Information Technology Reference
In-Depth Information
targeting, and to preserve as well as improve quality. In some sense, the direct appli-
cation of Big Data analytics for healthcare is late to the game and is making a transi-
tion that several other industries—such as airlines, weather prediction, and actuarial
sciences—made several years ago.
7.3
Discussion: Hot Topics and Future Directions
As the use of data in health care grows, there are several sources of data that need to
be synthesized and integrated. The key sources to watch are:
￿
Clinical: Electronic Health Records
￿
Genomic data
￿
Research and clinical trial data
￿
Patient reported data (via personal health records, surveys or mobile apps)
￿
Social media data (Twitter, Facebook, Google Plus)
￿
Billing, claims and fi nancial data
￿
Sensors collecting data on people and their environment
￿
Public health surveillance and health care utilization data (AHRQ; DHHS)
Approaches that go across data-sources and that attempt predictive data-
mining—such as predicting falls before they happen (via carpet sensors), readmis-
sions [ 19 ], suicides [ 21 ], depression [ 22 ], abuse [ 23 ] —are fruitful directions to
apply Big Data analyses.
Along with the increase in data availability, there is increased participation, own-
ership and stewardship in using data by the “engaged patient”. Instead of being a
by-stander in one's own care, patients are empowering themselves with raw data
(e.g., Quantifi ed self) and other individual's experiences (e.g., www.patientslikeme.
com ) as well as the wisdom of other patients (e.g., on forums such as www.smart-
patients.com ) .
The idea of using user generated content for enhancing health is exemplifi ed by
the Quantifi ed Self collaborative, which lists over 500 modalities of collecting raw
data from an individual for self-tracking. 9 Given the popularity of such efforts, the
application of Big Data analysis for mass phenotyping to discover patterns and cor-
relate those patterns with health and well-being, is bound to increase [ 12 ].
Finally, it is important to realize that just because vast amounts of data are avail-
able we are not guaranteed to fi nd better insights. The results based on Big Data will
only be as good as the analysis methods employed, and there is therefore an urgent
need for new formal science methods as advocated by the Data Science movement 10
to help with Big Data interpretation.
9 http://quantifi edself.com/about/ .
10 http://bigdatablog.emc.com/2012/11/09/openchorus-project-the-dawn-of-the-data-science-
movement/ .
Search WWH ::




Custom Search