The Sibyl technology comes from the developers of macro-eyes, which is a machine learning company centered on personalizing patient care. The new technology allows care practices to manage the points in the day when patients do not turn up for appointments. In the U.S., for example, it has been estimated that patients not turning up for medical appointments accounts for some 15 percent of all appointments made, although for ‘worst case’ situations ‘no shows’ can account for up to 40 percent of the day’s bookings.
Costing medical centers money
According to Benjamin Fels, CEO of macro-eyes, in a message sent to Digital Journal: “No-Shows and lack of optimization in scheduling costs healthcare providers billions, hits morale, strains operations and has implications on care that can cost lives.” macro-eyes is a Seattle-based machine learning company creating products that give healthcare providers context and insight that simplifies personalized patient care.
This was the reason why his company developed Sibyl, aiming to solve the appointments gap problem with machine learning. Sibyl harnesses artificial intelligence that is capable of learning when to schedule individual patients in order to increase utilization. While much of healthcare is increasingly becoming data-driven, Fels noted that scheduling was not and there was a clear need for predictive analytics.
Predictive scheduling
Through predictive scheduling the Sybil platform can learn the appointment times that are best-fit for both the patient and provider, which helps to increase the utilization overall. The software can be fitted as an add-on to existing scheduling systems and it can guide the practice manager by making appointment recommendations for each patient. This is through using patterns in behavior to learn when patients are most likely to show based on statistics.
The learning component of the artificial intelligence was ‘trained’ over the course of several years at leading academic medical centers in New York and California. Here appointment histories and data points across different providers, relating to patient, location, time and type of care, were analyzed. To add to this other variables were factored in, which can impact upon whether a patient decided to show for an appointment of not. These variables included weather patterns, air quality, and traffic and transport data.
For the late-stage development of the software, macro-eyes worked with 20 clinics across the U.S. to review 2 million anonymous appointment records. What was missing from the data given to Sybil was whether the appointment was met or not. Sibyl predicted actual patient outcomes with 76 percent accuracy.
The use of technology like Sibyl shows how every aspect of healthcare can utilize data more efficiently and improve workflow interruptions.
