Personalized Medicine - Translational Informatics: Realizing the Promise of Knowledge-Driven Healthcare - page 41

patients with a virus can be designated as such, unnecessary prescription of antibiot-
ics may be avoided [ 17 ]. One of the most promising areas of application for disease
stratifi cation is in cancer, where molecular information can provide clues regarding
which pathways have been dysregulated, and thus what therapy is likely to work
best, or not at all. Less specifi c, but still potentially quite valuable, is grouping
patients by prognosis. Knowing that someone has a very poor prognosis can help to
inform decisions regarding how aggressive to be with treatments that are known to
have unpleasant side effects.
Even when a distinction is not medically actionable through a specifi c therapeu-
tic decision, biomarkers to differentiate between different groups can be useful. If a
biomarker signature can be used to predict symptomatic fl are-ups in advance, then
even if no treatment is available, a patient can use this information to inform deci-
sions regarding work schedule or recreational plans. Relapsing remitting multiple
sclerosis is a good example of such a condition [ 18 ]. Similarly, if a cohort with a
given diagnosis can be stratifi ed by likelihood of disease progression, this can help
increase the fi nancial feasibility of carrying out a clinical trial on potential leads. As
an example, currently available drugs for osteoarthritis treat only the symptoms and
not the disease [ 19 ]. This is true in part because it can be hard to predict which
patients' disease will progress and whose will remain static, thus requiring very
large numbers of participants in a clinical trial in order to obtain suffi cient statistical
power. In this case, biomarkers for likely disease progression can be used to enrich
the population being tested with likely progressors, enhancing statistical power and
ultimately lowering the high cost of bringing a drug to market.

3.3.4 Beyond Biomarkers

It is important to note that personalized medicine is not solely about physiological
biomarkers. Equally important are intangible factors such as personal preferences
and values. A risk-averse person may choose a low-risk intervention that will only
partially resolve a condition rather than a radical option that could effectively cure
the condition, but comes with a higher risk of mortality. Likewise, a person might
make different decisions, or differently timed decisions, if he or she is starting a new
job, about to get married, or about to become a grandparent for the fi rst time.
Unfortunately, fi nancial resources and insurance coverage factor in as well. Other
considerations might include past prescription compliance, environmental condi-
tions, literacy, and presence of caregivers and a support network.

3.3.5 Timing

One theme throughout the different approaches to personalized medicine, particu-
larly in the P4 version, is that of timeliness. Intervention before a patient is symp-
tomatic, or better yet before a person is sick, is far less expensive than the therapies