Population Health Management
at Partners HealthCare
Timothy Ferris, MD, MPHSVP, Population Health Management, MGH, MGPO and Partners HealthCare
Integrated Care ConferenceOctober 11, 2016
Center for Population Health Confidential – do not copy or distribute
The Ecology of Medical Care- Historic View
White KL, Williams TF, Greenberg BG. The ecology of medical care. N Engl J Med 1961;265:885-892
Monthly Prevalence of Illness and the Provision of Medical Care.
Adult population at risk
Adults reporting symptoms
Adults consulting a physician 1 or more times per month
Adults admitted to hospital 1 or more times per month
Adults referred to academic medical center per month
Adults referred to another physician per month
Center for Population Health Confidential – do not copy or distribute
The Ecology of Medical Care- Current View
Monthly Prevalence Estimates of Illness and the Provision of Medical Care.
Adult population at risk
Adults reporting symptoms
Adults visiting alternative medical care
Adults receiving home health care
Adults visiting outpatient clinic
Adults considering seeking medical care
Adults visiting MD office
Adults visiting ED
Adults hospitalized
Adults hospitalized at AMC
Adapted from Green, LA, Fryer GE, Yawn BP, Lanier D, Dovey SM, Green LA. The Ecology of Medical Care Revisited. N Engl J Med 2001; June 28, 2001
Center for Population Health Confidential – do not copy or distribute
The Ecology of Medical Care- Future State
Adult population at risk
Adults reporting symptoms
Adults receiving home health care
Adults visiting outpatient clinic
Adults considering seeking medical care Adults visiting MD office
Adults visiting EDAdults hospitalized
Adults hospitalized at AMC
Adults consulting physicians through e-visits
Adults consulting physicians through virtual visits
Adults receiving home observationAdults receiving tele-monitoring
Adapted from Green, LA, Fryer GE, Yawn BP, Lanier D, Dovey SM, Green LA. The Ecology of Medical Care Revisited. N Engl J Med 2001; June 28, 2001
Adults admitted directly to SNF
Adults seeking palliative care
Monthly Prevalence Estimates of Illness and the Provision of Medical Care.
Center for Population Health Confidential – do not copy or distribute 5
• The number of adults reporting symptoms has remained unchanged but the breadth and specificity of services has grown dramatically.
• Payment system creates the ecology of medical services.
• FFS system requires highly specified services- difficult to customize to patient needs and still get paid.
• Risk contracting does not require specification of services.
Key takeaways from the changing health care ecology.
Center for Population Health Confidential – do not copy or distribute 6
Health care “crowds out” other spending
Massachusetts, FY 01 vs. FY 14, in billions
Center for Population Health Confidential – do not copy or distribute 7
What we’re facing…
• The market is returning to techniques used during managed care in the 1990s [closed networks, budget-based risk, cost sharing, restriction of choice] – will this generate the same backlash?
• But...
