The ethics of using predictive analytics in medical decision making. A mobile app that captures clinical data from Parkinson’s patients. Using predictive analytics to help inform end of life decisions. These are just three of the topics discussed at the Big Data & Healthcare Analytics Forum in Boston on November 20-21, 2014. Here’s what you missed:
“Everything I do is predictive analytics,” said John Showalter, MD, chief health information officer at University of Mississippi Medical Center. He has no ethical concerns about using predictive analytics in healthcare.
“My entire job is predictive analytics. It’s about looking at you and determining what is the most likely diagnosis and what is the most appropriate test to confirm that diagnosis,” he said. “Somehow, when it gets so big that the computer has to do it, it’s scary. I don’t understand that mindset at all.”
We get into ethical concerns when we automatically act on those predictions, said Showalter. “Every once in a while someone comes in with the sniffles and those sniffles are the precursor to some awful diagnosis that we miss.”
Suchi Saria, assistant professor in the department of health policy and management at Johns Hopkins University, told the crowd about a pilot project with Parkinson’s patients at Johns Hopkins that uses a mobile app to capture information about the frequency of patients’ tremors, gait quality and voice changes.
Researchers are trying to determine if they can use this data to quantify day-to-day changes in severity over time. The goal would be to trigger communication with a neurologist or a primary care physician to adjust medication dosage if a patient’s health has deteriorated, said Saria.
In Jeanne Huddleston’s hospital’s mortality review process, they call it the “freight train,” she said. Dr. Huddleston is the medical director of the healthcare systems engineering program at the Mayo Clinic. She’s referring to “the care that keeps going and going and going. It feels like everyone’s just loading the train up. It’s got three engines on it, and no one just stops and steps back and looks” at how they’re caring for patients at the end of their lives.
Conversations about end of life issues will always be challenging. Access to information generated by big data about the probability of success with certain procedures can help inform these conversations, though, according to Huddleston. Of course, trained clinicians must continue to rely on their medical training in navigating these sensitive issues with patients and their families.
“It’s important to understand what patients want,” she said. “Getting the research done so we can populate that data visualization and show them the results of that research is vital, and it contributes to shared decision making between the clinical team, the patient and their family.”
Aine (“ONya”) Cryts is an on-staff contributing writer for MedTech Boston. She's a political scientist by education, a writer and marketer by trade. She has written for various healthcare technology publications and also served as marketing director at several healthcare software companies in the Boston area. Cryts is an avid volunteer, pet lover and long-distance runner. Story ideas are always welcome.
Send this to a friend