Big data analytics and clinical decision support is the wave of the future in healthcare and is already taking shape.  The industry has vast amounts of data it generates, and to date, most has been stored as hard copies or in unstructured formats rather than in easily usable digitized formats.  This data includes clinical data, written notes and prescriptions, medical imaging, labs, insurance information, demographic and administrative data, sensor data from vital sign monitoring, and information from social media streams..

Information explosion

Some reports note that in 2011, the US healthcare system’s data reached 150 exabytes.   At this rate of growth, it is estimated big data for US healthcare will reach the zettabyte scale ( 10 ^ 21 gigabytes) and thereafter, the yottabyte (10 ^ 24 gigabytes)! (1)  As a point of reference, it’s interesting to note that Kaiser Permanente, with more than 9 million members, is thought to have anywhere from 26.5 to 44 petabytes of rich data from EHRs, including imaging and annotations (1).  

How can healthcare’s massive, complex and rich data sets be leveraged to create innovative, cutting edge and usable systems? In other words, how can we maximize the use of this data for both clinicians and patients to optimize patient health outcomes, bring the joy back to doctoring, and eliminate tedious tasks to free up time for invaluable personal interaction between doctor and patient?

Is data science the solution?  

Sifting through clunky EMRs is no longer acceptable.  It’s been estimated that big data analytics can save healthcare $300 billion per year in the US. (2)  At Valeet Healthcare, we believe a combination of applying data science and leveraging the human perceptual system is key to discovering associations, patterns, and trends that can best guide us in clinical decision-making.  We also acknowledge that data science alone will not save the day. Humans, who are adept at  recognizing visual patterns, synthesizing information, and applying critical thinking, are an essential part of this equation, provided the data science can present the results in a visual way so that humans can then make better decisions.

I have personal experience where such a system could have helped at least two of my patients. One patient, with documented end stage COPD, had been in and out of healthcare systems over a dozen times in 7 months for COPD exacerbations.  Had there been a unified platform which clearly shared his FEV1 and other clinical data, he would likely not have been admitted to the hospital on so many different occasions, which put his hospital costs over $800,000 in just 7 months.  In this case, having an alarm that automatically notified hospital administration, attending physicians and community partners of his presentation to the ER could have prevented unnecessary hospitalizations.  This patient should have been cared for in the community in a more peaceful and less invasive setting.  In the hospital he received labs and other unnecessary testing which did not change his overall prognosis or even provide immediate short term relief.

Another patient of mine had a penicillin allergy and seizure disorder, was actively septic, and needed broad spectrum antibiotics.  As the clinical provider, I was aware certain classes of antibiotics can actually lower the seizure threshold; however, an alert did not appear in the EMR.   A clinician with whom I was working was steps away from ordering, and I reminded him of the interaction.  Unfortunately, these types of incidents occur frequently in healthcare systems — we need systems in place to avoid these sorts of avoidable errors.

Leveraging the Valeet Healthcare solution

No question, we need tools that combine a patient’s personal health profile with the latest medical guidelines to streamline development of the best treatment plans for each condition a patient has.  Development of the treatment plan is just the beginning.  Patients need support to ensure they are successful and that the treatment plan developed for them is in fact the best one.  Using these tools, we can also create models for risk stratification and predictions scoring.  Who will come back to the hospital and how can we prevent that?  How long will a patient stay in the hospital?   At Valeet Healthcare, we are determined to make sure patients succeed in their health and help clinicians work smarter by leveraging our technology solution.

If you are looking to leverage big data in your health organization please contact us today!  Visit us here or email us at



1.IHTT . Transforming Health Care through Big Data Strategies for leveraging big data in the health care industry. 2013.

2. Manyika J, Chui M, Brown B, Buhin J, Dobbs R, Roxburgh C, Byers AH. Big Data: The Next Frontier for Innovation, Competition, and Productivity. USA: McKinsey Global Institute; 2011.

This blog post was authored by Dr. Armand Prieditis and by Dr. Sima Pendharkar.

Armand Prieditis

Author Armand Prieditis

Armand is a distinguished Data Scientist and an entrepreneur who has been building large-scale data science systems for over 20 years and has successfully founded two machine learning companies, one targeting search recommendations and the other delivering industrial automation solution. In the past he also served as Professor of Computer Science at University of California at Davis. Armand currently holds three patents and has authored 3 books in the area of Artificial Intelligence, Machine Learning and Analogical Reasoning along with numerous publications.

More posts by Armand Prieditis

Join the discussion 2 Comments

Leave a Reply