• The economic imperative is stronger
• Government is proactive (Massachusetts 3.6% cap on healthcare cost growth)
• Rate of change is slower (caps on increases, not cuts)
• And we have…
• Better health IT and data for population management
• Strategies and tactics that we know will improve care and reduce costs
Center for Population Health Confidential – do not copy or distribute 8
Implications for providers
Challenges
1. We need tactics that will be successful under any new payment model
2. How to make external incentives meaningful to our clinicians
• Must not undermine trust
3. Moving at the right pace
Too fast: we will lose the docs in the rush to implement – MDs attitude often creates the patient's attitude (managed care backlash)
Too slow: will mean not succeeding under the contracts and forcing more blunt solutions
Center for Population Health Confidential – do not copy or distribute 9
Telling our story
~48 PHM publications since 2012
• Highlights
• A Modified “Golden Rule” for Health Care Organizations (2012) Mayo Clinic Proceedings
• Massachusetts General Physicians Organization's Quality Incentive Program Produced Encouraging Results (2013) Health Affairs
• Balancing AMC’s Missions and Health Care Costs –Mission Impossible? (2013) New England Journal of Medicine
• The “Medical Neighborhood’ Integrating Primary and Specialty Care for Ambulatory Patients (2014)
JAMA
• Initial Results of a Cardiac E-Consult Pilot Program (2014) Journal of the American College of Cardiology
• Operational Lessons from a Large Accountable Care Organization (2014) Journal of Clinical Outcomes and
Management
http://www.partners.org/Innovation-And-Leadership/Population-Health-Management/About-PHM/Publications.aspx
Center for Population Health Confidential – do not copy or distribute 10
• The largest integrated academic health care system in the United States
• In addition to two Academic Medical Centers, PHS includes:
• Community and specialty hospitals
• Managed care organization
• Community health centers
• Physician network
• Home health and long-term care services
Background on Partners HealthCare (PHS)
PHS by the Numbers •1.5M patients served annually
•6,500 physicians
•9,100 nurses
•64,000 employees
•$10.9B in revenue
• In 2012, we entered new contracts that put us at risk for medical expense trend
• PHS currently have over 500,000 lives in various accountable care relationships
Center for Population Health Confidential – do not copy or distribute 11
The path we’re traveling at Partners
Pressure to reduce cost trend
New contracts with risk for trend
Internal PerformanceFramework (IPF)
Investment in Population Management Infrastructure
Changes to Partnersorganization structure
Primary Care Specialty Care
Care Continuum Patient Engagement
IS/Analytics
Improved quality and lower cost trend
Implement new local incentives/compensation
Network Composition
New relationships with community hospitals
and doctors
Center for Population Health Confidential – do not copy or distribute
Over 500,000 lives in accountable care contracts(~60% of primary care patients; ~20% of system wide revenue)
Self Insured
Partners Plus
Commercial population, but savings accrue directly to Partners,
and improves our own lives
3
Covered lives: ~100k
Commercial
Alternative Quality Contract (AQC)
Younger population, specialists critical to
management
1
Covered lives: ~350k
Medicare
Pioneer Accountable Care Organization
Elderly population, care management
central to trend management
4
Covered lives: ~100k
Medicaid
Neighborhood Health Plan
Population with significant disability, mental health, and
substance abuse challenges
2
Covered lives: ~30k
12
Our contacts 4 years in…
Lives under the Accountable Care Model
Center for Population Health Confidential – do not copy or distribute
Managing incentives in contracts that include risk for total medical expense (TME) trend
Risk/Shared Saving
Performance on TME for thedefined population
Performance on claims-based quality and safety metrics
System-wide Performance programs
Cost-standardized medicalexpense trend
EHR based quality and safety metrics
Compensation incentives
System wide infrastructure investments
Tax on clinicalrevenue
Local infrastructure investments
Risk/Shared Saving
13
Payers
Partners
Physicians and Hospitals
Center for Population Health Confidential – do not copy or distribute
Payer 1
Payer 2
Payer 3
Payer 4
Payer 5
Payer 6
Partners
Internal
Performance
Framework
Adoption of
Tactical Programs
Lower the Cost
Standardized Trend
Improve Performance
on
Quality Metrics
Risk Adjusted Cost
Trend Benchmark
14
How the Internal Performance Framework (IPF) works
Strategic
Programs
Performance
Metrics
Local
Programs
Physician
Orgs.
