Transcript Disparities in Patient Ratings of Care: Why Race Matters

Overcoming Healthcare Disparities:
The Role of Patient-Centered Care
Lisa A. Cooper, MD, MPH
Professor of Medicine, Epidemiology,
and Health Policy and Management
Johns Hopkins Medical Institutions
Racial and ethnic disparities in
health are documented
• Life expectancy at birth – Blacks vs. Whites,10
year gap for men, 5 year gap for women
• Infant mortality rate – Blacks and Native
Americans vs. Whites: twice as high
• Death rate – Blacks vs. whites: greater for cancer,
diabetes, heart disease, HIV/AIDS, homicide;
Hispanics vs. Whites: greater for diabetes
• Morbidity – most ethnic minorities vs. Whites:
higher for cancer, diabetes, hypertension, obesity,
HIV/AIDS, tuberculosis, hepatitis
Potential Reasons for
Disparities in Health
Race
• Biologic factors
• Socioeconomic status
• Environmental factors
• Discrimination/Stress
• Cultural factors
• Health risk behavior
• Access to healthcare
• Quality of healthcare
Health
Access to Health Care
for Racial and Ethnic Groups
Barriers
Personal/Family
 acceptability
 cultural
 language/literacy
 attitudes, beliefs
 preferences
 involvement in care
 health behavior
 education/income
Structural
 availability
 appointments
 how organized
 transportation
Financial
 insurance coverage
 reimbursement
levels
 public support
Health Care Processes
Use of Services
 Visits
 primary care
 specialty
 emergency
Procedures
 preventive
 diagnostic
 therapeutic
Mediators
 Quality of providers
 cultural competence
 communication skills
 medical knowledge
 technical skills
 bias/stereotyping
Outcomes
 Health Status
 mortality
 morbidity
 well-being
 functioning
Equity of Services
Appropriateness of care
Efficacy of treatment
Patient adherence
Patient Views of Care
 experiences
 satisfaction
 effective
partnership
Modified From Access to Health Care in America (1993, Millman M, ed).
Cooper LA, Hill MN, and Powe NR. JGIM 2002; 477-486
Unequal Treatment: A Report
of the Institute of Medicine*
Whites
Ethnic minorities
Quality of Care
Difference
Clinical
Appropriateness and
Need, Patient
Preferences
Systems, Legal,
Regulatory
Disparity
Discrimination, Bias,
Clinical uncertainty
*National Academy Press, Washington DC, 2003
Racial and ethnic healthcare
disparities are pervasive
• Conditions: cancer, diabetes, heart disease,
kidney disease, HIV/AIDS, mental health,
respiratory diseases (e.g., asthma)
• Populations: young, old, urban, rural, men,
women, immigrants, non-immigrants
• Settings: primary care, emergency care,
hospital care, specialty care, nursing homes
• Levels and types of care: preventive, acute
care, chronic disease management
• Dimensions of healthcare quality: timeliness,
effectiveness, safety, patient-centeredness
Dimensions of
Health Care Quality
• Structure: “characteristics of the settings in
which care is delivered…”
• Process: “ …the care itself, or activities
undertaken by the health care system…”
• Outcome: “the effect of care on the health
and welfare of individuals or populations…”
Donabedian A. JAMA 1988;260:1743-1748
Process
interpersonal,
technical care, or
appropriateness of care
Structure
race concordance,
staff expertise, availability,
organization, coordination,
Outcome
patient ratings of care,
equity of services
death, complications
Examples of Structure, Process, and Outcome Variables
Disparities in Process of Care
• Technical care – many studies
– Ethnic minorities receive fewer preventive services,
diagnostic and therapeutic tests and procedures, and
fewer appropriate medications
• Patient-centered or interpersonal care – fewer
studies
– Ethnic minority patients rate interpersonal care from
physicians more negatively than whites
– It is unclear whether this is due to ethnic/racial
discordance, poor communication, bias, or mistrust
• Few disparities studies make links between
structure, processes, and outcomes
Process
Interpersonal or
Patient-centered Care
Structure
Race
Concordance
Outcome
Patient ratings
of PDM
* physicians’ participatory decision-making style
Concordance
• What is it?
– a structural dimension of health care quality
– shared identities between patients and health
professionals
• Why do we care?
