1. Accountable Care and Evidence-based decision-
making
Eugene Rich MD
Director, Center on Health Care
Effectiveness

2. Inside the DC Beltway
 65 miles surrounded by reality
 Beltway view of the current US Health Care
System

4. CBO on promoting evidence based care to
address rising health care costs
 “…little rigorous evidence is available about which
treatments work best for which patients”
– Solution: CER/PCOR
 And”…financial incentives … tend to encourage
the adoption of more expensive treatments and
procedures, even if evidence of their relative
effectiveness is limited”
– Solution: provider payment reform
Orszag and Ellis, NEJM, Nov 2007

5. 2009- $1.1 Billion Investment in CER thru ARRA
Stakeholder Human & Comparative Dissemination
Needs Input & Data Scientific Effectiveness and
Identification Involvement Infrastructure Capital Research Translation
Horizon Community Research Evidence
Data base
Scanning Forum Training Generation
projects
Career Evidence
Electronic Data
Development Synthesis
Management
Forum Methods
Development
Populations
Conditions
Ty p e s o f I n t e r v e n t i o n s
Items in blue represent components derived from the HHS CER framework, those in yellow represent components taken from the
FCCCER strategic framework, and items in green represent components taken from both the HHS and the FCCCER frameworks.
5
5

6. Support for CER/PCOR after ARRA
Patient Centered Outcomes Research Trust
Fund
 grows to $600 M per year by 2014
– $50 M in 2011; $150 M in 2012
– Funding thru mandatory appropriation, and tax on
private health insurance
 20% to HHS and AHRQ for PCOR data
infrastructure, methods development and research
translation
 80% controlled by PCORI
– “not an agency … of the federal government”
– 21 member Board of Governors
– Establish and support agenda for “patient centered outcomes
research”

7. Public Investment in Health Care Effectiveness Research
AHRQ-ACA, PCORI, NIH, AHRQ-ACA, NIH AHRQ,
NIH, PCORI AHRQ CMMI, (PCORI) CMMI
Comparative Research
Research Evidence-
Infra- on polices
Effectiveness on using based, Affor
structure to
Research CER dable Health
for CER promote
Studies findings in care
using
practice
CER
•Medications
•Medical devices and
•Stakeholder technologies •Providers •Payment and
input •Medical and surgical •Patients
services, regulation
•Databases •Delivery
•Behavioral change •Monitoring
•Training strategies, Systems and feedback
•Methods •Delivery system
•Reviews interventions

8. Patient Centered Outcomes Research Definition
“Patient-Centered Outcomes Research (PCOR) helps
people and their caregivers communicate and make
informed health care decisions, allowing their voice to
be heard in assessing the value of health care
options.”
1. “Given my personal characteristics, conditions and preferences, what
should I expect will happen to me?”
2. “What are my options and what are the benefits and harms of those
options?”
3. “What can I do to improve the outcomes that are most important to
me?”
4. “How can clinicians and the health care system they work in help me
make the best decisions about my health and healthcare?”
www.PCORI.org
8

9. PCORI Funding Opportunities
 Assessment of Prevention, Diagnosis, and Treatment Options – projects
that address critical decisions that patients, their caregivers and
clinicians face with too little information (CER)
 Improving Healthcare Systems – projects that address critical decisions
that face health care systems, the patients and caregivers who rely on
them, and the clinicians who work within them
 Communication and Dissemination Research – projects that address
critical elements in the communication and dissemination process
among patients, their caregivers and clinicians
 Addressing Disparities – projects that will inform the choice of
strategies to eliminate disparities
 Accelerating Patient-Centered Outcomes Research and Methodological
Research – COMING FALL 2013
9

10. Patient Centered Research Questions
 A 47-year-old woman with rheumatoid arthritis has learned that her
primary care doctor recently joined a large medical group
…becoming part of the system’s patient-centered medical home.
– What should this woman know about the potential benefits or possible risks of
this new way of primary care practice compared to her current or other care
approaches?
 A world-class athlete has been advised by her sports medicine
physicians that she needs arthroplasty in each knee. She is
referred to a group of orthopedic surgeons …that are part of an
accountable care organization.
– How will this organizational model impact her care,
– and what information about the ACO should she know to determine whether
they will be likely to honor her strong preference for treatment that will return her
to maximal function as quickly as possible?
10

