The Leading Physician Network worked with PSCI to develop a population risk stratification tool using their EMR data to identify high-risk patients for chronic conditions and reduce costs. The tool calculates individual "state of health" risk scores to target care management programs at those most likely to be hospitalized. This approach helped reduce hospitalizations and ER visits while improving case manager productivity.
Exploring the Future Potential of AI-Enabled Smartphone Processors
PSCI Case Study - Population Predictive Risk Analytics from PSCI
1. Leading Physician Network lowers Per
Member Per Month (PMPM) costs by
reducing acute care admissions for
chronic disease conditions through
effective care management
2. Patient centric medical home (PCMO )
leverages unique “state-of-health” population
risk stratification approach from PSCI.
PCMO uses Population Predictive Risk Analytics from PSCI.
3. PCMO SITUATION
• The successful patient centric medical home
(PCMO) is a leading provider of Primary Care
management services and is known for its
network of outstanding physicians in the local
market.
• The innovative, growth-oriented management
team made the decision to proactively acquire
the capabilities required to prosper in the
emerging climate of pay-for-performance.
4. OPPORTUNITY
• The ACO, bundled payment and pay-for-
performance models require transformational
process improvements in the primary care setting
to avoid unnecessary hospitalizations and ER
visits.
• The PCMO’s growth strategy was to offer the
local leading self-insured employers a compelling
value proposition with their focus on preventive
care and chronic care management, to minimize
the total cost of care to their membership across
the continuum-of-care.
5. OPPORTUNITY…
• The value proposition needed to be credible
and measurable in order to negotiate higher
rates for physician services and also increase
market share in the local market.
6. CALL TO ACTION
• After careful analysis of their patient
population healthcare costs, it was clear that
the highest cost population category was
chronic disease care and unnecessary ER
visits.
7. CALL TO ACTION…
• The management team, together with its
physician “think tank” came to the conclusion
that the key driver to manage chronic care
costs was to minimize hospitalizations and ER
visits with proactive, targeted care and case
management programs
8. CALL TO ACTION…
• To accomplish this, they needed analysis tools
to continuously identify and monitor “high
risk” patients proactively by major chronic
condition along with the risk drivers.
• They also wanted decision support tools to
measure patient risk based on current “state
of health” using clinical data from their
existing EMR systems on a monthly basis.
9. CALL TO ACTION…
• High risk chronic patients were defined as
those with a high probability for admission to
acute care facilities within the next 12-18
months due to complications.
• Furthermore, the team wanted physicians to
have the ability to analyze which processes
were needed to fill any gaps in care
management that may lead to
hospitalizations.
10. CALL TO ACTION…
• The required tools had to be
comprehensive yet provide easy-to-
absorb information with a clinical
perspective.
• The client insisted that physicians be able
to quickly and easily identify the key risk
drivers and prescribe appropriate care
and case management programs at
patient and population levels.
• However, the client were adamant that
these tools not be used for physician
profiling or as clinical outcome
predictors.
11. THE CHALLENGE…
• The team searched the market for a vendor to
provide decision support tools. They reviewed
risk adjustor applications, and determined the
tool did not adequately meet their requirements.
• Furthermore, the evaluation team learned that
most risk adjustment tools were primarily built to
address payer needs.
• They reported that claims-based risk predictor
tools did not serve their objectives for the
following reasons:
12. THE CHALLENGE…
• Acute care cost centric Risk adjustor models
are extremely complex and heavily skewed to
acute care costs and past resource utilization.
Models incorporate many variables that are
cost-focused and not under primary care
management control.
13. THE CHALLENGE…
• Claims-based Models are heavily based on
claims data with a payer-centric perspective,
whereas the physicians wanted clinical-centric
models.
• These models are very controversial and have
a negative connotation with clinical teams
because they are commonly used for physician
profiling.
14. THE CHALLENGE…
• Cost-prohibitive These tools are very
expensive and it is difficult to interpret results
from a care management perspective. Near
“real-time” analysis with weekly/monthly
frequency is prohibitively expensive.
15. THE CHALLENGE…
• Population-based models. Baseline models
are built at a population level and require a
large population mix for credible results – they
are not appropriate for smaller populations.
16. THE CHALLENGE…
• These models perform regression analysis at a
population level, then attempt to take scores to a
patient level.
• Risk scores at patient levels were based on relative
scores aligned with the population, therefore
individual patient scores would vary with population
changes, with no change in the individual state of
health.
