A University of Florida (UF) Department of Anesthesiology team intends to streamline chest and urine drainage management with automated, remote, and continuous monitoring.
Samsun Lampotang, Ph.D., the Joachim S. Gravenstein Professor of Anesthesiology, is at the forefront of this project along with Nikolaus Gravenstein, M.D., the Jerome H. Modell, M.D., Distinguished Professor of Anesthesiology, and Jennifer M. Pruitt, MSN, administrative director of nursing research at UF Health Shands. Their efforts are supported by the UF W. Martin Smith Interdisciplinary Patient Safety Awards Program, which provides start-up funding to faculty and staff who create projects to reduce the likelihood of adverse events or claims and/or improve patient safety and clinical processes.
The chest and urine drainage systems used at UF Health are mechanical devices that require manual and labor-intensive monitoring, calculating, and transcribing. While there are digital drainage systems, they can be very costly. Lampotang and his colleagues’ solution is to create automated, remote, and continuous monitoring for chest and urine drainage systems. This type of monitoring could allow health care providers to receive timely clinical data and alerts, thus making the drainage process quicker and less prone to complications.
Lampotang says the automated, remote, and continuous monitoring could “provide timely clinical management information and importantly also alarm functions by converting stand-alone devices into smart and connected devices.” The team believes that this type of monitoring can be achieved by using computer vision, a branch of artificial intelligence.
The idea for their project first arose years ago while they discussed issues with traditional systems. Gravenstein proposed the idea of using a camera to estimate and track numbers and data. With the help of Chris Samouce, Ph.D., assistant scientist in the department, Lampotang began testing the capabilities of various cameras. They also experimented with OpenCV, a computer vision application, and realized that chest and urine drainage systems could be monitored using computer vision. In addition, the team has utilized HiPerGator, the UF supercomputer, to train the computer vision model to accurately read clinical settings.
“The remote, automated and continuous monitoring of drainage systems should improve quality of care and patient safety,” said Lampotang. “The decrease in bedside provider workload frees up time to do other patient care or documentation.” There is also the possibility of reducing claims against medical professionals and litigation that arise from the complications associated with traditional drainage systems.
Currently, the team is integrating their two computer vision models into one model that will monitor both urine and chest drainage systems. Verification and fine tuning in a simulated environment will take place next year, including the installation of a computer vision system at a UF Health patient care location following UF Institutional Review Board approval, with data analysis and publication of their results to follow in 2027.