Researchers from the University of Bonn and the University Hospital Bonn (UKB) have thoroughly examined the body of research on AI's impact. They demonstrated that AI does not always result in faster work processes. Their findings have now been published in the journal npj Digital Medicine.
Image Credit: everything possible/Shutterstock.com
Artificial intelligence (AI) is being increasingly used in hospitals and patient care. AI has long been a part of routine clinical practice, particularly in specialized fields like radiology, which involves a lot of imaging. However, there is still much to learn about how much AI affects clinical workflows.
The true effects on work processes are unknown, even though AI is frequently viewed as a solution for managing repetitive tasks like patient monitoring, care task documentation, and clinical decision support. There is a dearth of trustworthy data on efficiency gains, especially in data-intensive specialties like radiology, pathology, and genomics, where AI is already being used to identify patterns in massive amounts of data and prioritize cases.
We wanted to find out to what extent AI solutions actually improve efficiency in medical imaging. The widespread assumption that AI automatically speeds up work processes often falls short.
Katharina Wenderott, Study Lead Author and Doctoral Student, Institute for Patient Safety, University of Bonn
Consistent Evaluation of Studies is Difficult
The research team carried out a systematic review of 48 studies that looked at the application of AI tools in clinical settings—specifically in radiology and gastroenterology. No notable efficiency gains were found in the meta-analyses of the 33 studies that examined the processing time of work processes. However, 67% of them reported a reduction in working hours.
“Some studies showed statistically significant differences, but these were insufficient to draw general conclusions,” Wenderott added.
The team also examined how well AI is incorporated into current processes. It was discovered that the particular circumstances and procedures at the location significantly impact the implementation's success. However, it was challenging to carry out a consistent evaluation due to the diversity of the technologies and study designs.
Our results make it clear that the use of AI in everyday clinical practice must be considered in a differentiated way. Local conditions and individual work processes have a major influence on the success of implementation.
Matthias Weigl, Director, Institute for Patient Safety, University of Bonn
The study offers crucial preliminary information about the potential effects of AI technologies on clinical work procedures.
Prof. Weigl concluded, “A key finding is the need for clearly structured reporting in future studies in order to better evaluate the scientific and practical benefits of these technologies.”
Journal Reference:
Wenderott, K. et. al. (2024) Effects of artificial intelligence implementation on efficiency in medical imaging—a systematic literature review and meta-analysis. npj Digital Medicine. doi.org/10.1038/s41746-024-01248-9