11. CER/PCOR Purpose II
 Feb 2009 ARRA CER investment
was driven in part by need to
increase value of Medicare and
Medicaid spending,
 but policymakers became sensitive
to public fears that CER will be used
to ration care
 This 2009 debate affected ACA
language on what PCOR is and how
it can be used
 E.g. PCORI not to fund work that
calculates “dollars-per-quality
adjusted life year (or similar
measure that discounts the value
of a life because of an individual's
disability)”
Rich EC, Docteur E, MPR CHCE Issue Brief 2010
11

12. Using CER: What’s Allowed by ACA
 Dissemination:
– AHRQ charged with disseminating findings
published by PCORI and other CER/PCOR
 Clinical decision support:
– PCORI/AHRQ expected to promote use of findings
via automated clinical support tools
 Use of PCORI findings in coverage and
reimbursement decisions by public programs
– findings can’t be sole input to Medicare coverage
decisions but ACA does not prohibit use
Rich EC, Docteur E, MPR CHCE Issue Brief 2010
12

13. CMMI: Learning How to Improve Care Delivery
CMS- Center for Medicare and Medicaid Innovation
(CMMI)
 $10B mandatory appropriation over 10 years
 Goal: better care and better health, at reduced costs through
improvement…
 identifying, testing and spreading new models of care and
payment
– Patient Care Models- eg interventions to reduce
healthcare-acquired conditions
– Seamless Coordinated Care- eg identifying and deploying
the best advanced primary care and health home models
– Community and Population Health Models- eg test new
care models that impact underlying drivers of heath
(smoking, obesity)

14. PCOR- Research Development Challenges
 New understanding of the consumers of CER study
products
 CER/PCOR focus on answering questions relevant to
typical clinician and patient decision makers
Rich EC, Bonham A, Kirch D, Academic Medicine 2011
14

15. Incorporating the clinical decision-maker perspective
 Soliciting the insights of patients and clinicians
 Incorporating the diversity of
– communities,
– cultures,
– patient perspectives,
– practice settings
 Recruiting representative research participants
and settings into research networks
Rich EC, Bonham A, Kirch D, Academic Medicine 2011
15

16. CBO on promoting evidence based care to
address rising health care costs
 “…little rigorous evidence is available about which
treatments work best for which patients”
– Solution: CER/PCOR)
 And”…financial incentives … tend to encourage
the adoption of more expensive treatments and
procedures, even if evidence of their relative
effectiveness is limited”
– Solution: provider payment reform
Orszag and Ellis, NEJM, Nov 2007

17. Fee for Service Payment
 Longstanding approach to physician reimbursement
 Risks well recognized
– Code of Hammurabi, Heraclitus, Ben Franklin, GB Shaw
 Physician as “seller of services”
– Buyer does not have physician’s specialized knowledge
– Buyer further disadvantaged by pain, anxiety, cognitive
impairment
 Principle-agent theory
– Physician contracts to act as patient’s agent
– Patients interests are advanced when the physician
(clinician) recommends services with evidence of benefit
17

18. 18

19. Decision-making at the Point of Care
 Patients seek physicians to address their
health concerns
– And relieve their symptoms/distress
 Each patient encounter generates numerous
decisions
 Physicians make these decisions in the face of
extensive and conflicting relevant evidence
– Many studies, few answers
 All diagnostic tests are imperfect
– Inherent risk of over-diagnosis and under-diagnosis
– Multiple sequential tests do not help
19

20. Clinician incentives can bias decisions
 Clinician beliefs about their professional role
 Assessment of “prior probability”
 Interpretation of clinical findings (eg over-diagnosis)
 Recollection of clinical research evidence
– 23,000 clinical trials published in English each year
 Maslow’s Hammer
– reputational bias
– pseudo-consensus
 Facilitating Adherence

21. Too Little? Too Much? Primary Care Physician Views
 95%- believe physicians vary in what they do for
identical patients
 42%- patients (in their practice) receive too much care
 6%- patients receive too little care
 Most important factors for aggressive practice
– Malpractice concerns - 76%
– Clinical performance measures- 52%
– Inadequate time to spend with patients-40%
– Financial incentives
• 62% subspecialty diagnostic testing could be reduced
• 39% primary care diagnostic testing could be reduced
B E Sirovich, S Woloshin, L Schwartz, Arch Intern Med. 2011;171(17):1582-1585

22. FFS and Point of Care Decision-making
 FFS offers straightforward method to
encourage delivery of services at the point of
care
– Patients have greater trust under FFS payment
 FFS may not provide consistent incentives to
promote evidence- based practice
– Poor calibration of fees- eg high margins for services
of limited effectiveness
 Potential impact of FFS imbalance on point of
care decisions
– Over or under-testing
– Over or under diagnosis
– Over or under treatment
22

