Forecasting trends of rising emergency department chest imaging using machine learning.

Introduction: Imaging studies in the acute care setting, such as the emergency room, have been increasing. In this report, we use the Centers for Medicare and Medicaid services (CMS) database to assess trends in ED chest CT and chest CTA imaging in ED from 2010 to 2021. In addition, we forecast the utilization of these imaging modalities until 2030 to predict volume of studies using machine learning models.

Methods: CMS data regarding volume of procedures from 2010 to 2021 for CPT Codes: Chest CT: 71,250, 71,260, 71,270; and Chest CTA 71,275 was obtained. IBM SPSS Statistical Software was used to assess compound annual growth rate (CAGR) to identify trends in ED center (ED) imaging. For predictive time series forecast, and autoregressive integrated moving average (ARIMA) model was constructed using R Studio statistical software. Ljung-Box test was used to check the model's residuals.

Results: Chest CT imaging in the ED setting has steadily increased from 273,063 in 2010 to 540,047 in 2021. Chest CTA imaging also saw an increase from 175,554 in 2010 to 705,727 in 2021. ED Chest CT imaging has increased from 2010 to 2021 with a CAGR of 7.5% (p < 0.001). Similarly, chest CTA imaging has also increased during this period with an average CAGR or 13.0% (p < 0.001). The ARIMA model forecast for ED CT chest imaging predicts a continued increase in volume of CT imaging reaching 614,995 in 2024 and 760,851 in 2030. Chest CTA is predicted to have an even higher increase up to 850,320 in 2024 and 1,139,505 in 2030 assuming the trend observed over the past decade is consistent.

Conclusion: From 2010 to 2021, we found an increasing trend in ED chest CT and CTA imaging with forecast predicting volume up to twice of volume in 2030 compared to a decade ago. Understanding the anticipated volumes of forthcoming ED imaging can aid in planning hospital infrastructure, personnel, and funding amongst other required healthcare resources. It is important to consider that while imaging holds a significant role in patient care, inappropriate utilization can result in unnecessary radiation exposure, prolonged wait times, and an increase in burden of healthcare resources.