Much of the focus on artificial intelligence (AI) in healthcare has revolved around clinical insights, but a new report from Frost & Sullivan points out other ways hospitals can use the technology to enhance their operations.
The report, commissioned by Intel, claims that artificial intelligence is often “misperceived as a tool that could eliminate or reduce certain types of staff positions.” Further, that overwhelming hype can strip potential users, particularly those in healthcare, of a true understanding of how they can use it, according to the study.
Workflow optimization, waste reduction, and data interpretation are all key areas that AI can help hospitals refine in order to save time, money, and improve patient safety, the authors said.
Stanford’s Lucile Packard Children’s Hospital is piloting AI as a means to improve a seemingly simple yet essential safety function: hand hygiene protocol. By placing advanced sensors near sanitizer dispensers and processing captured images through a convolutional neural network, the hospital is tracking the frequency and effectiveness of staff hand sanitization.
That’s just one example mentioned by the authors. Other ongoing applications include a pilot at Brigham and Women’s Hospital to evaluate an AI-enabled continuous home monitoring system for chronic conditions; multiple natural-language processing efforts to streamline documentation digitization; and the ELLI.Q system that Intuition Robotics is developing to integrate care and family socialization for aging-in-place patients.
And there is a warning embedded in the whitepaper: “Traditional healthcare providers reluctant or unwilling to embrace technology solutions in a similar way could see significant attrition as these other entrants target certain types of primary care services.”
Before engaging any number of the possible AI applications in a healthcare system, the report makes a few recommendations. Organizations need to undertake a thorough internal review to find weaknesses that it could help address. IT infrastructure must be prepared to support necessary data capture and storage, alluding to the adage, “Garbage data in, results in, garbage results out.”
Another misconception that Frost & Sullivan’s work seeks to address is that AI technologies, and the IT upgrades they may require, are prohibitively expensive. “In reality, depending on the age and sophistication of systems in place, most facilities are able to implement programs with minimal upgrades to IT hardware,” they write, adding that the increasing proliferation of open-source platforms and innovative pricing models create affordable options.