Ahead of the curve, Orion Health looks to introduce machine learning

Published on the 29/09/2016 | Written by Donovan Jackson


Machine learning just starting Hype Cycle journey…

With plenty of talk around machine learning, NZX-listed healthcare software vendor Orion Health is looking to turn that into action within the next 12 months. In its quest to do so, it has recently hired a new lead for its analytics team and said it is ‘investing significantly’ in analytics and machine learning applications. The application of the technology in healthcare is not unprecedented, but it is in its very early days; analyst IDC anticipates that 30 percent of healthcare providers will use cognitive analytics with patient data by 2018.

iStart asked Orion’s VP of integration Dhaya Sivakumar what machine learning can do in the healthcare environment. “There is a push globally around population health management; when you do that, you’re building cohorts, assessing high level risk with the right care plans for specific individuals. Risk stratification and assessment can be done using machine learning, and that will happen in the background. It will do scoring on data presented and create groups of patients to target with more appropriate interventions,” he said.

The company’s new head of analytics, Peter McCallum, said machine learning advances the goal of precision medicine. “Traditionally, healthcare has worked with a ‘broad brush’ approach. There is so much data from so many sources involved in treating patients, but clinicians don’t have access to all that information. They are working on care plans with limited information.”

Through machine learning – and alluding to recent comments by SAS Institute’s Oliver Schabenberger (big data – intelligently ignore most of it) – McCallum said clinicians can be rapidly presented with contextual information which can make them more efficient, while also improving the delivery of quality care.

In simpler terms, this means analysing data from thousands of other patients with similar afflictions, then providing suggested courses of action to improve the management of that person’s medical issues. It also means the 15 minutes the doctor spends with you now, could become 7 minutes or even less, enabling the limited number of doctors to handle more patients – and better, at that. (As an aside, in our experience, patients visiting doctors with Google diagnoses are rarely appreciated; could Orion’s machine learning be a predictive ‘Google for doctors’?)

How does that differ from traditional BI, we wanted to know? “Traditional BI is a rear view mirror; you’re reporting on what has already happened. Machine learning and predictive analytics gives you an idea of what is likely to happen. It is a scoring process; there is a big distinction between the two,” Sivakumar explained.

With ‘hypy’ concepts like machine learning, it is always a good idea to reference Gartner’s Hype Cycle to get an idea of how likely production applications of the technology are. The analyst has been kind enough to release one just weeks ago, and on it, machine learning finds itself just past the ‘Trigger of Innovation’. That makes Orion Health’s efforts at putting it to work ambitious.

But putting it to work is exactly what the company is committed to doing. Sivakumar said the product roadmap is established and machine learning should be included in Orion’s products, which include the Amadeus open data platform, within the next 12 months. Asked how mature machine learning is right now, he added, “There’s been a huge amount of investment in the foundations; we’ve done the research and are now getting the talent on board. We’re busy with lots of prototyping, and are getting to the space where we are ready to deliver product. Now is the time get to market.”

New features in a product are one thing, getting users to adopt them another. McCallum said there is no easy answer to that, but the anticipated benefits of improved care and reduced time to decision-making for patients provide strong drivers. “If you can deliver better care in a shorter time frame without guessing, that’s probably a strong motivation.”

Post a comment or question...

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

MORE NEWS:

Processing...
Thank you! Your subscription has been confirmed. You'll hear from us soon.
Follow iStart to keep up to date with the latest news and views...
ErrorHere