by John R. Fischer
, Senior Reporter | January 30, 2020
While expected to increasingly adopt AI for their practices, providers and imaging centers still face a number of barriers when it comes to implementing it.
That’s what healthcare leaders and radiologists indicated in their responses to the 2019 AI Use in Imaging Survey
conducted by Definitive Healthcare, a marketing firm specializing in data, intelligence and analytical insights on the healthcare provider market. One third of organizations who responded indicated using AI as part of imaging in their practices.
"There is a very real opportunity and developing trend to utilize AI technology to augment the knowledge and experience of radiologists and imaging professionals," Matt Valley, senior healthcare analyst at Definitive Healthcare, told HCB News. "Clinical expertise combined with AI technology opens the door for faster imaging read times, the potential for more thorough analysis, and early disease detection or diagnosis, thus allowing for better patient care. Responding organizations identified disease state detection as the primary use for both current adopters, and those with plans to adopt AI technology, indicating the trend will continue to grow.
The survey was completed between October and December of 2019, and focused on adoption rates, primary areas of use and greatest challenges associated with implementing AI for imaging. The choice to examine these areas stemmed from the abundance studies and use cases showing the correlation between the effectiveness and progress of using AI in imaging and increases in adoption. Both acute hospitals and imaging centers are beginning to implement the technology as advancements, with successful use cases continuing to emerge.
Among respondents, 93 percent reported using AI for computer-aided image detection of disease states, according to the study. Process or workflow improvement was the next area in which it was used the most, by 27 percent. Following behind it were technological monitoring for equipment maintenance (16 percent); computer-aided image detection for fractures and musculoskeletal injuries (15 percent); care guideline consultation and suggestive care options (13 percent); and financial or revenue-related performance assistance (13 percent).
The presence of these solutions is reported by 57 percent to improve or assist in the accuracy of diagnosis. Another 18 percent have observed improvements in operational tasks and workflow, and 15 percent say it has reduced time of diagnosis and enabled early detection of diseases. Nine percent indicated the greatest benefit to be specific improvements to existing technologies such as CT, MR and mammograms, while one percent claimed reduced costs as the best.
Despite these positives, implementing AI is still a challenging task, with 55 percent viewing cost as a major barrier. Thirty-five percent report a lack of strategic direction as an issue, while 33 percent said more IT personnel and data scientists are needed to provide technical expertise. Other challenges are lack of necessary IT infrastructure (32 percent); regulatory guidelines (26 percent); lack of clinical expertise (22 percent); lack of leadership buy-in (19 percent); current use cases being too narrow (18 percent); and cybersecurity concerns (16 percent).
"Education is the primary factor; understanding the cost, staffing needs, and expertise required are all important components, and that can only be achieved through communicating with experienced organizations and professionals, reading up on the matter, and making it a priority," said Valley. "Organizations need to begin planning for the adoption of AI technology. Regardless of if it is in the near future or not, it appears AI is looming as one day becoming a necessary component of imaging operations."