Is AI-generated advice too artificial? Tue Jul 5 Algorithms shuffle through music and recommend memes for us, but are we ready for them to give us health advice? Some insurers have started using artificial intelligence (AI) to scan patient medical records to efficiently detect ways to improve their help. So far, the AI can help identify the risks for a potential or worsening condition including obesity, depression, heart disease, and more. It then contacts the patient through phone call, email, or text to enroll them in a treatment or prevention program. Problem is, hardly anyone responds. While medical professionals recognize that patients have no obligation to enroll, they fear the AI-based approach isn't personalized enough to capture patient attention. One company began training nurses to contact patients directly with sensitivity and respect to their condition. This additional work resulted in a 10% increase in enrollments to the clinical programs. That said, it might not be the AI's fault entirely. A number of prior studies have shown that patients often ignore medical advice altogether. One decade-old study showed that failure to follow doctor advice leads to 10% of annual hospitalizations and over 125,000 deaths each year. And research is continuing to show how it's not enough to simply hear about health risks and proper care. Humans need to internalize the information and understand it to be motivated enough to follow it. So while AI may make it easier to match patients to treatment plans, communication is still a strictly people business. | |
No comments:
Post a Comment