Information Technology Reference
In-Depth Information
benefi t from using Big Data, it is important for stakeholders to appreciate the pace
at which the benefi ts are realized unfolds on vastly different time scales. Research
is inherently exploratory, with false starts and frequent dead ends. Using Big Data
can speed the process along, but it is possible that the basic nature of the activity
(i.e., the high failure rate) does not change. At its core, the practice and delivery of
health care, is an operational problem and Big Data approaches may have their fi rst
impact in optimizing the conveyer belts of healthcare. It is imperative that scientists
do not oversell their research use cases that advance the science and that the enter-
prise of health care delivery pay attention to the use of Big Data to improve the
practice of medicine.
7.5
Conclusion
In summary, Big Data, which is widely used in retail industries as a means to under-
stand consumers' purchasing habits and preferences, is increasingly being used in
health care—both for advancing medical science as well as improving the delivery
of healthcare. In order to achieve the desired goals from the analysis of Big Data, it
is imperative for stakeholders to be clear on the end goals and the sources of data
that need to be integrated. Privacy and data ownership are key issues that should not
be ignored. The axis along which data are voluminous—samples, features, or
time—affects the kind of analyses that are possible and is a key question to ask
along with the goal of the analysis. This is further complicated by the rapid rate at
which data are available (velocity), the heterogeneity of the data (variety), and the
ability to trust the inferences made from the data (veracity).
It is important for all stakeholders to recognize that research is inherently explor-
atory, while the practice and delivery of care is an operational exercise. Big Data
approaches may have their lasting impact in optimizing the delivery of health care.
Finally, data will be king; therefore initiatives such as the “blue button”, which
empower patients in controlling the generation and the access to their healthcare
data, are developments that stake-holders should monitor.
Discussion Points
Review the dimension on which data can be big: Samples, Variables, and Time.
Discuss what kind of data are big along which axes.
Discuss how the applicable analytic methods change as datasets differ along the
dimensions of: number of samples, number of variables, and number of measure-
ments over time.
Big Data analysis can be approached in a problem-fi rst or methods-fi rst manner.
Each approach results in different solutions to the question at hand. Discuss the
trade-offs.
What problems would you solve with Big Data analysis?
Search WWH ::
Custom Search