• It was difficult to interpret the clinical drivers and
their impact on the risk scores
17. THE DECISION…
The evaluation committee realized
that risk adjustor tools were not
built to address primary care
provider-driven care management
programs. The team decided to
build an application in partnership
with an innovative healthcare
decision support provider.
18. THE DECISION
PSCI was selected to build a Population
Predictive Risk tool with the following
capabilities:
19. • PSCI, with the help of clinical teams, conducted
extensive research and identified nationally accepted
“state-of-health” models for each major chronic
condition to start with.
• PSCI developers worked with physician teams to make
the models more pragmatic in context of available
data, with standardized assumptions, and
simplification in agreement with larger expert teams.
• The solution collected clinical data from existing
ambulatory EMR, lab, pharmacy, and claims systems
on a regular basis to refresh patient “state-of-health”
risk scores.
THE APPROACH
20. PSCI’s EMR-based Population Risk Predictive Model
PSCI uses a patent pending, transformational
approach for predicting risk of hospitalization that
takes into account 6 dimensions. No one in the
industry has put all of them together to predict
risk of hospitalization/re-admission.
21. THE APPROACH…
• Calculate patient “state-of-health” scores by
chronic disease condition for the most
common chronic conditions for the target
population mix using latest patient records
from EMR
• The score would indicate the probability of
hospital admission for any given patient due
to complications within 12-18 months.
23. THE SOLUTION…
• Identify evidence-based best practices based
on data analysis and physician input for each
chronic condition.
• Provide insight and data for optimal care-
management programs for patient risk groups.
24. THE SOLUTION…
• Help physicians maximize pay-for-
performance and Shared Savings Model
(ACOs) and help physicians proactively
manage patient population risk.
• Not a point-of-care solution.
• Not an outcome prediction tool.
25. • Provides easy-to-understand risk score
drivers, and pinpoint which variable
(demographic, clinical, etc.) is contributing
to an adverse state-of-health at any given
time.
• Physicians and clinical teams then
determine what diagnosis, treatments,
and care management strategies to focus
on to improve the specific patient risk
scores.
THE SOLUTION…
27. RESULTS…
PSCI delivered Population Risk Analyzer, a care
management decision support tool that:
• Helped in reduction of hospitalizations & ER
visits with an increase in case manager and
care manager productivity.
28. RESULTS…
• Provides a state of health risk score for each
chronic condition for a patient or a population
based on current clinical information.
The risk scores are calculated at the patient level
and then rolled up to the population level.
• The solution enables physicians and
administrators in their local setting – ACOs, clinics
in an integrated health care system, etc. to look
at the information and identify clinically high-risk
patients ER visits/hospitalization/readmissions.
30. Target right patients (High Risk Patients) at right time
Strong individualized care management programs
Intensive, multi-level, multi-dimensional, high contact programs
Provider-driven programs
Broad programs have no impact
Data-driven care management analytics
16
RESULTS…Customized Care Management Programs
31. “BlueCross BlueShield has been running
medical home pilots since 2010 with Village
Health Partners in Plano and the 42 offices of
the Medical Clinic of North Texas. The pilots
improved care and saved an average of
$10.50 a month for 25,000 patients, said
Scott Albosta, a division vice president with
the insurance company.” - (Dallas Morning
News June 23, 2012).
OUTCOMES
32. By using near real-time patient health
records from EMRs along with financial
claims and demographics data, PSCI
presents clinical teams information that
allows them to understand the risk drivers
associated with patient care across the
patient population. By understanding the
clinical cost, quality and risk drivers,
physicians make interventions to have a
dramatic impact to lower the healthcare
cost curve.”
– Karen Kennedy, CEO – Medical Clinic of
North Texas
TESTIMONIALS
33. ABOUT PSCI
• PSCI is an innovative provider of predictive
population risk analytics for care management and
contract optimization leveraging EMR, Claims &
Demographics data for medical homes, physician
groups, ACOs, hospital systems, IDNs, and shared
savings programs.
34. ABOUT PSCI
• PSCI delivers predictive chronic disease models for
population state-of-health risk stratification, quality-
cost-risk visibility, "what-if" modeling and ACO
demand planning for improving overall healthcare
provider and payer performance.
• PSCI is critical to managing “At-Risk” populations and
pay-for-performance objectives. For more
information, please visit http://www.PSCIsolutions.com