23. Imaging for Low Back Pain*
 High margin for imaging studies for back pain creates
incentives for physician/clinician to …
– Promote increased patient awareness of medical services for the
problem
– Increase patient access for evaluation
– Perceive higher likelihood of conditions that require testing
– Provide services to help patients adhere to testing
recommendation
 If imaging study is an efficient means of diagnosis
candidates for a high margin treatment – then additional
incentives for physician/clinician to …
– Diagnose the condition that warrants the high margin treatment
– Provide services to help patients adhere to testing
recommendation * Overused service identified by “Choosing Wisely” program
23

24. Antibiotic prescribing in Sinus infection
 No direct FFS incentive of ABX RX (in US)
 FFS incentive to recommend an approach that
satisfies patient expectations
– ABX plausibly effective in addressing the likely
diagnosis
– Patients prior belief regarding ABX efficacy
– Patient desire to avoid missed work/school
– Patient preferences and shared decision-making
 Current FFS provides inadequate incentive to
educate patients regarding risks and benefits
 Patient satisfaction may not be enhanced by
efforts to discourage antibiotic prescribing
24

25. Treatment “under-management” for GERD
 GERD Rx “should be titrated to the lowest
effective dose needed to achieve therapeutic
goals”
 Evidence-based care requires:
– physician must contact asymptomatic patients on
chronic therapy for GERD,
– reduce medication dose as appropriate,
– Follow-up on symptom response and further adjust
medication
 Not easily rewarded via FFS
 May be viewed as unwelcome distraction by
asymptomatic patients
25

26. Payment reform options: potential impact
on evidence-based care
 Revised FFS
 FFS + P4Q
 Episode-based payment
 Global payment (capitation)
26

27. Revised FFS
 Advantages-
– Many current fees not reflective of physician work (some over-valued, some
under-valued)
– If margins for physician services are high, practices will increase use
– Increased payments can address underuse of highly effective services
 Disadvantages
– Reducing payments for overused services may not consistently reduce demand
• Inertia, prior beliefs
• “physician induced demand”
• Risks of payment reductions below actual cost
• Many overused services not driven by FFS incentives (antibiotic use)
• Under-management of chronic illness not easily addressed by encounter-
based FFS
– Challenges in adjusting ffs payments based on evidence of effectiveness
• Services often proven effective for one patient subgroup- benefits unclear
for others
• Ever-changing clinical research evidence
27

28. FFS w/ P4Q
 Advantages-
– Monitor/reward better chronic care management (eg
GERD management)
– Monitor/reward appropriate use of test or treatments (e.g.
back imaging, antibiotic use)
 Disadvantages
– Focus P4Q on high priority services
• physicians make numerous decisions per
encounter, 1000s of decisions per day
– Rectify conflicting P4Q signals from multiple payers
– Assuring salience to real-world decision-making
• Attribution to the correct clinician decision-maker
• Patient risk adjustment, benchmarking
– Quality measures ≠ evidence-based practice
28

29. 23,000 clinical trials/yr = Enough Evidence?
Robert Califf, IOM Meeting, 12 December 2008. Less than 20% of AHA/ACC heart disease
management recommendations are based on a high level of evidence and over 40% are based on
the lowest level of evidence: Level A evidence (multiple populations and risk strata) to Level C
(very limited population risk strata). The proportion of recommendations with high evidence levels
has not increased over time.

30. Heterogeneity of Treatment Effect
Diagram by J. Meddings
(Vijan & Hayward, Ann Intern Med 1997)

31. Stringent Dichotomous Measures
Don’t target patients most likely to benefit
– Ignore the heterogeneity of patient risk factors
Don’t help providers do the “right” thing
– Blunt instruments with little or no clinical nuance
Don’t take into account patient
preferences
– Often mandate care not wanted by well-informed patients
Could result in unintended consequences
– Polypharmacy, hypoglycemia, worse outcomes, wasteful
spending
R Hayward, 2012

32. FFS w/ P4Q
Overused +/-
test
Underused √
test
Over DX +/-
Under DX √
Overused +/-
Rx
Underused √
Rx
Under- +/-
mangd Rx
32

33. Episode-based payment
 Advantages-
– Single payment for all services needed during an
episode of illness
– Removes “piecework” incentive of FFS
– Incentive for constraining volume of services during an
episode of illness
– Over testing example-
• Physician discretion to make “evidence-based” use of
imaging for diagnosis and management of back pain
• Testing represents cost, not additional profit
33

