# P-617. Multi-Parametric Diagnostic Algorithm for Early Detection of Severe Acute Respiratory Infections in Children: A Prospective Multi-Center Study in Resource-Limited Settings

## TL;DR

A new diagnostic tool combining clinical signs and biomarkers helps quickly identify severe respiratory infections in children in areas with limited resources.

## Contribution

A novel multi-parametric algorithm improves early detection of severe respiratory infections in children using point-of-care biomarkers and clinical indicators.

## Key findings

- The algorithm achieved 92.4% sensitivity and 89.7% specificity for identifying severe respiratory infections.
- It reduced treatment initiation time from 8.4 to 2.7 hours and cut unnecessary antibiotic use by 41.6%.
- The algorithm distinguished bacterial from viral infections with 87.3% accuracy.

## Abstract

Pediatric respiratory infections remain a leading cause of morbidity and mortality in resource-limited settings, with diagnostic challenges due to limited laboratory infrastructure and delayed access to healthcare. This study evaluates a novel multi-parametric diagnostic algorithm combining clinical indicators with point-of-care biomarkers for early identification of severe acute respiratory infections (SARIs) in children under 5 years.

We conducted a prospective, multi-center observational study across 12 healthcare facilities (8 rural, 4 urban) in five states of India from June 23 to February 25. A total of 3,240 children aged 3-59 months presenting with respiratory symptoms were enrolled. Our diagnostic algorithm integrated clinical parameters (respiratory rate, oxygen saturation, chest indrawing) with point-of-care biomarkers (CRP, PCT, and neutrophil-lymphocyte ratio) in a mobile application-based scoring system. The algorithm's performance was compared against a reference standard comprising radiological findings, comprehensive laboratory testing, and 14-day clinical outcomes.

The multi-parametric algorithm demonstrated 92.4% sensitivity and 89.7% specificity for identifying severe respiratory infections requiring intervention, significantly outperforming standard clinical assessment alone (sensitivity 78.6%, specificity 72.3%; p< 0.001). Implementation of the algorithm reduced time to appropriate treatment initiation from 8.4 to 2.7 hours (p< 0.001) and decreased unnecessary antibiotic use by 41.6%. The algorithm effectively distinguished bacterial from viral etiologies with 87.3% accuracy. Cost analysis revealed a 38.2% reduction in overall healthcare expenditure per case. Field health workers demonstrated 96.8% proficiency in algorithm application after a standardized 4-hour training program.

Our novel multi-parametric diagnostic algorithm provides an accurate, cost-effective approach for early identification of severe pediatric respiratory infections in resource-limited settings. The integration of clinical parameters with accessible point-of-care biomarkers enables rapid triage and appropriate management decisions where comprehensive laboratory facilities are unavailable.

All Authors: No reported disclosures