Cost-Effective Diagnostic Algorithms for Ugandan Patients with Solid Tumors Who Develop Chemotherapy-Associated...

In low- and middle-income settings, microbiologic evaluation of chemotherapy-associated febrile illness is limited by cost. Cost-effective diagnostic algorithms could streamline evaluation for chemotherapy-associated febrile illness where comprehensive testing is not possible, particularly in HIV- and tuberculosis (TB)-endemic settings. In this study, we created a decision analytic model to evaluate costs, diagnostic yield, and cost-effectiveness of diagnostic algorithms for adult inpatients with solid tumors who developed chemotherapy-associated febrile illness in Uganda. Given the high prevalence of HIV, TB, and malaria, diagnostics included serum cryptococcal antigen (CrAg) lateral flow assay, Alere TB urinary lipoarabinomannan (LAM), malaria rapid testing, and aerobic blood cultures. We considered the following testing algorithms: 1) a comprehensive approach where all tests were performed on all participants and 2) stepwise algorithms where tests were performed in series and stopped with the first positive result. Prevalence data and test performance were taken from the published literature. A comprehensive testing strategy yielded 32.2% correct diagnoses at a cost of $97.85 per correct diagnosis. Of the stepwise strategies, testing CrAg first yielded the cheapest strategy at a cost of $85.49 per correct diagnosis and the highest number of appropriate diagnoses (29.9%). The incremental cost-effectiveness ratio was $234 per additional correct diagnosis for the comprehensive strategy when compared with the sequential strategy. Although comprehensive diagnostic testing is optimal for patients who experience chemotherapy-associated febrile illness, if this strategy is unavailable, stepwise testing with CrAg first and then, TB LAM is optimal to maximize correct diagnosis and minimize costs.