PAY-FOR-PERFORMANCE, PUBLIC REPORTING AND RACIAL DISPARITIES How are programs being designed?
Alyna T. Chien MD MS Marshall H. Chin MD MPH Andrew M. Davis MD Lawrence P. Casalino MD PhD The University of Chicago Academy Health, Walt Disney World Dolphin Hotel June 5, 2007
“P4P” PROGRAMS ARE PROLIFERATING
• Medicare
• •
Medicaid Commercial HMOs
Centers for Medicaid and Medicare Services 2004 & 2005, JAMA Just this month Centers for Healthcare Strategies 2005, Commonwealth Fund 2007 Rosenthal NEJM 2007
INTENDED EFFECT OF PROGRAMS
Quality
Time
EFFECT ON DISPARITIES
Quality
?
Time
EMPIRICAL EVIDENCE
Intended effect (~14 studies) Unintended consequences (4 studies)
1 rewards those already doing well 2 “cherry-picking” for disease severity 1 “cherry-picking” for racial/ethnicity
EMPIRICAL EVIDENCE
Intended effect (~14 studies)
2 documentation improved 5 significant effect 4 mixed effect 3 no effect
Unintended consequences (4 studies)
1 rewards those already doing well 2 “cherry-picking” for disease severity 1 “cherry-picking” for racial/ethnicity
Dudley 2004; Peterson 2006; Lindenauer 2007 Rosenthal;,Hofer;Shen; Werner
EMPIRICAL EVIDENCE
Intended effect (~14 studies)
2 documentation improved 5 significant effect 4 mixed effect 3 no effect
Unintended consequences (4 studies)
1 rewards those already doing well 2 “cherry-picking” for disease severity 1 “cherry-picking” for racial/ethnicity
Dudley 2004; Peterson 2006; Lindenauer 2007 Rosenthal;,Hofer;Shen; Werner
POTENTIAL IMPACT ON DISPARITIES
NEUTRAL NARROWING WIDENING
More of the same
Shores up differences
Has differential impact
Identifies minority sub-groups in need of more tailored programs
Widens resource gaps between “rich” and “poor” organizations
Induces cherry-picking
POTENTIAL IMPACT ON DISPARITIES
NEUTRAL NARROWING WIDENING
More of the same
Shores up differences
Has differential impact
Identifies minority sub-groups in need of more tailored programs
Widens resource gaps between “rich” and “poor” organizations
Induces cherry-picking
POTENTIAL IMPACT ON DISPARITIES
NEUTRAL NARROWING WIDENING
More of the same
Shores up differences
Has differential impact
Identifies minority sub-groups in need of more tailored programs
Widens resource gaps between “rich” and “poor” organizations
Induces cherry-picking
POTENTIAL IMPACT ON DISPARITIES
NEUTRAL NARROWING WIDENING
More of the same
Shores up differences
Has differential impact
Identifies minority sub-groups in need of more tailored programs
Widens resource gaps between “rich” and “poor” organizations
Induces cherry-picking
LEADER INTERVIEWS
DESIGN:
Semi-structured interviews
LEADERS – 3 SECTORS:
Nationally prominent programs (ALL 5) Medicaid programs (ALL 11) Commercially-sponsored programs (RANDOM 10)
TELEPHONE INTERVIEWS
4 Domains Constant comparative methods of qualitative analysis
LEADER INTERVIEWS
Nationally Prominent
Response rate 5/5
Financial +Reputational
Achievement + Improvement + Risk Adjustment + Tournament Clinical processes
State Medicaid
4 / 11
Financial
Achievement + Improvement
Commercial Health Plans
6 / 10
Financial +Reputational
Achievement + Improvement
Incentive type
Incentive trigger(s)
Performance measures
Clinical processes +Clinical access + Patient satisfaction
Clinical processes +Outcomes +Clinical access +Patient satisfaction +Efficiency +Formulary use
LEADER INTERVIEWS
Nationally Prominent
Response rate 5/5
Financial +Reputational
Achievement + Improvement + Risk Adjustment + Tournament Clinical processes
State Medicaid
4 / 11
Financial
Achievement + Improvement
Commercial Health Plans
6 / 10
Financial +Reputational
Achievement + Improvement
Incentive type
Incentive trigger(s)
Performance measures
Clinical processes +Clinical access + Patient satisfaction
Clinical processes +Outcomes +Clinical access +Patient satisfaction +Efficiency +Formulary use
LEADER INTERVIEWS
Nationally Prominent
Response rate 5/5
Financial +Reputational
Achievement + Improvement + Risk Adjustment + Tournament Clinical processes
State Medicaid
4 / 11
Financial
Achievement + Improvement
Commercial Health Plans
6 / 10
Financial +Reputational
Achievement + Improvement
Incentive type
Incentive trigger(s)
Performance measures
