Intelligent Health Lab
Personally Controlled Health Records
and the
App Store for Health
Kenneth D. Mandl, MD, MPH
Director, Intelligent Health Laboratory
Children’s Hospital Informatics Program
Harvard Medical School Center for Biomedical Informatics
Intelligent Health Lab
 $2.5 Trillion
 17% GDP
 Low return on investment
 24th Life expectancy at birth
 29th Infant mortality
 37th System performance
 1/3 spent on activities that do not improve patient
outcomes
 Inconsistent use of effective interventions
Intelligent Health Lab
US Spending per capita vs. Life Expectancy
Intelligent Health Lab
Biased Evidence—two examples
 Publication bias
 Negative studies aren’t published
 Industry funded trials
 Are less likely published within 2 years of completion
 Are more likely to publish reported favorable outcomes
Annals of Internal Medicine 2010
Intelligent Health Lab
As per the National Coordinator . . .
 New England Journal of Med 2008: Low uptake of
HIT in ambulatory setting
 New England Journal of Med 2009 Low uptake in of
HIT in hospitals
Conclusion: $48B
investment, pushing
the technology
Intelligent Health Lab
Medicare Meaningful Use Incentive Payment Schedule
Cap applies for any eligible professional with at least $24,000 in Medicare Part B allowable charges in each
payment year
Intelligent Health Lab
The Goal:
A Learning Health System
Intelligent Health Lab
But investment is in current stage technologies:
No data in or out, no communication, terrible UIs
Intelligent Health Lab
March 1, 2009
“There’s no way small practices can effectively
implement electronic health records on their
own.”
“This is not the iPhone.”
Intelligent Health Lab
Later in March
Intelligent Health Lab
$15M ONC-FUNDED
RESEARCH PROJECT
Intelligent Health Lab
Clinical use case 1
Med-tastic is a well-funded NewCo which has developed
an elegant medication list application that has
physician and consumer facing functionality
To work, Med-tastic needs
 Prescribing history
 Dispensed medication history
 Allergies
 Problem list diagnoses
Intelligent Health Lab
Use case 2
Intelligent Health Lab
Domestic Abuse
British Medical Journal 2009
Intelligent Health Lab
Use case 2 (cont)
 The application would require
 Comprehensive diagnostic data from primary site of
care for each patient (to work well)
 Comprehensive diagnostic data from all sites of care
(to work very well)
Intelligent Health Lab
 MedTastic may be able to develop apps adapted to
several APIs (Cerner’s Mpages etc)
 Academic group cannot.
 THEREFORE, focus is on an API that enables a single
apps store for
 Cerner Install
 Hospital with homegrown system
 Physician practice
 Open source EMR
Intelligent Health Lab
We imagine EMRs as an iPhone-like
platform where Medtastic could
create and widely distribute an app
across many disparate EMRs
Intelligent Health Lab
EHR as an “iPhone-like” Platform
 There is a common
interface that enables
application
programming
 Software developers to build SUSTITUTABLE
applications
Push innovation to the edges
Nimbly evolve functionality
Avoid vendor lock
Shrink switching costs
 Enable disruption
Intelligent Health Lab
Our vocabulary:
 Data Sources (managed by containers)
 Containers (present data from data sources to
apps in a uniform fashion)
 Apps (completely substitutable)
Intelligent Health Lab
Substitutability
works both ways—
the containers can also be swapped out
Intelligent Health Lab
Governance
 code:
open-source,
open
formats,
led by SMArt team
 app store: one app exchange to start, but
others can be built. Installations manage their
app gallery. Users manage their dashboards.
 brand: compliance test to ensure that “SMArt”
is meaningful
Intelligent Health Lab
“App Store”
 The SMArt App Exchange will feature apps
approved by the SMArt committee
 Other
organizations
can
operate
and vouch for alternate app exchanges
 Each
SMArt
container
installation
will decide which apps it wants to feature
in its App Gallery
 Each user may select his preferred apps
placed in his App Dashboard
Intelligent Health Lab
It is not
the wild
west
Intelligent Health Lab
Intelligent Health Lab
SMArtPlatforms.org
 SMArt Health App $5,000 Challenge
 Announced by Aneesh Chopra during keynote with Bill
Gates at mHealth last week
 Opens in March and allows innovation in MODULAR
functionality
 Imposes discipline on us to create version 1.0 of the
API
 Judges:
Regina Herzlinger (Harvard Business School)
David Kibbe (AFP)
Doug Solomon (IDEO)
Edward Tufte (Yale)
Jim Walker (Geisenger)
Intelligent Health Lab
Ecosystem
NEJM 2008
Intelligent Health Lab
“We cannot overstate how
important PHRs are to the
efficient functioning of a
low-cost, high quality
health-care system . . . .
