Pilot study to determine feasibility of development of an influenza clinical diagnostic tool

Abstract

Background: Influenza is a costly healthcare burden worldwide. Diagnosing influenza in a timely and accurate manner allows the proper appropriation of scarce healthcare resources and antiviral medications. Unfortunately, current testing technologies are either insensitive (rapid antigen tests), nonspecific (clinical diagnosis) or untimely and expensive (rt-PCR). Ultimately, treatment and management decisions for influenza are suboptimal due to lack of diagnostic accuracy. The need for differentiating influenza from other viral infections led us to the hypothesis that influenza may have a large affect on certain vital signs allowing development of a clinical diagnostic tool. Methods: We prospectively collected data, including demographics, vital signs and rapid influenza antigen testing results from patients who presented to primary care clinics with fever and a cough or fever and a sore throat (influenza like illness) of 2 days or less duration. A total of 38 children were included in this analysis. Subjects’ heart rates were converted to a uniform unit (heart rate ratio) by dividing their heart rate by the median heart rate for age. Likelihood ratios (LR) were determined for apparently significant vital sign thresholds. Results: 14 children were influenza A positive by rapid antigen testing. Compared to rapid antigen negative controls, clinically significant likelihood ratios were determined for heart rate ratio greater than 1.4 (Positive LR = 10 (CI 1.4-76.9)) and a combination of current temperature less than 38C with a heart rate ratio less than 1.25 (Negative LR = 0.24 (CI 0.06-0.99)). Other vital signs yielded no predictive significance. Graphs of Likelihood Ratios Conclusion: Although limited by its small size and the use of the rapid antigen test as a proxy to a gold standard, we feel that this pilot study highlights the feasibility of developing a clinical influenza diagnostic tool. Children presenting for medical care with HR ratios >1.4 may be prioritized to treatment based on the higher post-test probability of influenza. Perhaps of more utility, patients with Temp <38C upon presentation with a heart rate ratio of less than 1.25 may have low enough post-test probability to avoid antiviral therapy or allow earlier return to school. We feel this issue deserves additional study. Nomogram Showing Pre-test to Post-test change in Probability Abstracts for SupplementInternational Journal of Infectious DiseasesVol. 14Preview Full-Text PDF Open Archive