Clinical processes +Clinical access + Patient satisfaction
Clinical processes +Outcomes +Clinical access +Patient satisfaction +Efficiency +Formulary use
LEADER INTERVIEWS
Nationally Prominent
Response rate 5/5
Financial +Reputational
Achievement + Improvement + Risk Adjustment + Tournament Clinical processes + Patient satisfaction + Efficiency
State Medicaid
4 / 11
Financial
Achievement + Improvement
Commercial Health Plans
6 / 10
Financial +Reputational
Achievement + Improvement
Incentive type
Incentive trigger(s)
Performance measures
Clinical processes +Clinical access + Patient satisfaction
Clinical processes +Outcomes +Clinical access +Patient satisfaction +Efficiency +Formulary use
Will your program impact disparities?
4 “YES”:
"We have an internal disparities working group. We're in the process of looking at demonstrations to see if there are lessons learned. Hopefully disparities will be avoided if we focus on 'gaps' as opposed to 'absolute thresholds'." -Leader of Federal Program #1
"We‟re trying to do some things with disparities and the [health] plans. We've had some people in to talk about this. …We have shared race/ethnicity information with plans when we wouldn't previously have. We have also looked at providers, plans, regions and neighborhoods -- it's another way of looking at the data." -Leader of Medicaid Program #2
Will your program impact disparities?
4 “YES”:
"We have an internal disparities working group. We're in the process of looking at demonstrations to see if there are lessons learned. Hopefully disparities will be avoided if we focus on 'gaps' as opposed to 'absolute thresholds'." -Leader of Federal Program #1 "We‟re trying to do some things with disparities and the [health] plans. We've had some people in to talk about this. …We have shared race/ethnicity information with plans when we wouldn't previously have. We have also looked at providers, plans, regions and neighborhoods -- it's another way of looking at the data." -Leader of Medicaid Program #2
Will your program impact disparities?
4 “YES”:
"We have an internal disparities working group. We're in the process of looking at demonstrations to see if there are lessons learned. Hopefully disparities will be avoided if we focus on 'gaps' as opposed to 'absolute thresholds'." -Leader of Federal Program #1 "We‟re trying to do some things with disparities and the [health] plans. We've had some people in to talk about this. …We have shared race/ethnicity information with plans when we wouldn't previously have. We have also looked at providers, plans, regions and neighborhoods -- it's another way of looking at the data." -Leader of Medicaid Program #2
Will your program identify minority sub-groups?
7 / 15 Measure race/ethnicity
4 / 15
Identify sub-groups in need more tailored programs
“We know that different areas within the State have quite striking differences in care regionally; we know where the less resourced minority areas are. …We‟re thinking of linking up [poorer performing groups] with better performing groups, maybe providing some infusion of resources for groups that wouldn‟t otherwise get „it‟ about best practices.”
- Leader of Commercial Program #5
Will your program identify minority sub-groups?
7 / 15 Measure race/ethnicity
4 / 15
Identify sub-groups in need more tailored programs
“We know that different areas within the State have quite striking differences in care regionally; we know where the less resourced minority areas are. …We‟re thinking of linking up [poorer performing groups] with better performing groups, maybe providing some infusion of resources for groups that wouldn‟t otherwise get „it‟ about best practices.”
- Leader of Commercial Program #5
Will your program identify minority sub-groups?
7 / 15 Measure race/ethnicity
4 / 15
Identify sub-groups in need more tailored programs
“We know that different areas within the State have quite striking differences in care regionally; we know where the less resourced minority areas are. …We‟re thinking of linking up [poorer performing groups] with better performing groups, maybe providing some infusion of resources for groups that wouldn‟t otherwise get „it‟ about best practices.”
