Mining the Bibliome - Translational Informatics: Realizing the Promise of Knowledge-Driven Healthcare - page 89

5.5 Conclusion: Mining the Bibliome of the Future

There is no end in sight with technological improvements that enable an increase in
available data at an exponential rate that far exceeds human ability for interpreta-
tion, understanding, or meaningful use. Within health care, biomedical literature is
an essential mechanism for transfer of knowledge, which is the result of transforming
raw data into meaningful information, as wisdom that can guide research inquiry,
clinical practice, and community understanding. Biomedical literature will increas-
ingly become the essential bridge between the volumes of data and their interpreta-
tions. As the medium for literature increasingly shifts from analog (paper) to digital
formats, the ability to leverage bibliome mining approaches will correspondingly
improve. Similarly, the ultimate utility of biobliome mining techniques for identify-
ing actionable knowledge will depend on appropriate evaluation and understanding
of limitations of inferred knowledge. There will undoubtedly be the need for contin-
ued improvements in bibliome mining techniques, which collectively will need to
shift from research-only environments to contexts for support clinical care decisions
(either by providers or patients). Nonetheless, the prospect of leveraging computa-
tional approaches to extract and identify potentially novel knowledge from bio-
medical literature offers some hope in the harnessing the value of next generation
biomedical data.
The process of knowledge discovery and cataloguing of wisdom is inherent in
the overall decision support lifecycle. A grand challenge remains the generalization
of this decision support lifecycle in the context of a real-world health care system
with targeted challenges (e.g . , identifying ways to reduce costs while keeping the
quality of care or clinical outcomes the same). The most signifi cant role of the bib-
liome, therefore, is to be the catalogue of wisdom onto which new knowledge dis-
covery approaches are used. Furthermore, the process of bibliome mining will help
unveil new potential hypotheses that can be demonstrated or tested within the con-
text of a real-world healthcare system.
As advances in technologies like natural language understanding continue, along
with improvements in literature and metadata representation, the potential role of
the bibliome will become an essential element central to any functioning and self-
assessing healthcare system. This is not entirely different from the ad hoc manner
in which clinicians may base clinical decisions or identify potential treatment regi-
mens - by scanning literature such as systematically catalogued in resources like
MEDLINE, a clinician can identify case studies that describe a particular patient
population and possible treatment options along with expected outcomes. By lever-
aging computational methods to further curate biomedical literature, it is conceiv-
able that new knowledge may be identifi ed that might have otherwise been
overlooked. This is because in contrast to a clinician seeking a particular answer for
a particular scenario, bibliome mining presents the full array of potential knowledge
to consider for a variety of scenarios. In this way, the application of bibliome mining
in the context of a learning healthcare system may identify potential hypotheses that
advance the overall system in a way that may not have been otherwise considered.