I remember one of my professors in nurse practitioner school emphasizing our duty to provide the “right treatment at the right time.” Unfortunately, this isn’t always possible: Sometimes we don’t have the right medication, and other times we don’t have enough of the right medication. The same concept applies to personnel: Sometimes we don’t have the right type of provider, and other times we don’t have enough providers at all. In the end, the right treatment at the right time seems to rely heavily on the luck of the draw.
Whether a patient receives the right treatment at the right time hinges on the healthcare system’s ability to allocate resources effectively. More than we might like to admit, luck is calling the shots. We aren’t going to eliminate luck, but the healthcare system could minimize luck’s impact by more successfully employing big data.
Historical Use of “Small Data”
We have already been capitalizing on the value of data for decades. Take flu season, for example. Influenza activity begins in October, peaks between December and February, and disappears by May. Hospitals prepare accordingly by stocking extra medications and scheduling additional staff.
Through the use of epidemiological data, the healthcare system can more intelligently manage its resources and ensure that flu patients receive the right treatment at the right time. In this case, data is calling the shots, not luck.
What Is Big Data?
While influenza activity is certainly data, it’s not big data. Big data refers to enormous data sets that are analyzed to reveal patterns, trends, or other associations. It differs from traditional data in its volume, velocity, and variety, which is commonly known as the “3Vs” model. Big data is much larger (volume), is generated much faster (velocity), and includes unstructured data (variety).
Where Does Big Data Come From?
Everywhere. It comes from researchers, patients, devices, insurance companies, and providers. It comes from registries, social media, the stock market, traffic and weather reports, electronic medical records, public health databases, credit card transactions, and fitness bands. It comes from Google and Bing searches, tweets, likes, RSS feeds, YouTube videos, and Facebook pictures. Today, big data accumulates with every digital interaction.
Using machine learning and natural language processing, computers crunch these exabytes of data to glean otherwise impossible insights. We employ these insights to make enlightened decisions. Today, big data is used to spot business trends, influence customer behavior, combat crime, gamble, and scout for athletes — to name just a few. In healthcare, big data can diagnose conditions, read radiographic images, predict infectious disease outbreaks, and thwart bioterrorism attacks.
Eliminating Luck From Heart Failure Treatment
Over the past few years, hospitals have begun to utilize big data to effectively allocate limited resources such as medications and personnel. The concept is similar to what we have been doing for decades with the flu season, except this time we have more data and can make more informed decisions.
For example, a hospital system in Texas is using an algorithm to identify which patients diagnosed with heart failure are most likely to return to the hospital within 30 days of discharge. This algorithm analyzes more data points from more sources faster and more accurately than a human ever could. As a result, the hospital system knows exactly which patients will need intensive follow-up care. Case managers (a limited personnel resource) can then be assigned to the patients who need them most.
In this case, luck doesn’t determine whether a patient receives the best treatment — big data does.
Barriers to Big Data in Healthcare
Big data in healthcare specifically has unique challenges. Our most important data points are spread across disparate systems that do not communicate with each other: electronic medical records, research databases, patient registries, public and private insurance companies, and human resources software. Furthermore, healthcare systems cannot afford to hire data science experts like the banking or high-tech industries. Finally, we haven’t found a good way to keep big data sufficiently private and secure, which is nonnegotiable when it comes to patient information.
Conclusion
Despite these challenges, the future of healthcare data is big data, and it will help ensure that patients receive the right treatment at the right time.
In his 2015 State of the Union address, President Barack Obama reminded me of my nurse practitioner school professor. He announced a new big data initiative that will usher in “a new era of medicine — one that delivers the right treatment at the right time.” Using insights from big data, the healthcare of tomorrow will certainly minimize the impact of luck on the safety, quality, and accuracy of a patient’s care.