– Because most ethnic minorities see physicians who
differ from them in key social characteristics
• Patients and physicians may be concordant in:
– Visible demographic factors such as race/ethnicity,
gender, age, education, social class, language
– Less visible factors such as beliefs, values,
expectations, preferred roles
Patient-centered Care
“Providing care that is respectful of and
responsive to individual patient preferences,
needs, and values, and ensuring that patient
values guide all clinical decisions…”
*Institute of Medicine, “Crossing the Quality Chasm, 2001
Race, Gender, and Partnership in
the Patient-Physician
Relationship
• Design: Cross-sectional telephone survey
• Subjects: 1816 adults (784 W, 814 AA, 218 Other)
who had seen their MD (n=65) within the past 2 weeks
• Setting: 32 primary care practices, large network style
managed care organization in Washington D.C. area
• Predictor variables: race and gender concordant or
discordant status in patient-physician relationship
• Main Outcome: patients’ ratings of their MD’s
participatory decision-making (PDM) style
Cooper-Patrick L et al, JAMA 1999;282:583-589
Measurement of
Physicians’ Participatory
Decision-Making Style*
Patient is asked:
• If there were a choice between treatments, how
often would this doctor ask you to help make the
decision?
• How often does this doctor make an effort to give
you some control over your treatment?
• How often does this doctor ask you to take some
of the responsibility for your treatment?
*Kaplan SH et al, Medical Care 1995;33:1176-1187
Each item contributes 33.3 points. Maximum score is 100 points.
Ethnic minorities rate their visits
with physicians as less participatory
78
77.1
PDM score
77
P=0.007
76
P=0.05
75
74
73.9
73.8
Whites
Blacks
Others
73
72
PDM scores range from 0-100. A higher score means visit is more participatory.
Cooper-Patrick L , JAMA 1999;282:583-589
Mean PDM Style Score
Patients in race-concordant
relationships rate their physicians as
more participatory
64
63.3
63
62
P=0.02
P-value NS
61.7
61.1
61
60
59
58.5
concordant
discordant
58
57
56
Race
Gender
Adjusted for patients’ age, gender, education, marital status, health status, length of the
patient-physician relationship, physician gender (race concordant analysis) and physician
race (gender concordance analysis). Cooper-Patrick L, JAMA 1999;282:583-589
Process
Interpersonal or
Patient-centered Care:
Communication
Structure
Outcome
Race Concordance
Patient ratings
of PDM* and
Satisfaction
* physicians’ participatory decision-making style
Patient-physician communication
is related to important outcomes
• Patient adherence
• Patient satisfaction
• Clinical outcomes
Glycemic
control
BP control
Pain reduction
Depression resolution
Roter 1988, Greenfield 1988, Kaplan 1989, Stewart 1995, Kaplan 1995
Patient-Centered Communication,
Ratings of Care and Concordance
of Patient and Physician Race
• Design: cross-sectional study using pre-visit and
post-visit surveys and audiotape analysis
• Participants: 458 African American and white
adult patients receiving care from 61 PCPs
• Setting: urban primary care practices serving
managed care and fee-for-service patients
• Patient recruitment: ~10 patients per MD
recruited consecutively from waiting rooms
Cooper LA, Roter DL, Johnson RL, Ford DE, Steinwachs DM, Powe NR.
Ann Intern Med 2003;139:907-915
Functions of Clinical
Communication
• Data-gathering
• Educating and counseling
patients
• Relationship-building
• Partnering with patients to
negotiate diagnostic and
treatment decisions
Lipkin, Putnam, & Lazare, 1995
Measuring Clinical
Communication*
• Content (questions and information-giving)
– Biomedical talk
– Psychosocial talk
• Affect
– Emotional Talk
- Negative talk
– Positive talk
- Social talk
• Process
– Orientation (directions or instructions)
– Facilitation (includes partnership-building)
*Roter Interaction Analysis System (RIAS)
Roter D, Larson S. Patient Educ Couns 2002;46:243-51
Examples from RIAS
Communication Categories
• Biomedical talk
“Your blood pressure is 100 over 70.”
“I was in the hospital last year for ulcers.”
• Psychosocial talk
“You really need to get out and meet more people.”
