Artificial Intelligence and Big Data in Orthopedics
In This Video
- Working within the Spine Service at Massachusetts General Hospital, Christopher Bono, MD, is researching how artificial intelligence (AI) can be utilized to improve and streamline surgical decision making
- Using vast quantities of data to develop an algorithm, doctors are able to input individual patient-specific data and predict patients' potential surgical outcomes and risks
- Another application of AI is identifying underreported conditions and diagnoses using natural language processing
- Dr. Bono suggests that the development of big data algorithms will transform clinical practice, since it will inform decision-making processes in orthopedic surgery that previously relied on individual surgeon knowledge and expertise
In this video, Christopher Bono, MD, orthopaedic spine surgeon, executive vice chair of the Department of Orthopaedic Surgery at Massachusetts General Hospital and associate program director of the Harvard Combined Orthopaedic Residency, describes his latest research into artificial intelligence decision making for spinal surgery.
Subscribe to the latest updates from Orthopaedics Advances in Motion
Transcript
My latest spinal research within our research team involves a lot of decision-making protocols being developed using artificial intelligence (AI). We are currently working on technologies and ways to use artificial intelligence in order to predict many different outcomes, teaching the artificial intelligence programs to recognize fracture patterns on X-rays, to recognize when there is a tumor or not a tumor in an X-ray or an MRI, as well as various other applications. This research will advance the field of spinal surgery, as it will enable practitioners to make decisions based on individualized statistics.
What is interesting about artificial intelligence, as well as the use of big data, is that huge amounts of data can be inputted into these algorithms. With the research and the projects that we're working on now, we will be able to input individual patient-specific data and have it calculate a customized prediction of the patient's outcome, their infection or complication risk, and ultimately, even guide individual surgical decisions to be made.
One of the more simple examples is called natural language processing. For instance, we did a study looking at the presence of a dural tear or a spinal fluid leak in surgical reports for patients who had undergone lumbar surgery. The actual reporting that was found through coding was not very reliable. The computer program actually learned the keywords and demonstrated that the dural tears were grossly underreported. This was an entirely new way of reporting complications in the field of medicine, and particularly within orthopedics. We are working through our artificial intelligence center on methods to implement this as a quality and safety standard in addition to a research platform. Artificial intelligence will change decision-making.
Now this is a point of contention within the field. There is going to be a struggle between the individual physician who wants to make decisions based on their information, maybe the evidence, maybe on how they were trained 20 or 30 years ago. This will be pit against the algorithms and the recommendations based on AI that will be updated and fine-tuned based on everyone's data who is collecting and contributing to these databases. So it may be a decision that is suggested that is very different than we would have selected on our own. It may be different than the patient's preference or values and this has to be considered when we are engaging in shared decision-making. But certainly, we expect artificial intelligence to give us a very objective and scientific analysis of a patient's situation and what treatment, or no treatment, might be recommended.
Learn more about the Department of Orthopaedic Surgery's Spine Service
Learn more about the Harvard Combined Orthopaedic Research Program