One of the biggest reasons I wanted to attend the HIMSS Connected Health Conference earlier this month was so I could see what new developments were happening in one of my favorite areas of health IT—mobile health. Sure, the new HIMSS Connected Conference is more than just the mHealth Summit from previous years (it combines that Summit with the just-added Cybersecurity and Population Health Summits, all under the HIMSS umbrella), but the way mobile has grown in the healthcare marketplace in recent years has fascinated me.
At the conference at the Gaylord National Resort in National Harbor, Md., one of the best sessions I covered was on the intersection of wearables and big data, presented by Ken Kubota, director of data science at The Michael J. Fox Foundation (MJFF), and Shahar Cohen, a data scientist and product visionary at Intel Corporation. Last year, the two organizations officially hooked up with the aim to improve research and treatment for Parkinson's disease. The collaboration includes a multiphase research study using a new big data analytics platform that detects patterns in participant data collected from wearable technologies—such as a Pebble smartwatch—used to monitor symptoms, the two organizations announced at the time.
The connection between the disease and leveraging real-time data from Parkinson’s patients originated long before that, however. Kubota began the presentation by telling his story: 13 years ago, while working for Microsoft as a systems engineer in San Francisco, his girlfriend, who was a personal chef, invited him to a July 4th barbeque for a family that she cooked for. Lo and behold, the host of the party was none other than Andy Grove, former president, CEO, and chairman of the board at Intel, and Silicon Valley icon.
Kubota and Grove hit it off that summer day, talking mostly about healthcare, oddly enough. About three days later Kubota said he got a call from Grove, who disclosed to Kubota that he had Parkinson’s disease. “He asked for help. It was personal and vulnerable for me, and it led to the perfect storm in my heart since I lost my father to cancer a decade before that,” Kubota said. “I wanted to help him for this, but didn't know how I could. He told me to learn quickly, and gave me a reading list.”
Parkinson's patients slowly lose the ability to move. According to Kubota’s learnings, there were no objective measures of the disease; it was a real data challenge, he said. “Clinical trials would take too long and cost so much. The samples were small, too. If you want to lose weight, you weigh yourself daily. But you cannot manage what you cannot measure,” Kubota said.
As such, Grove and Kubota worked together to chart different movement measure tests on Grove, the patient, for five years. During this time, Kubota became program director of the Kinetics Foundation, a private Parkinson’s disease philanthropy owned by Grove. Together, they asked the National Institutes of Health (NIH) about sensitive measures for Parkinson’s and took their inputs. Then they approached Eric Dishman, Intel’s digital health director, and they collaborated to make a wrist monitor product to measure movement electronically, looking at reaction time movement. A trial of 50 patients resulted in some trends, and the team was eventually able to detect deterioration in these patients, Kubota said. A major issue arose, however—the wrist device was bulky and heavy, and also too expensive, costing $3,500 per unit.
Wanting to move forward, the Intel advanced analytics team developed a big data platform that was needed to analyze the streams of data. Intel’s Cohen said that the platform has been running since the middle of this year for Parkinson’s patients, transmitting data 24/7 to their mobile devices. On the smartphone, there is an app that receives the data and sends it to the cloud. Real-time analytics can be run on the data via the platform as the patient records the sensory data, Cohen said.
In the project’s first phase, Intel engineers compared the device data to clinical observations and patient diaries to test the devices’ accuracy, and are developing algorithms to measure symptoms and disease progression, according to MichaelJFox.org. Wearable devices can capture up to 300 observations per second, so algorithms or formulas to interpret all that data and report what it means related to someone’s Parkinson’s can help individuals and their physicians monitor disease, the researchers said.
Specifically, Cohen said the Pebble smartwatch that the group chose is the first of its kind that has an accelerometer. From there, the researchers can extract activity levels, which can be shown against medication intake. “We can get accurate medication reports from this, which can help in motivating patients to be more active,” Cohen said. What’s more, tremor, one of the most obvious Parkinsonian symptoms, can be detected through frequency-domain analysis. Additionally, nighttime movement tracking—analysis of movement levels during the night—can indicate sleeping quality to some extent, Cohen explained.
Like with many wearables in the marketplace, there is still a ways to go. There are issues around cost, communication optimization, and data quality, Kubota and Cohen said. But they also said that the Parkinson's study could someday lead to the development of better wearable monitors for patients with Parkinson's and other diseases. Certainly, though, this approach is a significant step up from more traditional Parkinson’s research assessments that involve outdated movement tests and questionnaires. Kubota, who said he still sees Grove weekly, is “optimistic that the data can be gathered by wearables to the point where we will be able to divide and conquer this disease.”