34. Who to Give the Episode Payment To?
Practice environment and clinical decision-making
Collecting Data on Physicians and their Practices, AHRQ Report, 2012

35. Practice environment and clinical decision-making
 Patients  Physicians
– Age, gender, race/ethni – Personal characteristics
city – Clinical training
– Health concerns and – Current experience/
chronic conditions expertise
– Financial access to – Professional attitudes
care • Attitudes toward
– Education level, health evidence,
literacy • Attitudes towrd shared
– Patient decision-making, etc
preferences, expectati
ons, values

36. Practice environment and clinical decision-making
 Point of Care  Practice Organization
– Clinical focus – Practice organization size
• Inpatient, and specialty mix
outpatient, ASC – Practice ownership
– Clinical colleagues • Physician partnership
– Clinical workload • Private hospital
– Resources (support • Health plan
staff, examination • Academic medical center
rooms, patient – Practice governance and
educators) leadership
– HIT – Organizational culture
– Decision support – Sources of revenue, payer
– Care management mix
– Availability of DX/RX – Physician compensation
technology and incentives

37. HIT at the Point of Care
 How frequently do you use a computerized or
electronic system to perform the following
tasks at this practice location?
– Order laboratory tests
– Obtain clinical decision support
– Generate a list of patients overdue for tests or
preventive care
– Access standard order sets for a particular condition
or procedure
– Provide reminders for guideline-based interventions
or screening tests
– Electronically exchange patient clinical information
with any other clinicians outside your practice
organization or hospital
37

38. Practice environment and clinical decision-making
 Networks and  Market Environment,
Affiliations – Provider market
– Shared resources with concentration
other practice – Commercial payer
organization environment
(e.g., HIT, billing, equip
– Malpractice environment
ment, space)
– Community resources
– Formal relationships
with broader networks – Urban/rural
of providers
(e.g., IPAs, PHOs, ACO
s, etc.)

39. “Physicians don’t just work for money”
 Ability to do good- accessing/managing
resources related to what they care about (e.g.
innovative clinical programs, interesting clinical
problems or procedures)
 Ability to do important work- accessing/managing
high quality clinical program resources (e.g.
nurses, physicians
assistants, technicians, equipment, etc)
 Ability to do what they want- managing personal
time, personal administrative support, etc

40. Changing the employed clinician’s “margin”
Compensation Work environment
 % income at risk  Workload
– Work assignments
 Performance – call
measures – “hassles”
– “Productivity
measures”- eg billing  Support staff /space
– Quality metrics
– Patient satisfaction  Ease/difficulty obtaining
– Organizational financial tests, services
performance
 Recruitment /retention
 “Perks”
– Professional  Professional culture
development – leadership
40

41. Who to Give the Episode Payment To?
 Most physicians participating in episode-based
payment will be compensated by a larger entity
receiving the bundled payment
 The incentives presented to this larger entity will
be translated thru internal management to
influence clinical decisions at the point of care.
41

42. Physician Compensation Strategies and
Intensity of Care
 Highly capitated practice environments had
lower intensity of care for episodes of care
 Productivity payments had the highest
spending
 True for practice owners
 and for employed physicians
Landon, et al. The Relationship between Physician Compensation Strategies and the Intensity of
Care Delivered to Medicare Beneficiaries. HSR July 2011

43. Episode-based payment
 Disadvantages
– Episode-based payments may discourage evidence-
based testing and treatment during an episode of
illness
• PFTs in asthma
• Drug management in GERD
– Episode-based payments tied to diagnosis of
illnesses
• Potential incentive for over-testing to find episodes
• Potential incentive for over-diagnosis from test results
– Episode-based payments often tied to high cost
services (like surgical procedures)
• Potential incentives for over-Rx
 P4Q can help
– Same limitations as FFS
43

44. Episode-
based
payment
Overused +/- the role of episode-based
test payment reform in over-used
Underused +/- tests or treatments is highly
test contingent on how decisions
about these services are
Over DX +/-
incorporated into the definition of
Under DX √ episodes of care.
Overused +/-
Rx
Underused +/-
Rx
Under- +/-
mangd Rx
44

45. Global payment (Capitation)
 Advantages-
– Single payment for all services needed by a patient
during a year
– Removes “piecework” incentive of FFS
– Incentive for constraining volume of low- value services
for patients
– Incentives for providing services that are effective in
averting unnecessary spending on preventable illnesses
or illness complications
 Who to Give the Money to?
– Accountable care organizations
45