•We determine the content,
not payers
•Renewed annually
•Shared goals, but
participants can choose
different paths
•Forfeited funds pay external
contract losses
Center for Population Health Confidential – do not copy or distribute
2016 Internal Performance Framework (IPF)
Implementing PHS
Quality/Efficiency Strategies Reducing Medical Trend
Improving Select Quality
Measures
40% 40% 20%
Patient Centered Medical Home
• NCQA Recognition
iCMP (High Risk Care Mgmt)
• Process and outcome measures
• Innovation
Specialty Programs
• PCP/Specialty Collaborative Care
Agreements and E-Consults
• Specialty Programs (virtual visits,
PrOE, PROMs)
• Innovation
Hospital Metrics
• Medicare Spending per Beneficiary
• Nursing Process Measures
(pressure ulcers, falls with injury)
• Quality of Discharge Summary
Post-Acute Care Measures
• Readmissions
• Care Transitions
Trend Target (adult & pediatrics)
Big 3 Commercial: Cost
Standardized Medical
Expense (CSME)
Shared Risk: Hospitals and MDs
• Adult Diabetes Outcomes (3)
• Adult CVE and HTN Outcomes (2)
• Colorectal Surgery Protocols
• Menu of Process & Outcomes
Measures:
• Patient Experience (process &
outcomes)
• Infection Control Process
(CAUTI, CLABSI)
• CMS Measures (PSI 90,
NHQMs)
Adult MD Only
• Diabetes Screenings (3)
• Cancer Screenings (3)
• Depression Screening
• Patient Experience Composite
Pedi MD Only
• Asthma Composite
• Well Child Visits
• Patient Experience Composite
• Hospital Only
• Menu of Process & Outcomes
Measures (see above)
Center for Population Health Confidential – do not copy or distribute
Center for Population Health Confidential – do not copy or distribute
Center for Population Health Confidential – do not copy or distribute
Center for Population Health Confidential – do not copy or distribute
Center for Population Health Confidential – do not copy or distribute
Center for Population Health Confidential – do not copy or distribute
Center for Population Health Confidential – do not copy or distribute
Center for Population Health Confidential – do not copy or distribute
Center for Population Health Confidential – do not copy or distribute
Primary Care Strategy
Problem• Primary care as currently organized has
limited capacity to meet both acute and chronic illness needs.
• The work requirements of primary care are unsustainable.
Approach• Provide additional resources and match staff
skills with required work.• Phased adoption using milestones and
NCQA recognition.• Align compensation with goals
Progress• 77% of practices achieving milestones (n=221)
• 38% of practices achieved certification• Improved performance on primary and
secondary prevention metrics• Improved provider satisfaction
Partners Practices PCMH Readiness and NCQA Status
The 10 Building Blocks of High-Performing Primary CareThomas Bodenheimer, M.DAnnals of Family Medicine
March/April 2014 vol. 12 no. 2 166-171
Readiness
•77% at readiness for NCQA L3•38% with NCQA L3
Center for Population Health Confidential – do not copy or distribute 25
Integrated Care Management Program (iCMP)
Problem• Expenses are concentrated in a small % of
patients with multiple chronic conditions (9% of Medicare, 3% of Medicaid, 1% of commercial).
• Self-managing multiple chronic conditions challenging without assistance.
Approach• Identify high-risk patients and provide
care management and individualized care management plan.
• Demonstrated 7% cost reduction, reduced admissions, and 4% lowermortality
Progress• 13,972 high-risk patients actively
enrolled with a care plan (total iCMP patients)
• 84 care managers• 21 social workers • 6.6 pharmacists • 9 community resource specialists
• Lower rates of hospitalizations andED visits
Team
PMPM total medical expense lower after starting iCMP
Center for Population Health Confidential – do not copy or distribute 26
Behavioral Health Integration
Problem• High prevalence rate of depression in primary
care (10%) with half of patients receiving treatment from their primary care provider.
• Behavioral health issues increase the cost for patients with chronic illness 3-5x.
Approach• Consultations for primary care physicians
with behavioral health specialists. • Care management for patients with
depression and anxiety offered within primary care setting.
• Training and decision aids to support primary care.
Progress• 68% of primary care patients seen from Oct to
Dec 2015 screened for depression (n=143,438)
• Over 5,800 consults averaging over 300 per month in 2015.
• 36 Collaborative Care practices with 1,650 patients enrolled
Collaborative Care Model
Patients
Primary Care Physicians
New Role: Mental HealthCare Coordinators
Psychiatrist and Social Workers
406 640914 1157 1292 1458
4061046
1960
3117
4409
5867
0
1000
2000
3000
4000
5000
6000
7000
Q4 2014 Q1 2015 Q2 2015 Q3 2015 Q4 2015 Q1 2016
Quarterly Consults
Quarterly Cumulative
Total : 5,687
Center for Population Health Confidential – do not copy or distribute 27
Problem• Increase in demand for specialist services has
led to long wait times for appointments.
