Leveraging Electronic Health Records for Phenotyping - Translational Informatics: Realizing the Promise of Knowledge-Driven Healthcare - page 67

In a similar manner, policy makers must make decisions and set priorities based
upon the best available data and knowledge . To-date, such decision making,
when it pertains to clinical data sets, has been limited due to the lack of compre-
hensive and actionable representations of patient or population level phenotypes
in computational tractable formats. Thus, the provision of comprehensive and
rigorous phenotyping methods provides a means of enhancing such data-driven
policymaking.

Providers and Healthcare Organizations

Providing tailored and contextually appropriate decision support at the point-of-
care, such as that which would like clinical and bio-molecular phenotypes in
order to inform disease prevention and/or treatment planning needs, requires
computationally tractable representations of patient phenotype data . The use
of phenotyping methods overcomes the challenges of what is often critical and
unstructured data this is not well aligned with these types of information needs, in
order to enable such evidence-driven and personalized healthcare delivery.
As healthcare organizations seek to achieve the “triple threat” of lower costs,
increased quality, and improved outcomes of care, the ability to characterize
and manage populations of patients requires that we understand the pheno-
types of those individuals and groups . As such, the use of phenotyping algo-
rithms is central to such population management tasks, which are inherently data
analytic in their nature.

Patients and Their Communities

Patients often wish to be integral parts of the care delivery, and even better, wellness
promotion activities that make up healthcare management at the individual or
population levels. By phenotyping patients based upon the contents of their EHR
related information, and enabling the linkage of that data with patient-reported
outcomes, sensor data, and other non-traditional sources, we can enable patients
to become part of a “data fabric” the facilitates such shared healthcare deci-
sion making .
Finally, communities often wish to understand measures that can be taken to pro-
mote health and wellness. By using EHR-derived phenotypes for community-
based and/or participatory research paradigms, we can empower
communities to be part of the evidence-generation process that underlies
knowledge-based approaches to achieving optimal health outcomes.

4.5 Conclusion

For years, researchers have been recommending and attempting the use of EHR data
for research. These efforts have been met with moderate success on opportunistic
projects where the EHR data was complete enough and matched the research goals.
Recently, changes in research towards GWAS in biology and comparative