We think that the INDIVO
system, or something like
it is a good place to
start.”
--Clayton Christensen
Harvard Business School
2009
Intelligent Health Lab
Will disruptive innovation be or fostered in
healthcare
Intelligent Health Lab
Looping in the Patient
Intelligent Health Lab
In 1994 we observed that institutions rarely
share data
H1
x
H2
x
H3
 Proprietary
 Perceived competition
 Privacy
 Health Insurance Portability and Accountability Act
 No dedicated resources to do so
Intelligent Health Lab
What if we gave patients a tool to request their
records electronically?
H1
x
H2
x
Indivo Server
H3
Intelligent Health Lab
And create a personal health record
H1
x
Comprehensive
record
H2
x
Indivo Server
H3
Intelligent Health Lab
The collection of these records
is a population health database
H1
Indivo Records
x
H2
x
Indivo Server
H3
Intelligent Health Lab
Our original statement
on personal control
 A PCHR stored all of an individual’s medical history in
a container with:
 patient control
 interoperability
 open standards
 rules to protect patients
Intelligent Health Lab | Children’s Hospital Informatics Program
Intelligent Health Lab
Patient role
 Patients can
 access the record
 grant access to others
specific to their role
of selected portions of the record
 store their record in a location of their choice
 annotate in the record (but not delete)
 grant access to “apps” and to devices
Intelligent Health Lab
Ecosystem
NEJM 2008
Intelligent Health Lab
Intelligent Health Lab
Intelligent Health Lab
Intelligent Health Lab
“We cannot overstate how
important PHRs are to the
efficient functioning of a
low-cost, high quality
health-care system . . . .
We think that the INDIVO
system, or something like
it is a good place to
start.”
--Clayton Christensen
Harvard Business School
2009
Intelligent Health Lab
Intelligent Health Lab
Tectonic shifts:
PCHR vendors and users create large accessible
populations for public health study and intervention
New England Journal of Medicine 2008
Intelligent Health Lab
Individual contributions are accurate
JAMIA 2007
Intelligent Health Lab
JAMIA 2007
Intelligent Health Lab
Patient vs. Doc
Reports
Basch The Missing Voice of
Patients in Drug-Safety Reporting
NEJM 2010
Intelligent Health Lab
Individual contributions to drug safety data
 Patient reported outcomes
 Adverse effects
 Efficacy endpoints
 Adherence
 Satisfaction
 Quality of life
 Patient reported data
 Over the counter meds
 Complimentary/alternative meds
Intelligent Health Lab
THE GENOMICS APP
Intelligent Health Lab
Intelligent Health Lab
“What ever will we think about now
that the genome project is complete?”
Intelligent Health Lab
•
•
•
•
Genes
Environment
• Microbiome
Phenotype
Healthcare
NEED LARGE N
NEED data capture
at home and in
clinics
Intelligent Health Lab
Standard Biorepository
Phenome-Genome Database
(PGD)
1. Static phenotype
2. No return of
research results
to patients
3. No patient
engagement
Intelligent Health Lab
Disintermediation (MD’s not required)
Intelligent Health Lab
Dangers of Large N and small p(D)
Intelligent Health Lab
Why consent?
Without consent:
 Tend to be stuck with anonymized datasets
which are often cross-sectional and single
purpose
 Impedes study of phenotype over time
 Tend to lose the opportunity to follow-up
with the patient (public health imperatives
are an exception)
 Risk privacy backlash
Intelligent Health Lab
Tectonic shifts in the health information economy:
Enabling inference across the data of PCHR users
Population Database
poller
PCHR
Intelligent Health Lab
The Gene Partnership
Science 2007
Intelligent Health Lab
The necessary compact entails complexity
WSJ 2010
Intelligent Health Lab
www.smartplatforms.org
www.genepartnership.org
www.indivohealth.org