Using point-of-care diagnostic testing for improved antibiotic prescription : an economic model

Background: Antibiotics have been overprescribed to treat infectious diseases and have generated antimicrobial resistances that reduce their effectiveness. Following the rationale behind the new paradigm of personalized medicine, point-of-care diagnostic testing (POCT) has been proposed to improve the quality of antibiotic prescription with the aim of reducing antimicrobial resistances. Methods: In order to understand whether this recommendation is valid, we create a theoretical economic model to determine under which conditions the expected benefits of using POCT to guide antibiotic prescription are greater than for empiric prescription, where we define the expected benefits as the difference between the economic value of health and the costs of the treatment. We consider the interaction of a group of physicians who express differing levels of uncertainty when prescribing with a firm selling a diagnostic device, and analyse the firm's pricing policy and the physicians' prescribing decisions. We allow the physicians to internalize the external costs of antimicrobial resistances. Results: We find that the use of POCT reduces the number of antibiotic prescriptions. The reduction in antibiotic prescriptions is higher when physicians internalise the costs of antimicrobial resistances. Physicians with relatively high levels of uncertainty use POCT as they are uncertain about the right treatment for a large proportion of patients. Physicians with low levels of uncertainty prefer to prescribe empirically. The segmentation in the population of physicians regarding the uptake of POCT depends on the distribution of levels of uncertainty across physicians. For each test, the firm charges the marginal production costs of the inputs needed to administer the test, and makes its profit from the sales of the testing devices. Conclusions: From a theoretical perspective, our findings corroborate the fact that POCT improve the quality of antibiotic prescription and reduce the number of prescriptions. Nevertheless, their use is not always recommended as empiric therapy may be preferred when uncertainty is low.