Database Reference
In-Depth Information
1. Introduction
Healthcare includes ”efforts made to maintain or restore health, es-
pecially by trained and licensed professionals” [1]. These efforts are
performed by various entities within a large ecosystem composed of pa-
tients, physicians, payers, health providers, pharmaceutical companies
and more recently, IT companies. Medical informatics [2] is the sci-
ence that deals with health information, its structure, acquisition, and
use. A fundamental goal [3] of medical informatics is the improvement
of healthcare by acquiring and transmitting knowledge for applications
in a broad range of settings, across computational platforms, and in a
timely fashion.
Reaching this goal can have far-reaching consequences on our society.
Historically, healthcare has been provided in a reactive manner that lim-
its its usefulness. A major issue with this is the inability to detect early
or predict that a patient may be prone to develop complications associ-
ated with chronic diseases like cancer or diabetes. Even in an intensive
care environment, care is often provided in response to the adverse events
typically detected after the emergence of clinical symptoms, or after the
interpretation of a lab test. Quite often, reacting after the detection of
such events reduces the ability of physicians to drive patient trajecto-
ries towards good outcomes. As a result, there is an increasing push to
transform medical care delivery from reactive to proactive.
This transformation necessitates better monitoring and understanding
of the patients. Medical institutions and healthcare providers are collect-
ing large amounts of data on their patients, and organizing this data into
Electronic Medical Records (EMR) and Patient Health Records (PHR).
With the advances in sensor technologies, several new data sources, pro-
viding insights on patients, are emerging. For instance, Bluetooth en-
abled scales, blood pressure cuffs, heart rate monitors or even portable
electrocardiogram monitors are now available off the shelves for the col-
lection of important vitals that can be interpreted for early diagnosis.
Using these advances in sensor technologies, several remote health mon-
itoring solutions for chronic disease management, and wellness manage-
ment have been proposed [4].
While this rapid growth in healthcare sensor data offers significant
promise to impact care delivery, it also introduces a data overload prob-
lem, for both systems and stakeholders that need to consume this data.
It is, therefore necessary to complement such sensing capabilities with
data mining and analytical capabilities to transform the large volumes
of collected data into meaningful intelligence. In this chapter, we sur-
vey the application of sensor data mining technologies in medical in-
