Personally Controlled Health Records and the App Store for Health

Intelligent Health Lab

Personally Controlled Health Records and the

App Store for Health

Kenneth D. Mandl, MD, MPH

Director, Intelligent Health LaboratoryChildren’s Hospital Informatics Program

Harvard Medical School Center for Biomedical Informatics


$2.5 Trillion17% 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


US Spending per capita vs. Life Expectancy


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


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


Cap applies for any eligible professional with at least $24,000 in Medicare Part B allowable charges in each payment year

Medicare Meaningful Use Incentive Payment Schedule


The Goal:

A Learning Health System


But investment is in current stage technologies:

No data in or out, no communication, terrible UIs


March 1, 2009

“There’s no way small practices can effectively implement electronic health records on their own.”

“This is not the iPhone.”


Later in March


$15M ONC-FUNDED RESEARCH PROJECT


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


Use case 2


Domestic Abuse

British Medical Journal 2009


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)


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


We imagine EMRs as an iPhone-like platform where Medtastic

could create and widely distribute an app across many

disparate EMRs


EHR as an “iPhone-like” Platform

There is a common application programming interface that enables Software developers to build SUSTITUTABLE

applications Push innovation to the edges Nimbly evolve functionality Avoid vendor lock Shrink switching costs Enable disruption


Our vocabulary:

Data Sources (managed by containers)Containers (present data from data sources to

apps in a uniform fashion)Apps (completely substitutable)


Substitutabilityworks both ways—

the containers can also be swapped out


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


“App Store”

The SMArt App Exchange will feature appsapproved by the SMArt committee

Other organizations can operateand vouch for alternate app exchanges

Each SMArt container installationwill decide which apps it wants to featurein its App Gallery

Each user may select his preferred apps placed in his App Dashboard


It is not the wild west


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)


NEJM 2008

Ecosystem


“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


Will disruptive innovation be or fostered in healthcare


Looping in the Patient


In 1994 we observed that institutions rarely share data

H1 H2 H3 x xProprietaryPerceived competitionPrivacyHealth Insurance Portability and Accountability ActNo dedicated resources to do so


What if we gave patients a tool to request their records electronically?

H1 H2 H3 x x

Indivo Server


And create a personal health record

H1 H2 H3 x x

Indivo ServerComprehensive record


The collection of these records is a population health database

H1 H2 H3 x x

Indivo Records Indivo Server


Intelligent Health Lab | Children’s Hospital Informatics Program

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


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


NEJM 2008

Ecosystem


“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


New England Journal of Medicine 2008

Tectonic shifts: PCHR vendors and users create large

accessible populations for public health study and intervention


Individual contributions are accurate

JAMIA 2007


JAMIA 2007


Patient vs. Doc Reports

Basch The Missing Voice of Patients in Drug-Safety Reporting

NEJM 2010


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


THE GENOMICS APP


“What ever will we think about now that the genome project is complete?”


• Genes• Environment

• Microbiome• Phenotype• Healthcare

NEED LARGE NNEED data capture

at home and in clinics


Phenome-Genome Database (PGD)

Standard Biorepository 1. Static phenotype2. No return of

research results to patients

3. No patient engagement


Disintermediation (MD’s not required)


Dangers of Large N and small p(D)


Why consent?

Without consent:Tend to be stuck with anonymized

datasets which are often cross-sectional and single purpose

Impedes study of phenotype over timeTend to lose the opportunity to follow-

up with the patient (public health imperatives are an exception)

Risk privacy backlash


Tectonic shifts in the health information economy:Enabling inference across the data of PCHR users

Population Database


The Gene Partnership

Science 2007


The necessary compact entails complexity

WSJ 2010


www.smartplatforms.org

www.genepartnership.org

www.indivohealth.org

Personally Controlled Health Records and the App Store for Health

Documents

Transcript of Personally Controlled Health Records and the App Store for Health