Hospitals are increasingly adopting artificial intelligence-powered tools to streamline workflows and support clinical decision making, but not all patients are sold on the technology.
Researchers from University Park, Pa.-based Penn State and the University of California, Santa Barbara designed five chatbots for a study with 295 participants and paired them with either a human physician, AI chatbot, or AI-assisted physician.
The study's results, released in May, showed that when AI chatbots used the first name of patients and referred to their medical history, study participants were more likely to find the bot intrusive and less likely to take the bot's medical advice.
"Many patients feel that AI doesn’t take into account their distinctive characteristics and circumstances," Zafar Chaudry, MD, senior vice president and CIO at Seattle Children's, told Becker's. "However, AI-powered medical tools are primarily for decision support and have been shown to perform with expert-level accuracy."
Dr. Chaudry pointed out that the tools could bring significant benefits to patients, including better healthcare accessibility, early disease detection and healthcare cost reductions.
Still, many patients feel there is no way they can trust technology the same way they trust their physician, who went through years of schooling and training to be qualified to treat patients.
Replicating clinicians' thought processes results in complex neural networks that need vast amounts of data for training and years of pilot studies, according to Myra Davis, the chief information innovation officer at Texas Children's Hospital in Houston.
"Rest assured, the creation of these complex neural networks, which drive the AI-powered medical tools in question, are created by the clinicians themselves," Ms. Davis said.
She also pointed out that AI-powered healthcare tools take great advantage of data. If they're properly trained on enough EHR data, they can generate diagnoses with high accuracy and recommend treatments.
Patients should also feel more confident about clinical AI tools knowing they go through various quality assurance and FDA approval processes, according to Ms. Davis.
"AI in healthcare is a budding and exciting field, which will result in better informed decisions that will, if applied properly, improve population health," she said.