Diagnostic testing or empirical neuraminidase inhibitor therapy for patients with influenza-like illness: What a difference a day makes

Objective: To determine the clinical and economic tradeoffs of available diagnostic and treatment options for patients with suspected influenza infection. Design: Decision analytic model. Patients: A simulated cohort of adults with influenza-like illness of less than 48 hours' duration. Interventions: Patients received (1) no testing or treatment; (2) rapid testing for influenza, with neuraminidase inhibitors (NIs) prescribed for individuals who test positive; or (3) empirical NI therapy. A validated prediction rule to determine the probability of influenza infection based on patient signs and symptoms is available. The monetary value attributable to the incremental benefits of NI therapy must be quantified during the clinician visit. Outcome: Incremental cost per case of influenza treated with NIs. Results: The decision whether to treat, test, or prescribe NIs empirically for patients presenting with influenza-like illness of less than 48 hours' duration depends on the probability of influenza and the value of the benefits of NI therapy. If the incremental value of NI therapy exceeds $200 and the probability of influenza is greater than 35%, empiric NI therapy is the preferred option. The test option has a very narrow band of preference due to suboptimal test characteristics and cost. Conclusion: The decision whether to prescribe NIs based on clinical criteria or restrict NI use to patients with a confirmed laboratory diagnosis of influenza depends on the likelihood of infection, the accuracy and cost of the diagnostic test, and the benefits attributable to NI therapy. Clinicians must consider these factors when determining cost-effective use of NIs.