Predictive Analytics: Crunching the Numbers to Deliver Personalized Care
Picture yourself walking into a cardiologist’s office after you’ve been diagnosed with a dangerously high level of “bad” cholesterol. Since your father died at an early age from a heart attack, you know the importance of getting your cardiovascular risk factors under control as soon as possible.
But rather than writing a generic prescription for a statin drug to lower your cholesterol – a common course of action today – your cardiologist now has access to your genetic profile as well as dozens of clinical studies relating to patients with similar backgrounds. That allows her to use a technique called predictive analytics to determine which medication, dosage level or combination of drugs is likely to produce the best outcome for you.
In this case, she would use the health information technology (IT) tools at her desktop to find clinical studies involving high-cholesterol patients with similar genetic factors and family histories.
Predictive analytics is made possible by the growing ability of health researchers to tap into massive databases, extract relevant information and convey the results to clinicians and their patients – a field known as “big data” because of the number-crunching power and sophisticated data analysis tools necessary to find the “nuggets” of information.
It’s one of the many health IT advances being discussed during National Health Information Technology Week (Sept. 16-20), a collaborative forum focused on how technology will transform the delivery of care in the near future. Moving forward with predictive analytics is also a high priority for federal and state healthcare policymakers, as well as the U.S. government, which committed $200 million to data research projects in a variety of fields in 2012 as part of its “Big Data Initiative.”
In the near future, predictive analytics may be able to help physicians, hospitals and other providers in many ways:
• Identifying high-risk patients at an early stage of disease for better long-term outcomes
• Delivering personalized treatments based on a patient’s genetic or metabolic status
• Determining which patients are most likely to be readmitted to a hospital after treatment
• Developing wellness “prescriptions” personalized to specific individuals
• Charting the complex relationships between chronic health care disorders, such as obesity, diabetes and sleep disorders
But there is at least one important step necessary to capitalize on the promise of predictive analytics: integrating patient information now kept in “siloed” databases in order to get a more complete picture. While the electronic health record (EHR) is a remarkable IT tool for physicians to capture and analyze patient information, other clinical information now resides in the databases maintained by pharmacies, laboratories and non-traditional providers, such as acupuncturists. Prior insurance claims may also yield valuable insights into a patient’s medical history.
As a practicing physician, I look forward to the time when these databases can be brought together on behalf of the patient. That will allow clinicians like myself to review all the available evidence and make truly personalized treatment decisions. From my perspective, we need predictive analytics as soon as possible.
Tomorrow: Unlocking the potential of mobile technology