Objective This study aimed to understand hospital doctors’ priorities (target use cases and aetiologies) for the development of a new rapid diagnostic test for patients with fever. Design A cross-sectional online survey. Setting Europe-wide. Participants Secondary and tertiary care doctors involved in patient assessment and diagnosis across Europe. Intervention Online survey from April to September 2024. Main outcome measures Importance of developing a new test on a scale of 1–10 for up to 19 ‘use cases’ (types of febrile presentations in specific demographic groups): use case scores and ranks and differences across subgroups of respondents, with free text to capture additional suggestions; respondents’ preferences (multiple choice) regarding which aetiologies should be included in a new test. Results 265 respondents from 30 European countries (out of 270 starting the survey) were included in the analysis. Top priorities included febrile immunocompromised patients and fever without a focus for both paediatric and adult use cases, and 1–3 months old febrile infants. Rankings were similar across clinician subgroups despite some differences in average scores. 92% (243/263), 95% CI 89% to 95%, of respondents would find a ‘generic’ test for bacterial aetiology useful, even if it does not differentiate between Gram-positive and Gram-negative aetiologies. 54% (63/116), 95% CI 45% to 63%, of respondents would find a ‘generic’ test for inflammatory aetiology useful when seeking to diagnose children for whom Kawasaki’s disease (KD) is on the differential, even in the absence of any KD-specific test, 83% (96/116), 95% CI 75% to 89%, would find such a ‘generic’ test useful if they could use it alongside a KD test when desired. Conclusion Clinicians prioritise the most vulnerable patients (because of age or comorbidities) and unclear presentations (fever without a focus) for the development of a new fever diagnostic test. Even relatively simple (eg, bacterial, inflammatory) tests could provide added value to most clinicians.
Bonnet et al. (2026) studied this question.