Hospitals across the country are adopting AI at an increasing rate, but there are still many kinks when it comes to measuring the success of these tools and expanding them across the enterprise. Sutter Health System, based in Northern California.
“The challenge today is that most pilots don’t think about ROI. It’s ‘Let’s go – we’ll just solve the problem.’ The danger there is that it’s too far away without having to have a conversation about AI values. is. “In an interview this week at the Vive Conference in Nashville, he must come forward as soon as possible.”
Mysore noted that at least a rough estimate of the ROI of a particular tool is necessary before it is adopted. This information can shape decisions when investment hospital leadership is willing to make.
If a hospital predicts that technology will generate modest ROI, it probably won’t invest a lot of money in advance, but if the predicted ROI is much higher, the hospital may be explained Mysore.
Different tools require different metrics, so measuring after adopting the tool’s actual ROI can be difficult. There are also different ways to measure the success of a particular tool, he added.
For example, take a look at the surrounding listening tools equipped with AI.
“Do you save time for a doctor? It’s difficult to measure – when a doctor sees 10-12 patients in half a day, how do you actually measure it? The best thing we can measure is cognitive burden, but it’s not a scientific measure. It’s just a doctor, and I feel safe and relaxed. And instead of typing something , we can have conversations,” Mysore explained.
Qualitative metrics are extremely important in some tools.
The ambientry sinking tool is one of these tools. As the healthcare industry faces a serious shortage of clinicians amid historic burnout, doctors who feel less stressed in the workplace are an important tool to pay attention.
For other technologies, quantitative metrics are more important. For example, hospitals carefully track the average length of a patient’s stay after adopting AI tools that help automate the patient’s discharge process.
There is also a new set of challenges when it’s time to scale AI solutions that work well in the pilot phase, Mysore noted.
“Maybe you have a lot of primary care doctors and you deploy it to them first, but if you deploy it to a cardiologist, a nurse or someone else, it would be very different. Primary care physicians cannot always use the same scaling function as they ask specific questions and document specific ones. Cardiologists may do very different things, so they should not use AI. It’s really important to tailor it to the patient population and the physician population,” he said.
Without clear ROI metrics and customized deployment strategies, even the most promising AI tools risk stalling during the pilot phase, Mysore said.
Source: Ekahardiwito Subagio Purwito, Getty Images
