Charlotte task force to use data, analytics to study cancer hot spots
A task force expects to lean heavily on data analytics to better support care and predict instances of cancer in an area of North Carolina that’s seeing higher-than-normal levels of the disease.
The Lake Norman area near Charlotte is seeing clusters of thyroid and colon cancer at rates not generally seen in the region; it’s also seeing extreme rates of ocular melanoma, a cancer so rare that nationally there are only five cases per one million people—but in this metro area of 56,000 residents, there are currently more than 20 cases.
To research the alarming number of cancer cases, the local chamber of commerce created the task force, which will be led by Jeff Tarte, an ex-state senator and former CIO at New England Medical Center, Tufts University and Johns Hopkins Health System.
Scott Guilfoyle, former chief technology officer at PayPal, has been chosen by a coalition of providers, employers, technology firms and local and state officials to help collect and analyze data.
Guilfoyle and Tarte will be aided by the Atruim Health and Novant Health delivery systems, as well as physician groups, population health and genetics experts, machine learning and artificial intelligence professionals, data scientists, community leaders and major employers.
Data analytics will be the key, says Guilfoyle, and his business includes a deep learning analytics platform that can look back at data up to 10 years old.
“Collecting and processing as much health data from as many sources over a prolonged period of time is crucial to fully researching potential cancer hotspots,” he says.
“We don’t extrapolate results from sample data sets as many analytics companies do, because we can ingest data sets at a high rate,” Guilfoyle explains. “We can take in every bit of data available over any time horizon. Local individual healthcare providers have millions of data points, and along with data from the local and state level, we have potentially billions of data points with which to work.”
Claims data also is playing a role in the quest to understand what is going on in the region. More than 58 million claims, primarily from the ANSI X12 EDI 835 and 837 transaction standards for healthcare are being analyzed.
When additional providers join the fight and contribute their data, the effort gets a whole new set of intelligence that goes in the database, including contract data, Guilfoyle says.
The platform memorizes what parts of the work are changing because new data is always coming in, and the platform learns more, so the degree of manual work goes down.
That said, the organizations working on the cancer hot spots face a familiar problem. Despite the accelerated cancer cases in the region, some organizations still aren’t fully on board to contribute data, Guilfoyle notes.
“Our challenge is needing to have a more open approach to share data. We need to overcome the fear of loss of control of patient data. If expanded breast, colon and lung cancer types are caught early in Stage 1, we can have an 85 percent survival rate. That includes identifying younger women and making sure they get a mammogram screening.”
Anyone being screened for a mammogram or any other test also is getting their data back so they can share it with physicians.
Another available resource being employed is the Government Data Analytics Center (GDAC), which works with agencies to define their analytic needs, determine access to required data sources and measure the level of effort to support analytics development. As more organizations request the assistance of GDAC, prioritization of these needs will be based on the level of effort and cost, return on investment and available resources.
Further, the Gillings School of Global Public Health at Chapel Hill has a program to encourage colon screening in low-income communities to get people screened.