Machine learning to predict surgery for lumbar spine pain patients — key findings

Laura Dyrda -   Print  |

Nathan Xie presented research at the American Association of Neurological Surgeons Annual Scientific Meeting discussing artificial intelligence to improve referrals for degenerative lumbar spine conditions, according to Newswise.


Researchers examined data from patients at two Melbourne, Australia-based spine centers and reported wait times of more than 1,200 days for the first consultation. They recommended a strategy to evaluate primary care referrals based on likeliness the patient would go to surgery and redirect those who are more likely to undergo non-surgical management to reduce wait times.

Mr. Xie suggested using machine learning and artificial intelligence to triage spine patients. The researchers identified 55 factors associated with patients who eventually underwent surgery by reviewing medical records for patients with elective lumbar spine complaints between 2013 and 2018; they aimed to develop an artificial neural network.

The network they developed through back propagation learning was then able to predict likelihood of spine surgery progression. The researchers then compared their artificial neural network with a logistic regression model from the same data.

Both methods predicted surgery at a high degree, but the artificial neural network was more accurate, according to the report.

The paper won the Sanford J. Larson, MD, PhD Award.

More articles on spine surgery:
Dr. Joseph Schwab: Key concepts on AI and machine learning in spine
Florida Spine & Orthopedics sues patient for $49K+ in unpaid medical bills
Walmart can save $30K by flying workers out of state for spine surgery

© Copyright ASC COMMUNICATIONS 2021. Interested in LINKING to or REPRINTING this content? View our policies here.

Featured Webinars

Featured Whitepapers