• 20% of referrals are for relatively simple questions that can be addressed by email.
Approach • Develop clinician to clinician consult
program in which referring physicians can obtain input from specialists directly and rapidly, without requiring a face-to-face visit.
• Participating MDs are paid for their time.
Progress• 35 active specialty practices (BWH, MGH,
TCMA)
• 5,983 e-Consults performed
• ~4,891 visits avoided (~$1.2M in savings)
e-Consult Request Form
Total e-Consults: 5,983
e-Consults
Center for Population Health Confidential – do not copy or distribute 28
1) Relay Health Online Asynchronous Visits
In online learning envir onments, exchanges between students and teachers ar e frequently enacted asynchronously
rather than in simultaneous or face-to-face conversations. This type of communication taking place at dif ferent times
is a standard protocol for many online learning, auction, and business web services. W ith RelayHealth, a provider of
health-related web services, the Virtual Practice is testing a tool that conducts asynchr onous exchanges between phy -
sicians and patients to conduct online visits. V isits are available for about 100 non-ur gent symptoms and conditions
commonly seen in a primary car e practice. Patients login to the RelayHealth website and complete a r elevant online
interview using RelayHealth's web -
Visit®. The visit is conducted asyn -
chronously and allows patients to
request advice about non-urgent
symptoms and avoid unnecessary
office visits for minor pr oblems.
Using the RelayHealth e-visit plat -
form, patients participate in online
medical interviews that gather and
document key data about symp -
toms. Results of these online inter -
views are relayed to the physician.
A physician in the practice r eviews
the patient’s responses from the web
interview and determines a tr eat-
ment plan using communication
channels online. If the patient’s con -
dition requires further evaluation,
the physician will request that the
patient visit the office in person.
Currently, 357 patients have enrolled in the pilot study. Results of this pilot will be announced shortly .
2) Synchronous Communication | The Virtual Visit
We recently concluded two research studies which compared face-to-face visits with virtual visits using web cameras.
Though virtual visits ar e not meant to replace the traditional face to face visit in primary car e, virtual visits may be a
viable option in circumstances where patients need to be routinely monitored (i.e., in chronic conditions like diabetes,
hypertension, obesity or depr ession). Virtual visits may also be ef fective for triage of acute non-ur gent issues like up-
per respiratory infections or back pain.
In these studies5, a physician conducted visits with existing patients known to the physician's practice with comput -
ers equipped with web cameras and videoconfer encing software. We examined the feasibility, effectiveness and ac-
T h e Vi r t u a l P r a c t i c e | m g h v i r t u a l p r a c t i c e @ p a r t n e r s . o r g | Te l : 6 1 7 - 7 2 6 - 6 6 7 7 5
5 Dixon RF Stahl JE. Virtual Visits in a General Medicine Practice: A Pilot Study. Telemedicine and e-Health. August 1, 2008, 14(6): 525-530 .
RelayHealth | Sore Throat Interview
Problem• Increase in demand for in-person follow-up
visits results in long wait times and inconvenience (e.g. travel, time from work) and cost (e.g. parking, co-pays).
Approach • Develop two alternatives for in-person
follow-up visits for patients:
• Virtual Visits – real-time interactions between patients and providers using video.
• e-visits – web-based interactions using questionnaires to manage low acuity issues (i.e. cold, ear ache, etc) and chronic disease.
Progress• 395 clinicians conducted virtual visit/e-visit• 6,621 virtual visits performed • 11,664 e-visits performed
Virtual Visits
e-visits
•virtual visits: 6,621•e-visits: 11,664
Virtual Visit & e-visit Volume
Virtual Visits
Center for Population Health Confidential – do not copy or distribute 29
Problem• Overuse of surgical procedures, which is
difficult to track and document, is costly and may not result in providing the highest quality of care to patients.
