A system dynamics model of infection risk, expectations, and perceptions on antibiotic prescribing in the United States

Rationale, aims, and objectives: Inappropriate antibiotic prescribing is still a major concern that can lead to devastating outcomes including antibiotic resistance. This study aimed to simulate the antibiotic prescribing behaviour by providers for acute respiratory tract infections (ARTIs) and to evaluate the impact of patient expectation, provider's perception of patient's expectation to receive a prescription, and patient's risk for bacterial infection, on the decision to prescribe.

Methods: We developed a unique system dynamics (SD) simulation model based on the significant factors that impact the interaction between provider and patient during visits for ARTIs and the decision to prescribe antibiotics. In order to validate the model for different age groups and regions in the United States, we used the sample of 53 000 ARTI patient visits made at outpatient settings between 1993 and 2015, based on the National Ambulatory Medical Care Survey (NAMCS).

Results: Simulation results reveal that physician diagnosis for prescribing antibiotics is based on physician's experience from their prior prescribing behaviour, their perception of patient's infection risk, and patient's expectation to receive antibiotics. Also, there are some variations depending on patient's age and residential region. The simulation analysis also depicts the decreasing trend in patient's expectation over the past two decades for most age groups and regions.

Conclusions: Given the high number of unnecessary prescriptions for ARTI, we found that policies are needed to influence provider's prescribing behaviour through patient's expectation and provider's perception regarding those expectations. Our simulation framework can further be used by policymakers to design and evaluate interventions that may modify the interaction between health providers and patients to optimize antibiotic prescriptions among ARTI patients for different regions and age groups.