Background and Objective: Influenza-like illness (ILI) leads to a substantial disease burden every winter in Europe; however, oseltamivir is not frequently prescribed to ILI patients in the primary-care setting. An open-label, multi-country, multi-season, randomised controlled trial investigated the effectiveness of oseltamivir for treating ILI in 15 European countries.
We aimed to evaluate whether patients presenting with ILI in primary care and being managed with the addition of oseltamivir to usual care had lower average direct and indirect costs compared to patients with usual care alone. Methods: Resource use data were extracted from participants' daily diaries.
Itemised country-specific unit costs were collected through official tariffs, pharmacies or literature. Costs were converted to 2018 values.
The null hypothesis was tested based on one-sided credible intervals (CrIs) obtained by bootstrapping. Base-case analysis estimated direct cost and productivity losses using itemised costed resource use and the human capital approach.
Scenario analyses with self-reported spending rather than itemised costing were also performed. Results: Patients receiving oseltamivir (N = 1306) reported fewer healthcare visits, medication uses, hospital attendances and paid-work hours lost than the other patients (N = 1298).
Excluding the oseltamivir cost, the average direct costs were lower in patients treated with oseltamivir from all perspectives, but these differences were not statistically significant (perspective of patient: EUR17 [0-95% Crl: 16-19] vs. EUR24 [5-100% Crl: 18-29]; healthcare provider: EUR37 [28-67] vs.
EUR44 [25-55]; healthcare payers: EUR54 [45-85] vs. EUR68 [45-81]; and society: EUR423 [399-478] vs.
EUR451 [390-478]). Scenario and age-group analyses confirmed these findings, but with some between-country differences.
Conclusion: The average direct and indirect costs were consistently lower in patients treated with oseltamivir than in patients without from four perspectives (excluding the oseltamivir cost). However, these differences were not statistically significant.