• Payer utilization process burdensome and ineffective.
Approach • Develop web-based decision support tool to
assess the appropriateness of surgical procedures.
• Improve decision-making process for patients and provide personalized consent form with risks/benefits.
• Reduce administrative burden associated with prior authorization.
Progress• 16 practices implemented PrOE• 9,391 PrOE assessments performed
*Recently completed, in process of launching to practices.
0%
20%
40%
60%
80%
100%
Percent Appropriate by Procedure
Appropriate May Be Appropriate Rarely Appropriate
PrOE Procedures
Cervical Spine* Lumbar Spine
Total Hip/Knee Replacement* Carotid Endarterectomy
Vena Cava Filter Placement Carotid Artery Stenting
Coronary Artery Bypass Graft Prophylactic Mastectomy*
Valve Replacement/Repair Mohs
Diagnostic Catheterization Weight Loss Surgery*
Percutaneous Coronary
InterventionIncisional Hernia Repair
ICD/CRT Implantation* Prostate Biopsy
Hysterectomy*
Procedure Decision Support
Center for Population Health Confidential – do not copy or distribute 30
Patient Reported Outcome Measures (PROMs)
Problem• Traditional measures (readmissions,
infections) fail to measure value and improve symptoms, activities of daily living, and quality of life following an intervention.
Approach• Collect measurement of patient-reported
outcomes on mobile devices in clinics and from home.
• Use real-time trend data to inform patient care and aggregate data for decision-making, quality improvement, and demonstration of value.
Progress• 136,794 surveys collected• ~24 specialties, ~61 clinics across
Partners
Knee Replacement: Quality of Life
Knee Replaced
KO
OS
Qu
alit
y o
f L
ife
Sco
re(0
-100
, Hig
her
is B
ette
r)
Days Before/Since Surgery(From ~1y before to 1y after)
Total PROMS: 136,794
Total PHS PROMs CollectionMarch 2014- May 2016
361910051
19086
34703
52212
69850
93533
120094
136794
0
20000
40000
60000
80000
100000
120000
140000
160000
Jun-14 Sep-14 Dec-14 Mar-15 Jun-15 Sep-15 Dec-15 Mar-16
Center for Population Health Confidential – do not copy or distribute
Care Continuum: alternatives to the ED and hospital
Problem Approach Progress
Partners Mobile Observation Unit (PMOU)
Unnecessary ED visits and admissions/readmissions are high cost and do not always provide patients with coordinated care in the right place at the right time.
Provide patients with care at home by a nurse
•158 patients admitted to program (84% of referred patients)•~15% bounce back rate for hospitalization or emergency department
Congestive Heart Failure (CHF) Telemonitoring
Heart failure patients have high readmission rates.
Provide CHF patients with home monitor to track and record symptoms
•714 patients enrolled in the program•20 new patients (on average) monitored each month
Post-acute Care Transition Programs
Unnecessary hospitalizations due to Medicare’s 3-day inpatient rule and high rates of readmissions from Skilled Nursing Facilities (SNFs).
•ACO permits waiver of the 3-day rule.
•Developed a network of 61 SNFs.
•296 waivers utilized•59% of ACO patients discharged to network SNF• 3% lower length of stay• 2% lower 30 day
readmission rate
Urgent Care Access to services is a barrier for patients and ED services are costly
Formed a joint venture with MedSpring to offer urgent care sites across MA
•First urgent care site open •Next two locations opening in late 2015
Center for Population Health Confidential – do not copy or distribute 32
Patient Engagement Programs
Vidscrips, Shared Decision Making, and PCOI
Problem• Lack of easily accessible education materials and
decision aids, impacts patients’ ability to manage their care effectively and make informed decisions about their treatment.
Approach• Develop short, single-topic videos featuring a
patient's own healthcare provider.• Provide evidence-based decision aids (DAs) to help
inform patients making treatment decisions.• Develop a resource library of patient education
materials (Primary Care Office Insight).
