Where healthcare environments and digital futures converge

Health 2.0

Subscribe to Health 2.0: eMailAlertsEmail Alerts newslettersWeekly Newsletters
Get Health 2.0: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn

Health 2.0 Authors: Jason Bloomberg, Robin Woods, Progress Blog, OnPage Blog, Mat Rider

Related Topics: Health 2.0, Healthcare Innovation Journal, Big Data on Ulitzer

Blog Post

Attacking High Blood Pressure through Big Data

A shift from hypothetical models to data-mining approaches has begun and technology will only make data collection easier

Hypertension, or the state of having constantly high blood pressures within the arteries, affects almost 76 million people in the US. That is nearly 34% of the entire country's adult population. Needless to say, it isn't rare. Adding to that, it is the leading cause of cardiovascular death in the US.

Treatment of the disease is achieved by using drugs that slow the heart rate, dilate the arteries, or dilate the veins. There are actually over ten classes of these antihypertensive medications being used today. With so many available treatments, it would be assumed that the disease is under control. That is not the case, however.

Many of those who have been diagnosed with hypertension are not well-controlled with their current medical regimens, either due to poor selection by the clinician or by non-compliance. In fact, only 48% of patients who are aware that they suffer from the disease are being adequately treated by medications.

Attempting to Predict the Disease
An article in the Journal of Hypertension notes that until recently, much of what we did in healthcare was derived from a hypothesis-driven model. This provided clinicians with narrow sets of modifiable (lifestyle, diet, etc.) and non-modifiable (age, ethnicity, etc.) risk factors for hypertension. This has predictably lead physicians to attempt to intervene based on these risk factors, which has had limited success.

That is beginning to change. With the utilization of modern technology, we are transitioning to an "atheoretical, data-mining" approach. Initially, sequencing the human genome sparked our venture into using massive amounts of data to solve medical issues, which lead to the birth of the "big data" movement in healthcare. In a recent study, also published in the Journal of Hypertension, scientists utilized data-mining to build a predictive model for incident hypertension. The model was moderately better than the Framingham risk score, the currently established risk assessment for hypertension. The findings show great potential for the enhancement of screening processes used in primary care.

Improving Care with Big Data
The most notable example of big data improving the control of hypertension was found at the Kaiser Permanente North California (KPNC) hypertension program. The program enrolled 652,763 patients from 2001 to 2009. During that time, clinicians developed and shared performance metrics, used evidence-based guidelines, deployed medical assistants to patient homes for blood pressure measurements, and prescribed single-pill combination therapies. Adding to that, the guidelines used for treatment were updated every two years based on the most recent randomized controlled trials and national guidelines. These were regularly distributed via printed documents, email, and other media to all of the physicians in the program.

In essence, the entire program was functioning in a way that encouraged communication, simplicity, convenience, and evidence-based practice via continuous data collection. Predictably, the results were amazing. For the patients enrolled in the program, control rates rose from 43.6% to 80.4% over the 2001 to 2009 time frame. Compare that to the national average which increased from 55.4% to 64.1% in that same time. This heavily suggests that excellent communication and constant evaluation of treatment strategies will drastically advance the healthcare system.

The healthcare industry is still in the early stages of big data utilization. As a current example, EMR systems are taking advantage of big data collection, which could help individual practices provide better care to their patients. Nonetheless, a shift from hypothetical models to data-mining approaches has begun and technology will only make data collection easier. Furthermore, just as in any other industry, consistent measurement of effectiveness (of treatments or interventions) must be performed in order to improve outcomes.


More Stories By William Rusnak

William Rusnak is a fourth year student at Drexel University College of Medicine, financial investor, writer, and entrepreneur. He writes about topics such as healthcare technology, biotechnology, and nutrition. He is currently applying to residencies with plans to practice in Primary Care and Sports Medicine. Outside of his professional life, he is a family man, performing musician, and paleo-diet enthusiast.