Real-time big data analytics for clinical care

August 1, 2014 in Medical Technology

Over the summer, I’ve given many lectures about SMAC — social media, mobile, analytics and cloud computing.

The most popular analytics topics are business intelligence, big data, and novel data visualizations.

Recently, Dr. Chris Longhurst, chief medical information officer at Lucile Packard Children’s Hospital, and colleagues wrote an article in the Big Data Issue of Health Affairs, that suggests a very practical approach for enabling real time analytics within an EHR. They call it the Green Button.

The Blue Button is for patient view/download/transmit of medical records.

The Green Button is for instant access to outcomes, cost, and risk information for patients that match a given profile.

Here’s a personal example.

When my wife was diagnosed with Stage IIIA Breast Cancer in December of 2011, the biomarkers of the tumor were HER2 -, Estrogen +, Progesterone +.    Imagine that while in her record, a Green Button enabled access to the de-identified records of all 50 year old, Asian females with similar tumors and showed the treatment protocol used, the side effects, the cure rate, the cost, and the complications.

Although not completely scientific, such an approach does not identify causality, it does demonstrate experience and standard practices in the community. The Green Button idea is foundational to the learning healthcare system we’re all trying to build.

We do need to be careful. Here’s one example from our work with I2B2/Shrine.

Did you know that the average human white blood cell count is 5 at noon but 13 at 3 AM?

Scientifically we know that white count does not vary with circadian rhythm. However, who has a white blood cell count drawn at 3 AM?  Sick people.

You cannot conclude that white count varies over the course of a day because the data has confounding complexities.

However, there is an interesting possible conclusion. People who get white counts drawn at 3 AM, get blood cultures at 4am, and antibiotics at 5 AM. We can suggest that if you order a white count at 3 AM then you want to order a blood culture and antibiotics at the same time, since you’ll end up doing it anyway.

The Green button idea is to present valuable historical observational data at the point of care.

I2B2 is a great tool for clinical research and clinical trial enrollment, but imperfect for point of care advice.

How might the Green button be developed?

Emerging companies like QPID are creating new tools that summarize structured and unstructured data into unique visualizations.

The BIDMC experience with care management using a third party registry populated via the state HIE also provides promise.

I look forward to experimenting with the Green Button concept — another item on my to do list for the next year.

Be the first to like.
VN:F [1.9.22_1171]
Rating: 0.0/5 (0 votes cast)

Article source: http://www.healthcareitnews.com/blog/real-time-big-data-analytics-clinical-care

Be Sociable, Share!
Bookmark and Share

Leave a reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>