As the quantities of personal health data go from being measured in gigabytes to terabytes, a precision medicine action plan from the California Office of Planning and Research looks at the state of data sharing in the Golden State – and concludes the it can lead great advances in healthcare if it can get data sharing and standards right.
OPR commissioned the study, which looks at ways to advance precision medicine across both California and the larger U.S.
Health data exists in traditional forms such as electronic health records, lab results and clinical trials, of course – but also now comes in a variety patient-generated flavors from devices such as FitBits and via "passive" collection from smartphone and environmental sensors.
The report finds that there is "tremendous potential to use all of the data as input for algorithms and AI allowing for more real-time understanding and intervention," as long as there is a more concentrated effort across the state and the nation to develop common formats, metadata and ways to share the data.
Additionally, the shifting from siloed data to cloud storage and processing opens up greater possibilities for sharing data across a wide range of stakeholders and researchers. Advances in machine learning can leverage these newly-available troves of data to help clinicians arrive at new insights about individual and population health.
WHY IT MATTERS
Health data is growing in size rapidly and researchers are only uncovering the tip of the iceberg of what they can glean from it.
The report noted that in one experiment "deep learning methods accurately predicted multiple medical events, including in-hospital mortality and 30-day unplanned readmission," and that it is critical to be on the forefront of developing new tools and systems to analyze the information available.
Researchers are finding that the social determinants of health, environmental factors that are not specifically health-related but which directly influence a person or population's health outcomes, are incredibly important in developing precision healthcare approaches for patients.
While many determinants are captured in one way or another, there are no consistent standards for the data. Making SDOH data more readily available to health professionals through a common standard would allow practitioners to develop greater precision health approaches.
THE LARGER TREND
The value of massive datasets and the insights artificial intelligence can bring are already being recognized. Across Europe there is already a nascent framework for sharing anonymized data to allow machine learning to help researchers.
Greater insight driven by data sharing is critical to population health management, an industry that is expected to near $7 billion by 2022. Factors like AI-assisted research and greater collection of social determinant information can greatly help clinicians target both high- and low-risk populations for preventable and predictable health incidents with better data sharing and processing power.
Health data is accumulating at a brisk pace and shows no signs of slowing down. As large states like California begin to suggest common frameworks for sharing and researching this data, they are influencing a national conversation on how best to make patient data available securely, safely, and privately. These efforts are mirrored around the world, and are the harbinger of a new age in precision health.
ON THE RECORD
"We believe that much better integration of various data sources is possible (including the social, economic, and environmental data) and beneficial, and can be activated through state leadership," the authors of the action plan wrote on Dec. 26 in a letter to outgoing California Gov. Jerry Brown.
"We recommend exploring the feasibility of a California Patient Record that gives all Californians the ability to access their complete health record, with the ability to contribute their ability to contribute their own data and share their record with any provider or researcher. We also recommend protections for patients when they do share that data."
In addition, researchers suggested starting with a pilot project that "integrates the various components needed for precision medicine (e.g., technology tools, data integration and sharing, research and clinical partnership, patient-centered care) within a defined population.
"Piloting a precision medicine model of care for a high-needs population would serve the dual purposes of not only understanding precision medicine’s impact on patient outcomes, but also whether precision medicine can be a cost-effective model and reduce health disparities," they explained.
Benjamin Harris is a Maine-based freelance writer and and former new media producer for HIMSS Media.
Twitter: @BenzoHarris.