Modelling the Impact and Cost-Effectiveness of Biomarker Tests as Compared with Pathogen-Specific Diagnostics in the Management of Undifferentiated Fever in Remote Tropical Settings

Abstract

Background: Malaria accounts for a small fraction of febrile cases in increasingly large areas of the malaria endemic world. Point-of-care tests to improve the management of non-malarial fevers appropriate for primary care are few, consisting of either diagnostic tests for specific pathogens or testing for biomarkers of host response that indicate whether antibiotics might be required. The impact and cost-effectiveness of these approaches are relatively unexplored and methods to do so are not well-developed.

Methods: We model the ability of dengue and scrub typhus rapid tests to inform antibiotic treatment, as compared with testing for elevated C-Reactive Protein (CRP), a biomarker of host-inflammation. Using data on causes of fever in rural Laos, we estimate the proportion of outpatients that would be correctly classified as requiring an antibiotic and the likely cost-effectiveness of the approaches.

Results: Use of either pathogen-specific test slightly increased the proportion of patients correctly classified as requiring antibiotics. CRP testing was consistently superior to the pathogen-specific tests, despite heterogeneity in causes of fever. All testing strategies are likely to result in higher average costs, but only the scrub typhus and CRP tests are likely to be cost-effective when considering direct health benefits, with median cost per disability adjusted life year averted of approximately $48 USD and $94 USD, respectively.

Conclusions: Testing for viral infections is unlikely to be cost-effective when considering only direct health benefits to patients. Testing for prevalent bacterial pathogens can be cost-effective, having the benefit of informing not only whether treatment is required, but also as to the most appropriate antibiotic; this advantage, however, varies widely in response to heterogeneity in causes of fever. Testing for biomarkers of host inflammation is likely to be consistently cost-effective despite high heterogeneity, and can also offer substantial reductions in over-use of antimicrobials in viral infections.

Fig 1. Algorithms for antibiotic prescription for each of the strategies.

The yellow circular nodes in the pathogen-specific branches replicate current practice in terms of whether an antibiotic is prescribed and the choice of antibiotic. Patients with elevated CRP receive an antibiotic, with the choice of drug reflecting current practice.

Fig 2. Percentage of microbiologically confirmed diagnoses out of all outpatients, and of those the proportion prescribed an antibiotic.

JEV–Japanese encephalitis virus.

Fig 3

Antibiotic targeting using the Laos data (A) and across a range of simulated incidences and test accuracies (B). Figure A illustrates that CRP testing achieved the largest proportion of patients that are correctly prescribed an antibiotic (the bottom dark green segment of the bars show the percentages of patients with viral infections not prescribed an antibiotic; the light green segment shows the percentages of patients with bacterial infections correctly prescribed an antibiotic; the yellow segment shows the percentages of patients with viral infections prescribed an antibiotic; and the red segment shows the percentages of patients with untreated bacterial infections). The panel on the left shows that this advantage of CRP testing is consistent when modelling extensive heterogeneity in causes of fever, while the pathogen specific tests and the scrub typhus one in particular are more affected by such heterogeneity. ST–scrub typhus; CRP–C reactive protein; RDT–rapid diagnostic test.

Fig 4. Cost-effectiveness plane and cost-effectiveness acceptability curves for the three strategies when compared with a baseline of current practice.

The blue lines indicate a willingness to pay threshold of $1400, approximating the Laos GDP/capita, while the green line is a conservative willingness to pay threshold of $150. The Dengue test in most instances was associated with little or no advantage in terms of health outcomes while resulting in higher costs than current practice. The scrub typhus and CRP tests offered direct health benefits over current practice, but also at a higher cost. When accounting for parameter uncertainty, the scrub typhus and CRP tests are approximately 90% and 80% likely to be cost-effective at a willingness to pay threshold of $1400. ST–scrub typhus; CRP–C reactive protein; RDT–rapid diagnostic test.