Many hospital systems are turning to artificial intelligence to analyze patient data and identify people at risk of comorbidities. But there are pitfalls CIOs need to look for, according to a May 17 CIO report.
Having ethics and governance plans in place can prevent problems from popping up. One community hospital built a data system to predict whether patients might have comorbidities, Carm Taglienti, a distinguished engineer at IT company Insight, told CIO.
The work was important because a patient may visit the hospital for one reason, but have a separate potentially fatal health issue. If the hospital doesn't detect the second issue, the patient could be sent home and ultimately die.
But the hospital had to balance the AI forecasts with the risk of subjecting people to unnecessary tests that may strain resources. The facility opted to invest more in data collection to get more accurate predictions, Taglienti said.
Additionally, hospital officials asked whether it could lean more on nurse practitioners or automation to follow up with patients instead of relying on physicians. The hospital also added a way to schedule visits with ER physicians so patients without primary care providers could go at less busy times.