- Leader of Commercial Program #5
Will your program identify minority sub-groups?
7 / 15 Measure race/ethnicity
4 / 15
Identify sub-groups in need more tailored programs
“We know that different areas within the State have quite striking differences in care regionally; we know where the less resourced minority areas are. …We‟re thinking of linking up [poorer performing groups] with better performing groups, maybe providing some infusion of resources for groups that wouldn‟t otherwise get „it‟ about best practices.”
- Leader of Commercial Program #5
Program will Payor
Widen resource gap between ‘rich’ & ‘poor’ Cause cherrypicking
Federal government #1
Federal government #2*
•
•
•
State Medicaid #1
State Medicaid #2 State Medicaid #3 State Medicaid #4
Commercial Insurer #1
Commercial Insurer #2 Commercial Insurer #3
•
• •
•
•
Commercial Insurer #4
Commercial Insurer #5 Commercial Insurer #6 Private Coalition #1 Private Coalition #2
•
• • •
•
• • •
Private Coalition #3
•
•
Program will Payor
Widen resource gap between ‘rich’ & ‘poor’ Cause cherrypicking
Federal government #1
Federal government #2*
•
•
•
State Medicaid #1
State Medicaid #2 State Medicaid #3 State Medicaid #4
Commercial Insurer #1
Commercial Insurer #2 Commercial Insurer #3
•
• •
•
•
Commercial Insurer #4
Commercial Insurer #5 Commercial Insurer #6 Private Coalition #1 Private Coalition #2
•
• • •
•
• • •
Private Coalition #3
•
•
Program will Payor
Widen resource gap between ‘rich’ & ‘poor’ Cause cherrypicking
Federal government #1
Federal government #2*
•
•
•
State Medicaid #1
State Medicaid #2 State Medicaid #3 State Medicaid #4
Commercial Insurer #1
Commercial Insurer #2 Commercial Insurer #3
•
• •
•
•
Commercial Insurer #4
Commercial Insurer #5 Commercial Insurer #6 Private Coalition #1 Private Coalition #2
•
• • •
•
• • •
Private Coalition #3
•
•
Program will Payor
Widen resource gap between ‘rich’ & ‘poor’ Cause cherrypicking
Federal government #1
Federal government #2*
•
•
•
State Medicaid #1
State Medicaid #2 State Medicaid #3 State Medicaid #4
Commercial Insurer #1
Commercial Insurer #2 Commercial Insurer #3
•
• •
•
•
Commercial Insurer #4
Commercial Insurer #5 Commercial Insurer #6 Private Coalition #1 Private Coalition #2
•
• • •
•
• • •
Private Coalition #3
•
•
Program will Payor
Widen resource gap between ‘rich’ & ‘poor’ Cause cherrypicking
Federal government #1
Federal government #2*
•
•
•
State Medicaid #1
State Medicaid #2 State Medicaid #3 State Medicaid #4
Commercial Insurer #1
Commercial Insurer #2 Commercial Insurer #3
•
• •
•
•
Commercial Insurer #4
Commercial Insurer #5 Commercial Insurer #6 Private Coalition #1 Private Coalition #2
•
• • •
•
• • •
Private Coalition #3
•
•
RECOMMENDATIONS
1. Collect race and ethnicity data.
2. Emphasize conditions of higher prevalence in minorities.
3. Encourage nationally-prominent organizations to establish „disparity‟ guidelines and/or measures. 4. Reward improvement.
LIMITATIONS
1. Perception-based inquiry.
2. Sample bias.
2. Response bias.
3. Social desirability bias.
SUMMARY
1. Performance incentives can impact disparities.
2. Performance incentive program leaders:
– – – – Do not have disparities in mind Are not necessarily designing programs to narrow disparities Have features that can widen disparities Recommend collecting race/ethnicity as a starting point
SYSTEMATIC REVIEW
MEDLINE® MeSH and non-MeSH search terms for:
Articles related to “performance incentive programs”
N = 41,974
Excluded articles not related to “race”
N = 41,438
Articles related to “performance incentive programs” and “race”
N = 536
Excluded articles: not fitting our definition of “performance incentive program” that were non-empirical, non-English, or not conducted in the United States.
Empirical evaluations
N=1
N = 535