1 carilion clinic’s journey on the population health management and big data highways june 5, 2014...
TRANSCRIPT
1
Carilion Clinic’s Journey on the Population Health Management
and Big Data Highways
June 5, 2014
Tom Denberg, MD
Chief Strategy Officer
Executive Vice President
Carilion Clinic
Big Data and Healthcare
• Big data is a term used for massive amounts of information that can be interpreted by analytics to provide an overview of trends or patterns.
• Organizations leverage big data by gathering records and information captured and then interpreting it with analytics.
• Common in other industries, big data has only recently begun to become a factor in healthcare. It has applications range from provider-specific business intelligence to scouring over an entire state's health records to pinpoint people who are at risk for certain ailments.
• Many believe that big data can help target early warning signs and improve patient safety
Healthcare IT News 2014
77
Enterprise Data Warehouse
Claims Data
AetnaEmployee
Group,ACO
(Wholehealth)Claims
LabRx
Eligibility
TMGMedicare
AdvantageClaims
CMSMedicare Shared Savings
SAP/BusinessObjects
Enterprise
EPIC EMROperationalDatabase(Cache)
QNTXMedicare
HMO (Majesticare)
OtherPlans - TBD
CLAIMS/PlanData Sources CARILION CLINIC
NIGHTLY
ETL
ET
L
Clarity Relational Database
Cloud-Based/ASP services
TemporaryClaims Staging
Database
Care Conerns/Gaps,
Risk
Stratific
ation Data
PopulationAdvisor
Premier/Verisk
Web-based User Interface
Enterprise Data Warehouse
EPIC EMR
88
Healthcare IT and ACOsThe Critical List
• Population identification - attribution• Identification of care gaps – Decision Support • Risk Stratification• Cross Continuum Care management • Quality and Outcomes measurement• Patient engagement• Telemedicine • Mixing claims and clinical data • Predictive modeling • Clinical information exchange
Excess Cost Domain Estimates
$210
$130
$190
$105
$55
$75
Cost in Billions of $$$
Unnecessary Services ($210 B)
Inefficiently Delivered Services($130 B)Excess Administrative Costs($190 B)Excessive Pricing ($105 B)
Missed Prevention Opportunities($55 B)Fraud ($75 B)
IOM. The Healthcare Imperative, 2010.
Clinician-Driven Sources of Excessive Health Care Costs (Population Health Management Focus)
• Preventable/avoidable hospital (re-)admission and ED visits (Case Management, Readmission Reduction)
• Missed prevention (Pay-for-performance)
• Unnecessary care (Utilization Management)
Key patient populations
Key Strategic Initiatives
Ambulatory Case ManagementPatient engagement, care coordination, Extensivists, palliative care, transitions of care protocols
Ambulatory Quality / Pay for Performance (P4P)Cancer screening, BP, lipid, A1c, etc.; various patient engagement and contact components
Sickest and/or highest-utilizing 5-10%Advanced CHF, COPD, IHD, DM, asthma, cancer, psychosocial problems
Rising-risk 40-50%Patients with less severe chronic illnesses or behaviors that significant elevate morbidity or mortality risks; HTN, DM, hyperlipidemia, tobacco use, obesity
Low risk 45-55%Patients without medical problems; focus on prevention, wellness, and connectivity to health system
Behavioral Health / Psycho-social
Pay-for-performance
• Core measures, value-based purchasing (Hospital)• HCAHPS (Hospital)
• HEDIS, NQF (Ambulatory) • CGCAHPS (Ambulatory)
•Pro
fusio
n of
met
rics
•Prim
ary c
are
emph
asis
•In
crea
singly
shar
ed p
rimar
y-ca
re/sp
ecial
ty ac
coun
tabil
ity
•In
crea
sing
num
ber o
f spe
cialty
-spe
cific
met
rics
Ration
ale:
- Reim
burs
e fo
r valu
e, n
ot ju
st
volum
e (m
axim
ize sh
ared
savin
gs)
- Dem
onstr
ate
not s
kimpin
g on
care
CLBSICAUTICHF Readmission rate……BP controlA1c controlBreast CA screening…
“Off hand, I’d say you’re suffering from an arrow through your head, but just to play it safe, let’s get an echo.”
Utilization Management
The Future- Proactive Care
• Identify patients at risk before they develop symptoms of heart failure • Maximize treatment of underlying conditions• Closer follow up• Delay or prevent the onset of severe heart
failure • Bend the disease curve
CHF Onset Project
• Collaboration ( Carilion, IBM, Epic) • 3 years data / 500,000 records reviewed• NLP used to obtain unstructured data (20M)• 8500 patients at risk
• 3500 identified with NLP• Risk score generated based on clinical , social
and demographic data • Score available in EMR • Develop treatment protocols to address at risk
patients.