Desirability of outcome ranking (DOOR) has been developed for assessing desirability of outcome in interventional studies. However, its possible use in observational studies of the diagnosis and early treatment of infectious diseases has not been explored so far, and it might introduce interesting features in specific scenarios. This was a post hoc analysis of a prospective observational study in intensive care unit patients with sepsis and at risk of candidemia. The probabilities that a randomly selected patient would have a more, less, and equally cost-effective early therapeutic choice following a BDG-based diagnostic strategy rather than the empirical administration of antifungals to all patients were calculated using DOOR methods. The probability of a more cost-effective therapeutic choice following the BDG-based rather than the empirical strategy was 67.81% (95% CI 67.32-68.30), whereas the probabilities of a less and equally cost-effective early therapeutic choice were 19.68% (95% CI 19.27-20.10) and 12.50% (95% CI 12.16-12.85), respectively. The application of DOOR methods to observational studies focused on diagnosis and early treatment is a novel field that could merit further investigation.
Keywords: Biomarker; Candidemia; DOOR; Diagnosis; Glucan.