642
document and interpret the results, how to apply the information for treatment deci-
sions and prevention screening, and when to refer patients for genetic counseling.
Such automated guidance is vital for both health care workers and their patients.
EHRs also facilitate research with large cohorts, a factor that is especially valuable
for prospective studies of genetic and environmental effects on health in which the
linking of phenotype to genotype is essential. Translational research is converting
new knowledge gained from EHRs into insights about prevention and treatment of
diseases. The challenge is to ensure that innovation in research and medicine is
equaled by policies that foster science while protecting and respecting research par-
ticipants and patients (Hudson 2011 ).
Use of EHRs for Personalized Drug Discovery and Development
Although the application of EHRs is benefi cial for patient care in a hospital or clini-
cal setting, it can also aid drug discovery. EHR databases are already being used in
the pharmaceutical industry for market research, pharmacovigilance, clinical bio-
marker validation and drug safety evaluation. Because EHRs provide observational
data for a large population over long periods of time, it is possible to utilize them for
a better understanding of how drugs affect patients through changes in diagnosis,
disease progression and laboratory measurements. Specifi c applications of EHRs in
a drug discovery include fi nding novel relationships between diseases, re-evaluat-
ing drug usage and discovering phenotype-genotype associations (Yao et al. 2011 ).
In the near future EHR systems and related databases will have a signifi cant impact
on how we discover and develop safe and effi cacious medicines.
Personalized Prognosis of Disease
Computational and Applied Genomics Program of the Duke University (Durham,
NC) has developed a comprehensive modeling approach to combining genomic and
clinical data for personalized prediction in disease outcome studies. This integrated
clinicogenomic modeling framework is based on statistical classifi cation tree mod-
els that evaluate the contributions of multiple forms of data, both clinical and
genomic, to defi ne interactions of multiple risk factors that associate with the clini-
cal outcome and derive predictions customized to the individual patient level. Gene
expression data from DNA microarrays is represented by multiple, summary mea-
sures termed metagenes; each metagene characterizes the dominant common
expression pattern within a cluster of genes. A case study of primary breast cancer
recurrence demonstrates that models using multiple metagenes, combined with tra-
ditional clinical risk factors, improve prediction accuracy at the individual patient
level, delivering predictions more accurate than those made by using a single
genomic predictor or clinical data alone. The analysis also highlights issues of com-
municating uncertainty in prediction and identifi es combinations of clinical and
genomic risk factors playing predictive roles. Implicated metagenes identify gene
20 Development of Personalized Medicine