
Conclusion
This study supports the integration of simple clinical parameters with pulse oximetry and host biomarker testing to improve referral decision-making for febrile children in resource-limited South and Southeast Asian communities. Such approaches hold promise to reduce mortality by enabling earlier recognition and management of severe infections.
References
WHO Integrated Management of Childhood Illnesses (IMCI) Guidelines
Study Dataset and Analysis, 2020-2022
Pulse Oximetry and Host Biomarkers in Pediatric Infection Prognosis
