Flu Triage With Online Test for Kids Not Specific Enough

Emma Hitt, PhD

December 13, 2012

A web-based strategy to determine whether a child with influenza-like illness requires immediate emergency department (ED) care showed good sensitivity, appropriately selecting children who did need care, but low specificity, selecting many children who did not really need ED care, according to a new report.

Rebecca Anhang Price, PhD, from the RAND Corporation in Arlington, Virginia, and colleagues report their findings in an article published online December 12 in the Archives of Pediatric and Adolescent Medicine.

According to the researchers, "self-triage" using web-based decision support may be useful in helping patients and caregivers seeking appropriate medical attention. "However, the feasibility and safety of this strategy have not previously been evaluated," they write.

Dr. Price and colleagues assessed the usability and safety of a web-based decision support tool called Strategy for Off-Site Rapid Triage for Kids. The tool is designed to translate clinical guidance developed by the Centers for Disease Control and Prevention and the American Academy of Pediatrics, and specifically their clinical algorithm produced during the 2009 novel influenza A(H1N1) pandemic, to help triage patients to determine whether they need immediate care in an ED.

The test assigns patients into 3 groups: high risk (should receive immediate care in an ED), intermediate risk (warrants expedited evaluation by a primary care provider), and low risk (can recover at home if symptoms do not get worse).

To assess the web-based test, a study was conducted in 2 pediatric EDs in the Washington, DC, area. Researchers recruited 294 parents or adult caregivers bringing in a child aged 18 years or younger with signs and symptoms meeting Centers for Disease Control and Prevention criteria for influenza-like illness: temperature of 37.8°C or more and cough and/or sore throat.

Once at the ED, the parent or caregiver was asked to take the online test, which included questions about the child's symptoms, the usability of the Web site, and the sociodemographics of the user.

Of the participants, 90% said the Web site was "very easy" to understand and use, and this held true regardless of the sociodemographics of the user.

Of the 15 patients whose initial ED visit did meet explicit criteria for clinical necessity, the web-based tool identified 14 of them as being at high risk, resulting in an overall sensitivity of 93.3% (exact 95% confidence interval [CI], 68.1% - 99.8%).

In contrast, the specificity of the online test was poor. The overall specificity was only 12.9% (95% CI, 9.2% - 17.5%); out of 271 patients whose initial ED visit was not medically necessary, only 35 were correctly classified as low (n = 28) or intermediate (n = 7) risk; the remaining 240 children were classified as high risk when they were not.

Study limitations include recruitment from the ED waiting rooms rather than from home and poor post-ED follow up. Also, because of limited funding, Spanish and other translation services were not available for patients and only 2 institutions were evaluated.

"Our findings present a cautionary tale regarding the potential effects of self-triage tools," the authors note. "Had it been made available to the public in its current form, it might have led more, rather than fewer, parents to bring their children to an ED, thereby worsening, rather than ameliorating, ED crowding."

Support for this pilot study was provided by a grant from the de Beaumont Foundation. In-kind support for the project was provided by Walgreens in the form of gift cards for participants. The authors have disclosed no relevant financial relationships.

Arch Pediatr Adolesc Med. Published online December 12, 2012. Full text