If you’ve ever watched an episode of House, the medical drama that aired on FOX for eight years, then you’ve watched the framework of differential diagnosis in action. Essentially, the process eliminates the possibility of certain afflictions one by one, until the medical team arrives at only one issue that could be causing all of the patient’s symptoms. The process is simple in theory, but in practice requires the ability to narrow down a patient’s prognosis from thousands of possibilities. As with so many other industries, it’s this complex analysis that can be aided by the help of big data. Technology in the health field now includes a plethora of apps, devices, and other materials meant to improve lives and aid medical professionals. But we’re really just at the tip of the iceberg in terms of what combining medicine and data can accomplish.
In late April, big data guru Bernard Marr wrote a column for Forbes describing the implications of combining huge data sets with the medical expertise of doctors. In short, the possibilities for finding cures faster and prescribing effective treatments more quickly are virtually endless. Doctors are not robots. Even though we’d like to believe they have all of their med school textbooks memorized front to back, there are ways that recalling immense amounts of data in mere seconds can speed the health care process along.
Marr highlights one implication, by noting that the same data people are collecting about themselves through Fitbits and other devices could also be shared with their doctors, even before they get sick. When studied en masse, doctors could begin to spot troubling trends. “Even if there’s nothing wrong with you, access to huge, ever growing databases of information about the state of the health of the general public will allow problems to be spotted before they occur, and remedies – either medicinal or educational – to be prepared in advance,” Marr explained.
One test of that theory is already being put into practice by a partnership between three organizations in western Pennsylvania. It’s called the Pittsburgh Health Data Alliance, and pairs the medical research of the University of Pittsburgh, the computer science expertise of Carnegie Mellon University, and the hordes of deep data collected by the University of Pittsburgh Medical Center network. “Imagine a time when data gathered from routine medical appointments can predict when an individual is headed for the emergency room. Consider a future where doctors use genetic data to develop truly personalized cancer treatment programs,” the alliance explains on its website.
Though the alliance is in its infancy, it also recognizes the opportunity to lower health care costs through quick and expedient care. Largely funded by UPMC, the group also expects to begin two new research centers that will focus on analytics, personalized medicine based on data gathering, security and privacy models in the context of using big data, data-driven education for doctors and the public, and “a new general framework for big data in health care.” Essentially, the alliance hopes to create the mold for what data-driven health care looks like in the future.
In some cases, pieces of that future might already be here. IBM’s genius computer, Watson, has gone on from playing Jeopardy! in publicity stunts to helping doctors narrow down the proper cancer treatments in 14 cancer institutes across the United States and Canada. Oncology is one of the first fields where patients can receive direct care by matching genetic sequencing with specific treatment plans. Normally, it takes doctors weeks to identify the right drug for a cancer-causing genetic mutation. But Watson’s database is now equipped with the findings of scientific journal articles and clinical trial data that describe particular cancers and potential therapies. What takes professionals days to figure out takes the supercomputer minutes.
Because of the immense amounts of data involved, “the solution is going to be Watson or something like it,” oncologist Norman Sharpless of the University of North Carolina Lineberger Cancer Center told Reuters. “Humans alone can’t do it.”
Watson, which operates on the cloud, will start being used for this purpose by late 2015, including at the Cleveland Clinic, the Fred & Pamela Buffet Cancer Center in Omaha, and the Yale Cancer Center. According to Reuters, oncologists will upload the DNA fingerprint of a tumor, which can show which genes are mutating and causing the cancer. Watson will sift through the mutation possibilities and try to determine which mutation is driving the tumor, then suggesting a treatment based on which drug is most likely to stop that particular mutation.
It won’t work for certain types of cancer right away, and chemotherapy and radiation will remain the standard of care for many patients. In some cases, that’s because it is impossible to identify which mutation is driving the tumor, or because a targeted therapy doesn’t exist for a certain mutation. “When institutions do genetic sequencing, only about half the cases come back with something actionable,” said Steve Harvey, vice president of IBM Watson Health. Still, it’s a long way from starting a chemotherapy IV and hoping for the best.
Big data’s role in stopping epidemics
Big data is immensely helpful in other areas of health care, too. In many cases, the key is sharing data that’s already collected between doctors, pharmaceutical companies, and more. Researchers can use personal data to determine the best subjects for clinical trials, and personalized medicine based on genetics is becoming less science fiction and more a reality in doctors’ offices, Marr wrote in his column for Forbes.
What’s more, analyzing people’s movements using data analytics helped doctors to know the best areas to set up clinics to treat the Ebola virus in Africa, and is helping doctors discover the best ways to stem the tide of epidemics.
The key to all of this is the idea that doctors, patients, and other health care entities have to be willing to share information. In most cases, this seems like a no-brainer. If sharing health information starting in childhood will help choose effective medical treatments — or better yet, prevent issues before they arise — there’s very little downside. However, one of the largest concerns as with anything stored on the cloud or shared between networks, the potential for hacking and stolen information becomes much greater. In that case, the onus is on the medical field and those creating the technology to ensure that patient data becomes safe. Otherwise, there’s little point in moving forward.
Follow Nikelle on Twitter @Nikelle_CS