Examine Medical Data and Cure Problems Using Computer Algorithms Without Violating HIPPA
Machine learning is an area of quantitative science that allows computers to run algorithms to create predictions or disentangle causality. We are no longer limited by the ideas that the human brain can formulate because machine learning can uncover unconsidered possibilities. Medical breakthroughs are happening by data mining vast quantities of information and relying on computers to determine correlation. If the medical community partners with business intelligence architects, we can unlock medical knowledge in such a way that lives can be saved and cures can be found.
My company, IntelliTect, is a high-end software architecture and development consulting firm based in Spokane, WA. Recently, we started using machine learning to forecast production needs on a solar farm using weather data and historical energy production reports. Our algorithm allows utilities to more accurately predict when to purchase energy and helps our customers avoid paying last minute inflated rates unnecessarily.
I’ve seen hospitals in Europe share data across borders, and I struggle with the fact that machine learning isn’t prevalent in Spokane’s healthcare industry. I wonder if the medical community is reticent to try a partnership because they think HIPPA compliance isn’t possible. Let me assure you that HIPPA isn’t an issue. Computers can remove personally identifiable data, so programmers only see medical characteristics. If we can combine forces to make data available to be analyzed, we can bring about cures that haven’t yet been discovered.
Machine learning is already solving countless medical quandaries around the world. For example, Sherri Rose, Ph.D., uses statistical approaches to advance human health as an Associate Professor at the Department of Health Care Policy at Harvard Medical School. Rose’s team developed an algorithm to improve the accuracy of a prediction about the stage of lung cancer in a patient. They combined billing and registry data to create a staging algorithm that’s correct 93 percent of the time, an improvement of 75 percent! “We’re trying to solve problems we couldn’t solve before, and this is possible with machine learning and statistical tools,” Rose said. “We now have access to high-quality data, and the advantage is in the combination of these data and our statistical techniques.” She firmly believes that there is lifesaving potential to be found through machine learning.
I agree with Rose. At IntelliTect, our analytical experience started long before machine learning was even a buzz word. My team breaks down complex data to help people make strategic decisions. I am eager to partner with someone in the medical community that has the same goal of solving real-world problems locally. I know that Spokane can do better. We can combine data to correlate and produce knowledge that can save lives. The irony is that we currently have a people collaboration problem preventing this, not a technology problem.
Who’s interested in partnering with me?
If you would like to discuss machine learning further, please email me at firstname.lastname@example.org or join me at the upcoming Machine Learning in Healthcare conference – August 17-18 in Boston, MA.