P4C—From Avalanche to Aha! Successes in Collecting Patient Data for Research

November 15, 2016

We are in the midst of a data explosion. Kristen Schneeman of FasterCures described the current landscape where the patient health record has moved from the file room to the cloud, with 80 percent of office-based physicians using electronic health records. Despite this technological milestone, the results have been unsatisfying. FasterCures fellow Bray Patrick-Lake described her health data as amounting to two binders, six discs, and four patient portals, but still no longitudinal health record. Clearly we are not moving from avalanche to “Aha!” How are these data being used, and how do we identify and harness the possibilities? What is the role of the patient in these efforts? Representing a range of expertise, the panelists strive to support the integration of patient data while maintaining 100 percent of the focus on the patient from data collection to data use.

Getting the data is not the difficult part

Noga Leviner from PicnicHealth kicked off the session by noting, “We can gather 99 percent of the health data for an individual patient, that’s not the hard part.” Leviner explained that the difficult part is making the data useful and meaningful to patients, providers and the system. Still, data collection is no small effort; the approach at PicnicHealth involves fax machines and old-fashioned follow-up calls, emails and faxes—and persistence. Once collected, making meaning of the data involves use of outward-facing programs that display an individual’s data with graphs, as well as a data architecture that can be harnessed for research.

Ray Dorsey of the University of Rochester Medical Center and a White House Champion for Change in Parkinson’s disease spends a lot of time gathering data from patients who are not part of the conversation—the sickest, most distant, and most in need. Dorsey observed, “We the healthy ask the sick to participate in research on our terms.” He would like to see this tradition flipped on its head, so that patients are engaged on their terms. Reiterating previous statements that the data collection tools aren’t the problem, he explained that Apple’s ResearchKit enables Parkinson’s patients to continuously collect disease-relevant data themselves. He stated that the next stage of innovation is to move treatment from the doctor’s office to the patient’s home, which is something we can start to do today. Meaningful data informs understanding of what makes a patient’s disease worse or better.

Short-term fixes are useful when waiting for long-term solutions

When you encounter a problem, fix it in the near term while working toward long-term change. Gina Agiostratidou from The Helmsley Charitable Trust employed this approach when she led programmatic efforts to address Type 1 diabetes, which is developed by only 5 percent of the population with the genetic mutation for Type 1 diabetes. Agiostratidou requested data from the few research teams that had conducted Type 1 diabetes studies, with the goal of pooling the data and asking novel questions about disease etiology to move the field toward stronger prediction and prevention models. She quickly realized that the data were not uniformly collected across studies, making it impossible to draw conclusions from combined datasets. The Helmsley Charitable Trust Type 1 Diabetes program currently focuses on a platform to screen the general population to identify people with genetic risk and then follow them for three years. Ultimately the dataset will consist of varied data from more than 400,000 children and will support a long-term solution that makes the data available to the community, creates incentives for participation and data infrastructure, and develops governance for data storage, access and analysis. Using data for precision medicine and population health is the future goal.

Data is messy, so let’s get our hands dirty

Joel Dudley of the Icahn School of Medicine, Mount Sinai encouraged the audience to embrace  complexity, adding that in health there are rarely smoking guns to solve all our problems. It is difficult to move from knowledge to understanding, even in the best-case scenario when there is one disease and one gene, such as cystic fibrosis, which has no cure. Humans are complex adaptive systems; there is a constellation of billions of interactions inside a single cell. Dudley envisions a world where we use complex data to help us know the unknown. To get there, everyone, especially patients, needs to dirty their hands with the data. He drew a parallel to the PC revolution, which moved computers from untouchable servers to personal computers. He forecasts a future where the next generation of data can show everyone their own health journey. 

Sally Okun appreciates both the data avalanche, in her role in PatientsLikeMe, and the individual patient, as a former palliative care nurse. To avoid being knocked down by the avalanche, we need to find ways to systematically and methodically collect, understand, and use data to change care delivery. Okun described how PatientsLikeMe—with more than 30 million structured data points, 3 million pages of qualitative data, and 90 publications across more than 450,000 members and 2,500 conditions—places the patients at the center. She highlighted the organization’s patient community, who want information that allows them to get better, live better, and know they are not alone. Patients want to share their health data, because they want their experience to count. The opportunity to experience an aha moment is grounded in the premise that patients are the only ones who can validate the data that are collected about them.

The future is now

In many ways, the future is now with technology no longer the barrier and people both the barrier and the solution. The future depends on bringing more people to the table (#healthcitizenship), adding even more data to the system, and remaining focused on the patient at the center.

 Watch the Discussion