Progress• 454 total vidscrips filmed by 361 participating
physicians. There have been 31,321 unique views.• 16,261 decision aids provided to patients with 880
providers trained to prescribe DAs. • 192,078 hits to Patient Education handouts.
Total : 31,321
Total :16,261
Total : 444,516
Center for Population Health Confidential – do not copy or distribute 33
Analytics and IS Infrastructure
Problem• Disparate information and clinical data
systems impairs our ability to effectively redesign care delivery systems and measure our performance.
Approach• Single Electronic Health Record (EHR)• Enterprise data warehouse (EDW)
• hospital volume and margin• assessing variation in episodes of care• monitoring medical expenses and
utilization • monitoring quality metrics
Progress •eCare implemented at BWH, NWH, and MGH •693 EDW users
Total : 693
0
20
40
60
80
100
120
140
160
EDW Users by Quarter
New this Quarter Revoked
Center for Population Health Confidential – do not copy or distribute 34
• Executive Dashboard
– Contract Performance
– Key Quality and Outcomes Indicators
– Implementation Goals (rolled up view)
• Management Dashboard
– Local Performance Comparisons
– Quality and Outcomes Comparisons
– Implementation Goals
• Local Practice Dashboards
– Care Gap Registries (prevention, CVD, DM, etc)
– Physician Utilization Variation
How are we measuring progress?
Center for Population Health Confidential – do not copy or distribute 35
Population Health Management Index
• The Population Health Management Index (PHMI) calculates our progress in program implementation and healthcare cost savings.
• PHM programs “touched” 45,000 patients in 2014 and 105,000 patients in 2015.
• By the end of 2016, PHM is predicted to assist 258,000 patients with a cumulative total savings of $113 million.
• Annual savings in 2016 ($52 million) is the same amount as the PHM budget.
• Currently, PHM is predicted to generate $300+ million cumulative savings by end of 2018.
Estimated health care cost savings 2014-e2018
Center for Population Health Confidential – do not copy or distribute 3636
PHM implementation dashboard
Center for Population Health Confidential – do not copy or distribute 3737
PHM implementation dashboard
Center for Population Health Confidential – do not copy or distribute
3 Essential Trusting Relationships in HealthCare
38
Patient
Clinician
Management
Clinician
Management
Payer
Center for Population Health Confidential – do not copy or distribute
Closing gaps in service delivery
Reach/engage
Find opportunities
for improvement
Intervention
Identification
Realized Improvement
Adapted from Eisenberg J. Transforming insurance coverage into quality healthcare:Voltage drops from potential to delivered quality.JAMA 2000;284:2100-07.
Potential Improvement
39
Center for Population Health Confidential – do not copy or distribute 40
Rogers’ Law of Diffusion
Buy-in for change follows Rogers’ Law
Generating a Sense of Urgency and Commitment is about leveraging EarlyAdopters to engage the Early and Late Majority
Co
mm
itm
ent
Center for Population Health Confidential – do not copy or distribute
Lessons Learned
1) Risk contracting imposes a much needed goal on providers for evaluating their services through a cost-benefit lens in addition to a volume lens
• Fee-based systems lack this perspective
• Budget-based systems lack this perspective
2) Risk contracts themselves cannot adequately adjust for the complexity services and/or differences in health status, therefore:
• Penalties must be limited
• Leadership should be accountable for
delivery system performance
• Physicians should be accountable for:
• Using the tools available to them
• Having excellent patient experience metrics
3) Not possible to write aspirations into a contract
4) Pace
41
Both systems operate in silos. Risk contracting on total cost of care provides incentives for silos to collaborate
Physician Tools for Managing Patient Care
Primary Care Specialty Care
• Care managers• Gap registries• E-consults• Behavioral Health• Mobile Obs Unit• Palliative Care
• E-consults• Appropriateness• PROMs• Variation data• Mobile Obs Unit• Palliative Care