A four-year research study conducted across 82 hospitals in the United States has showcased the benefits of integrating artificial intelligence (AI) in healthcare systems. The study demonstrated the use of an AI tool in rapidly identifying potential infection outbreaks, thereby equipping clinicians with sufficient time to contain and treat the patients inflicted. The automated tool was tested against traditional methods of identifying infection outbreaks and exhibited a significant reduction of 64% in potential outbreaks. The AI tool successfully identified an average of three potential outbreaks per annum for each hospital.
Given that infection outbreaks are frequently overlooked or discovered too late, post preventable infections, the implementation of this study’s findings and the associated AI tool are instrumental in providing a standardized and practical approach to identifying early transmission of infections and preventing further outbreak escalation. The study was made possible through funding from the U.S. Centers for Disease Control and Prevention (CDC) and conducted between 2019-22 within the HCA Healthcare system. A collaborative effort by a team of investigators from HCA, Harvard Pilgrim Health Care Institute, and University of California, Irvine (UCI) Health, the research aimed to equip a healthcare industry that is still grappling with COVID-19’s effects.
The study’s results, though impacted by the pandemic, underline the potential for AI in managing infection rates in healthcare settings. Researchers created an ‘algorithm-driven statistical detection tool’ that analyzed laboratory data for more than 100 bacterial and fungal infection signs. Key to the process was automated patient clinical culture review and statistical analysis of increasing patient numbers with these infections. While the efficiency of the automated alert system was affected due to the pandemic as staff couldn’t timely respond to the alerts, the research team focused on pre-pandemic results. The tool’s underlying software will be made accessible to all healthcare systems, provided that it can be integrated into Electronic Health Records (EHR) and other clinical workflow platforms.
Source: https://www.healthleadersmedia.com/technology/researchers-use-ai-identify-healthcare-based-outbreaks