2. The Problem
Mental Health issues
often go undiagnosed,
sometimes leading to;
Suicide (Self Harm)
•As many as 22 Veterans per day
•With 80% of Suicidal Veterans are not in VA
care
•And while The Clay Hunt act provides more VA
oversight & operational reform…not innovative
action
Mass Shootings (Harm of others)
•Often problems are detected
•But the proper response is lacking
•Action could be taken Before problems arise
Insider Threat (Theft of property)
•With National Security implications
•Multi-type: e.g. Germanwings Slide 2
4. • The project was named in honor of Emile
Durkheim,
a founding sociologist whose 1897 publication of
Suicide defined early language analysis for suicide
explanations.
• Our team was comprised of a multidisciplinary
team of artificial intelligence, &
medical experts (Dartmouth & VA)
• Funded by DARPA from mid 2011- early 2013
• Based on mathematics funded prior by IARPA
• Delivered a data-powered mental health
screening system
Slide 5
The Durkheim Project
5. The math is like that of Moneyball
(Lewis), where instead of statistics on
your players, your looking at statistics on
your consenting users.
Big Data + Analytics provides a;
•Greater risk assessment (over time) for
individuals
•Greater number of individuals that can be
reached (e.g. rural communities)
•Cost Reductions. As we can;
• Reach over 100,000 veterans with
CURRENT systems
• Cut health care costs by isolating
those at TRULY risk
Slide 6
Predictive Health Care
6. Slide 9
• Red terms are from
Suicide Positive Cohort
• Yellow terms are from
Psychiatric/Non-Suicidal
Control
• Green terms are from
Control Group
Example Features
7. • We developed linguistics-
driven prediction models
to estimate the risk of
suicide.
• These models were
generated from
unstructured clinical notes
• From the clinical notes, we
generated datasets of
single keywords and
multi-word phrases
• We were able to initially
predict suicide with
65%* accuracy on a
small dataset.
• * We have since reached 70%
Slide 18
Machine Learning (with VA)
9. Dashboard
• Provides real-time monitoring of
patients
by Clinical Professionals
• Also enables a buddy-system
• (Live at Dartmouth Hitchcock
Medical Center)
Comparative Risk
Monitoring
• By Cohort
• Daily Tracking (think Fitbit™ for
mental health)
Slide 11
Dashboard Solution (for DARPA)
10. • Simplified user participation for consent, and privacy control
• Consent based access to online resources:
Slide 12
Opt-In Interface
11. • The Data is Secure/Protected
• Specifically, the data persists at a medical center, behind a HIPAA
compliant
firewall (e.g. Dartmouth Hitchcock Medical Center below)
• The highest professional security standards
Slide 13
Storage Layer
12. On July 2013, in conjunction with
Facebook, Inc., we publically launched
our project, with the aim of raising
awareness, as well as direct recruiting
of subjects.
https://www.facebook.com/notes/us-military-on-facebook/durkheim-project-la
… And since, we have been
positively covered by NPR,
NBC.com, Time.com, Fast Company,
The Boston Globe, Daily Mail UK,
Mashable, CIO, and CNN.com
Slide 15
Facebook Partnership
13. Intervention
Automated systems are coming online for potential patients and
families seeking
treatment, as well as passive intervention strategies (i.e. ‘safety
plan’narratives).
Interventions that
are designed to be
timely and
appropriate
Slide 16
15. Germanwings Flight 9525
"Germanwings co-pilot Andreas Lubitz
saw 41 doctors in five years;..." -ABC
"Germanwings Co-Pilot Andreas Lubitz
Was Treated for Suicidal Tendencies"
-WSJ
"Andreas Lubitz repeatedly set the
same plane for an unauthorised
descent earlier
that day." -BBC
"He told former girlfriend he was
planning an act so horrifying his name
would be remembered forever."
-Dailymail.co.uk
Should pilots be tested?
Slide 4
Insider Threat: The Murder-Suicide Scenario
16. Why isn’t DARPA funding this further?
The perception is that now that the science of the detection problem is
proven, the next steps are 'too medical' for their agency mandate
Why isn’t NIH funding this?
Heavy clinical/incremental medical studies focus & lack of dedicated
technology funding for NIMH
Why isn’t VA funding this?
The VA has scattered /uncoordinated attempts at similar. But the VA has a
core outreach problem (well known statistics). And leading VA researchers
are not (yet) taking Big Data approach
Why not Homeland Security?
Pilot studies (of similar, not same technology) have been conducted.
However, core policing engagement with suspects remains unchanged
Slide 19
Q&A
17. Suicide Risk:
An opt-in risk assessment and intervention routing system ALREADY exists.
We just need to deploy this at scale, through the responsible interface with
medical practitioners and social workers.
Mass Shootings:
Public surveillance and risk prediction, that ALSO protects personal rights IS
possible through proper training. One that enables law enforcement and
community preservation, where citizens are innocent before proven guilty.
Insider Threat:
Agencies can incorporate motive models into their behavioral
characterizations, rather than be blindsided.
Funding:
No government agency has an existing mandate to rise to the challenge.
Slide 19
Takeaways
18. Fund Appropriations for comprehensive technology initiatives utilizing
big data to detect mental health risks
• Research Programs
• Pilot Projects
Fund training programs for professionals that can use technology
for proper intervention, such as;
• Medical Staff
• Police and Firefighters
• Veterans and Social Workers
Public Awareness
• Spotlight on mental health/suicide prevention
• Congressional Hearings
• Community involvement in District
Slide 18
How Congress Can Help
19. Thank you
Chris PoulinChris Poulin
Patterns and PredictionsPatterns and Predictions
+1 617.755.9049+1 617.755.9049
Slide 19
20. • Opt-In is critical
• Technical Problem: How to build a system that collects, stores,
analyzes, and allows clinicians to react at Internet scale?
Phase 1:
• Machine Learning (Basic)
• Scalable Machine Learning
Phase 2:
• Opt-In Interface Layer
• Data Collection Layer
• Storage Layer
Phase 3:
• Automated Intervention
Slide 7
Appendix: Our Technology Approach