46. Incentives for Care of Low Back Pain
 incentive to reduce patient access to expensive
clinical services
 Incentive for clinician to perceive a lower likelihood
of conditions that require costly testing or
treatment
 Incentive for convincing patients of the risks of
additional imaging studies or interventions
 Incentives for promoting adherence to low cost
options
 Incentives for discouraging adherence to costly
interventions like advanced imaging or surgery
46

47. Global payment
 Disadvantages
– Capitation may encourage reduced access and under-
diagnosis
– Capitation may discourage evidence-based testing and
treatment
• PFTs in asthma
• Drug management in GERD
 P4Q can help
– Daunting limitations
• How to properly measure and reward myriad decisions at
the point of care
47

48. Global
payment
Overused √ •For some chronic conditions the
test intermediary receiving the capitated
payment can realize near-term financial
Underused +/- gains through improved chronic disease
test management.
Over DX √ •In many patients more evidence- based
point of care decisions confer near term
Under DX +/- costs, with savings realized only many
years hence, or not at all
Overused √
Rx
Underused +/-
Rx
Under- +/-
mangd Rx
48

49. Revised FFS w/ P4Q Episode- Global
FFS based payment
payment
Overused +/- +/- +/- √
test
Underused √ √ +/- +/-
test
Over DX +/- +/- √
Under DX √ √ √ +/-
Overused +/- +/- +/- √
Rx
Underused √ √ +/- +/-
Rx
Under- +/- +/- +/-
mangd Rx
49

50. Policy Goals for Payment Reform
 Promote evidence-based decision-making at the
point of care
– Patients seek clinicians they can trust to recommend “what
is best”
– Professional societies and policy makers want clinicians to
recommend evidence-based services
– Incentives that do not consistently reward evidence-based
care will prove unacceptable to both patients and clinicians
 Other purposes for broader payment reform
– Correcting clinician specialty imbalances
– Addressing care fragmentation
– Enhancing the role of primary care clinicians
– Promoting new modes for addressing patient concerns
50

51. Incentive Reform to Promote Evidence-based Care
 There are many mechanisms for paying
physicians; some are good and some are bad.
 The three worst are…
 fee for service, capitation, and salary.
– James Robinson

52. Incentive Reform to Promote Evidence-based Care
 Recalibrate productivity measures to
recognize physician costs (margin) at the point
of care
 Monitor patterns of care relative to highly
effective services
– Overused and underused tests
– Over- and under-diagnosis
– Overused and underused treatments
– Under management of chronic conditions
 Choosing Wisely Program
– One place to start
52

53. For under-used, highly effective, tests or treatments
 Address clinical issues
– Knowledge, diagnostic skills
– Conflicting interpretations/ professional standards
– Easy access to knowledge resources and decision support
 Incentive reform
– Re-evaluate for mis-calibrated physician costs
– If productivity measures look appropriate consider
• Compensation plan
– Increased FFS payment (to jumpstart increased use for highly
effective services)
– P4Q incentives to increase awareness of appropriate use
• Work environment
– Workload, Support staff , Ease of ordering/obtaining
– Professional culture
– ??Reminders (recent surveys show reminder burden)
53

54. For over-used in-effective, tests or treatments
 Address clinical issues
 Incentive reform
– Re-evaluate for mis-calibrated physician costs
– If productivity measures look appropriate consider
• Compensation plan
– Eliminate production incentive for this service
– P4Q incentives to increase awareness of appropriate use
– Production incentives/targets based on expected utilization
• Work environment
– Ease of ordering/obtaining
– Referral process
– Workload, Support staff
– Professional culture
54

55. Policy Goals for Payment Reform
 Promote evidence-based decision-making at the
point of care
– Patients seek clinicians they can trust to recommend “what
is best”
– Professional societies and policy makers want clinicians to
recommend evidence-based services
– Incentives that do not consistently reward evidence-based
care will prove unacceptable to both patients and clinicians
 Other purposes for broader payment reform
– Correcting clinician specialty imbalances
– Addressing care fragmentation
– Enhancing the role of primary care clinicians
– Promoting new modes for addressing patient concerns
55

56. Promoting evidence based care to
address rising health care costs
– “…little rigorous evidence is available about which
treatments work best for which patients”
 Solution: CER/PCOR
– And”…financial incentives … tend to encourage the
adoption of more expensive treatments and
procedures, even if evidence of their relative effectiveness
is limited”
 Solution: Provider payment reform
Orszag and Ellis, NEJM, Nov 2007