Mobile Health Technology for Personalized Primary
Care Medicine
Current primary care delivery models often revolve around a
series of episodes, rather than functioning as a continuum.
Patients make serial visits to a clinic, where clinicians
collect discrete and isolated health data. These single data
points, collected where patients spend little “life” time, are
compared with the patient’s history and analyzed to make
presumptive diagnoses and care recommendations. This
model neglects significant amounts of potentially meaningful
data from patients’ daily lives and results in lessinformed
treatment and scheduling of follow-up visits.
Lack of meaningful data further blinds clinicians to patients’
health outside of the clinic and can contribute to unnecessary
emergency department visits and hospitalizations.
Personalized care through mobile health technologies
inspires the transition from isolated snapshots based on
serial visits to real time and trended data. By using technologies
from cell phones to wearable sensors, providers
have the ability to monitor patients and families outside of
the traditional office visit. The ability to objectively “see” a
patient’s biological, behavioral, environmental, and social
environment in “real time” can allow for higher level of
analytics, such as predictive modeling, to occur. These data
analytics could provide notifications to primary care providers
of a deterioration in a patients’ health status, which
would allow for more appropriate office visit scheduling.
Scheduling visits in response to real-time data allows individualized
interventions and medication adjustments when
the patient needs it most.
Mobile devices, notably smartphones and other wearable
sensors, can collect this real-time health data from the
patient directly, as well as indirectly from their family or
care takers. Moreover, mobile health technologies overcome
geographic barriers faced by rural patients, allowing
increased provider access.1 These same technologies are
beginning to transcend socioeconomic barriers to care.1 This
new, continuous stream of data has the potential to yield
new insight into disease processes and can enhance our
understanding of the longitudinal effects of care delivery,
medications, and health behaviors.2
Increasing access to mobile technology platforms may be
especially useful for complex chronic illnesses, including
diabetes, obesity, and cardiovascular disease. These are
illnesses for which behavior change becomes a daily, if not
hourly, undertaking that involves multiple overlapping
factors with strong social influence.3 Self-management is
integral to control diseases such as diabetes, for which
patients currently provide 99% of their own care.4 Effective
self-management by patients requires: (1) real-time information
on their health status and behaviors and; (2) ongoing
health professional facilitation of the patient as they monitor
and perform self-care. However, accurate and timely information
for these activities are notably absent from the
current healthcare system.5 Providing real-time data would
facilitate patients and their care providers to better understand
illness dynamics, develop adaptive approaches to
improve health outcomes, and deliver personalized care
when it is most needed.6
NEW MODELS OF CARE DELIVERY FOR MOBILE
HEALTH
The bombardment of health data will require new models of
care delivery with advanced computing capabilities and
analytic tools to filter and present information in a meaningful
way. The full potential of mobile health technologies
will require automation and a care team approach. This will
require 2 levels of monitoring and real-time interaction by
the clinic. The first level will be automated algorithms that
allow a clinic computer to guide the patient to collect correct
data and, within specific parameters, to take health-related
actions. This will circumvent any concerns of the need for
increase in personnel time to analyze and respond to the
data.
The second level of monitoring and real-time interaction
will use personnel in the event that abnormal data points are
not able to be corrected by the automated algorithm. For
example, healthcare team members would be notified if
data, such as blood glucose levels, are consistently out of a
target range or inconsistently reported. The healthcare team
will become involved when a nurse or similar level provider
interprets the data and triages the patient appropriately.