A systematic review and meta-analysis finds that the definition of severe acute respiratory infection (SARI) used by the World Health Organization (WHO) demonstrated reduced sensitivity and low specificity in children, suggesting that surveillance systems that rely on SARI case definitions may under-report the burden of diseases like influenza and respiratory syncytial virus (RSV) infection in children.

For the analysis, published recently in JAMA Network Open, researchers in Ontario identified 1,144 studies and included 13 that assessed surveillance data from 65 hospitals in eight countries (Canada, Egypt, Ghana, India, Jordan, Kenya, New Zealand, and South Africa). Data were from pediatric patients hospitalized any time from 2007 to 2023. 

The authors wrote, “Several studies have suggested that SARI case definitions are inaccurate at detecting pediatric disease burden. Understanding the performance of SARI case definitions in children is important for pandemic preparedness.”

Sensitivity decreased greatly in younger kids

The most common SARI definition was the one developed by the WHO in 2014 (used in nine studies). Other studies used the WHO 2011 SARI definition (two studies), a general Integrated Management for Childhood Illness (IMCI) pneumonia definition (one study), and a combined definition consisting of five other definitions (one study). For children younger than five years, the WHO recommended using the IMCI definitions of pneumonia and severe pneumonia in addition to its 2014 SARI case definition.

Ten of the studies assessed the SARI definition for flu, and six for RSV.

The investigators found that the WHO 2014 SARI definition yielded a sensitivity of 75.7% and specificity of 30.6% for flu and a sensitivity of 70.6% and specificity of 38.7% for RSV. They also found that sensitivity decreased greatly in younger age-groups.

Sensitivity measures how well a test can correctly identify people with a disease (true-positive rate) and minimize false-negatives. Specificity measures a test’s ability to correctly identify patients without a disease (true-negative rate) and minimize false-positives.

Two studies evaluated the WHO 2011 SARI definition for flu and RSV. For influenza, one study reported that the 2011 definition yielded a sensitivity of 68.7% and a specificity of 18.0%. For RSV, another study reported that it yielded a sensitivity of 51.2% and a specificity of 80.9%.

In summing up the data, the authors of the review write, “These findings suggest that existing viral surveillance resources that rely on SARI case definitions may underestimate disease burden in young pediatric cohorts.”

Absence of COVID-19 assessment

The authors note, “The included studies evaluated case definitions across a range of countries of various regions and wealth categories, most prior to the COVID-19 pandemic, with a focus almost entirely on the younger than 5 years age group. The diagnostic accuracy of the WHO SARI case definition varied.”

They add, “Our study highlights the limited understanding of the diagnostic accuracy of the SARI case definition in pediatrics. We further highlight the near complete absence of viruses other than influenza and RSV, notably SARS-CoV-2.”

A key implication of the observed lower sensitivity of SARI case definitions among younger cohorts is the potential for the underestimation of disease burden.

The study authors conclude, “A key implication of the observed lower sensitivity of SARI case definitions among younger cohorts is the potential for the underestimation of disease burden. SARI sentinel surveillance systems are routinely used to estimate viral respiratory disease burden by calculating the incidence of virus-associated SARI for a given pathogen and extrapolating impact on the larger population, per WHO guidelines on this practice. 

“Our findings confirm that such calculations of incidence among younger cohorts should be adjusted to consider the sensitivity of the case definition applied.”