It is well known that big data will change healthcare as we know it. But, how will the next decade transform patient care? Dr. Pratik Shah develops new products that exist in the intersection of healthcare and machine learning by creating technology algorithms that run off of structured and unstructured data from a variety of sources. This kind of artificial intelligence advancement has the potential to revolutionize the healthcare industry. What can we learn from his success?
Dr. Shah stresses that the artificial intelligence architecture is critical first to strategically understand to eliminate red tape and potential barriers. He provides the example of both liver cancer and oral cancer patients that could benefit from deep machine learning approaches. To make better diagnoses he needed to reduce the number of images needed for machine learning and the expensive technologies so often required to screen patients.
Traditional methods would suggest 10,000+ images would be needed to provide a baseline of the disease, but Dr. Shah had a different vision. The solution involved starting with just one image of a patient that had the disease, and from that, he found a way to extract billions of data points just from that one image. Also, to reduce the burden of expensive machine screening, instead of using an expensive CT scanner for the undiagnosed patient, he started with just a photo from the undiagnosed patient’s mobile phone. He overlaid the pictures of the diseased mouth with the undiagnosed mouth to make a composite image. The results were that he only needed fifty of the original diseased images to compare with the one picture from the undiagnosed patient’s phone to get a higher diagnosis efficiency than currently exists today.
The forces at play include technological advancements like iPhones and new algorithms for interpreting data. Forces playing against the innovation include the economics of the health care system that has invested in these expensive machines and need volume to sustain their business models. Also, health plans currently do not know how to reimburse for these innovative kinds of services. There are some legal barriers as well, as the government is still trying to catch up in the mHealth regulation which could affect the success of these kinds of technologies.
However, generally speaking, the customer defines value. And, if the customer can “double check” their healthcare at a relatively low cost, there may be a market for these type of services as long as Dr. Shah’s business model is sustainable. Renowned big data scientist Dr. Obermeyer observed, “We are optimistic that patients, who generously – if unknowingly – donate the data underlying algorithms, will ultimately emerge as the biggest winners as machine learning transforms clinical medicine.”
#MHealth #MachineLearning #AI