Though machine learning has the power to dramatically change healthcare, physicians should not fear being overrun by robots, according to the Harvard Business Review.
Four healthcare leaders — Derek Haas, CEO of Avant-garde Health, Eric Makhni, MD, orthopedic surgeon at Henry Ford Health System, Joseph Schwab, MD, chief of spine surgery at Massachusetts General Hospital, and John Halamka, MD, executive director of the Health Technology Exploration Center at Beth Israel Lahey Health — debunked three common myths about machine learning in healthcare.
1. Machine learning can do much of what physicians do. Although machine learning can prevent patients from getting sick and diagnose a patient, the software can't provide care and treatment to patients. Additionally, machine learning lacks the human element in healthcare.
2. Big data and brilliant data scientists are always a recipe for success. While having data can be powerful, not all data is sufficient and necessary. Healthcare organizations must collect the right data and fully understand it. Questions to ask are:
- How was the data gathered?
- For what purpose was the data gathered?
- What are potential problems with or limitations of the data?
- Have circumstances changed?
3. Successful algorithms will be adopted and utilized. Just because a new software or algorithm is developed, doesn't mean it will be used among physicians. This is often because the tools aren't integrated into the workflows of clinicians. If an algorithm isn't part of the EHR, clinicians won't use it.
To read the full report, click here.