“I guess every marriage has its ups and downs.”
• Emotional talk
“This must be very hard for you.”
“I hope you’ll be feeling better soon.”
• Partnership-building
“Do you follow me?” “How does that sound to you?”
Measuring Emotional Tone of
Visits using the RIAS
Coders are asked to rate overall emotional
tone of the visit for patients and
physicians:
• Physician positive affect = (assertiveness +
interest + responsiveness + empathy) - hurried
• Patient positive affect = (assertiveness +
interest + friendliness + responsiveness +
empathy)
The Patient-Centered
Clinical Interview
• Visit duration is longer
• Speech speed is lower
• Physicians are less verbally dominant
• doctor talk to patient talk ratio is close to 1
• Patient-centeredness ratio is high: more
psychosocial, emotional, and partnership
talk than biomedical talk
• More positive emotional tone
Physicians communicate differently
with black and white patients
Communication measure
Whites
n=202
Blacks
n=256
p-value*
Physician verbal dominance
1.50
1.73
<0.01
Physician positive affect**
14.1
13.2
0.02
Patient positive affect**
16.7
15.8
<0.01
Patient-centeredness ratio
1.91
1.58
0.08
Adjusted for: patient age, gender, education level, and self-rated health status; and physician
gender, race, time since completing training, and report of how well he/she knows each patient.
*p-value from linear regression with GEE.** Patient and physician affect scores are derived
from audiotape coders’ impressions of the overall emotional tone of the medical visit.
Johnson RL, Roter DL, Powe NR, Cooper LA. Am J Public Health 2004;94:2084-2090.
Race-concordant visits are longer
with slower speech and more
positive patient emotional tone
20
P=0.01
17.5
P=0.05
19.2
18.2
P=0.03
16.4
15.4
15
15.8
P=0.19
13.2
12.7
concordant
discordant
10
Physician
Patient
Speech
Visit
positive
positive
speed per
duration,
affect
affect
minute
minutes
Adjusted for patient age, race, gender, and health status, physician gender & yrs in practice
Cooper LA et al, Ann Intern Med 2003;139:907-915
Patients in Race-Concordant Relationships
Rate Their Physicians Better
Mean Score/Probability
concordant
80
70
60
76.1
P=.01
68
73
discordant
P<.01
51
73
P=.03
57
50
40
30
20
10
0
Participatory
Decision-making
Overall Satisfaction
Recommend MD to a
friend
Analyses adjusted for patient gender, race, age, and health status, physician gender, years in
practice, and patient-centered communication.
Cooper LA et al, Ann Intern Med 2003;139:907-915
Summary
• African American patients experience visits in
which physicians are less patient-centered
• African Americans in race-discordant
relationships with their physicians experience:
– Lower levels of satisfaction
– Less participation in medical decisions
– Shorter visits with less positive emotional tone
• Differences in communication do not explain
why patients in race-discordant relationships
rate their care worse
• Other factors, such as physician and patient
attitudes, may play a role
Process
Interpersonal Care:
Bias
Structure
Outcome
Race Concordance
Patient ratings
of care
Explicit vs. Implicit Bias
• Explicit (conscious) bias: attitudes and
beliefs we recognize and know we have
• Implicit (unconscious) bias: attitudes that
are unavailable to introspection and
outside of conscious cognition
– Can unintentionally affect behavior
– Are better predictors of behavior than self
reported measures of prejudice,
stereotyping and discrimination
Clinician Racial Bias,
Communication Behaviors and
Patient Experiences of Care
• Design: Cross-sectional study
• Participants: 39 primary care clinicians and
213 of their African American patients
• Setting: 24 urban, community-based primary
care practices in Baltimore, Maryland and
Wilmington, Delaware
• Main predictor variables: Clinicians’ implicit
attitudes about race (race attitude IAT and
patient race/medical compliance IAT)
The Race Implicit Association Test
(http://www.implicit.harvard.edu)
• An indirect measure of an individual’s implicit
(unconscious) attitudes
• Images appear rapidly on computer screen and
subjects respond by sorting pairs of images and
attributes using right and left keys
• Premise: individuals will respond faster to
concepts that are strongly associated compared
to those that have weak associations
• If subjects match white+good/black+bad pairings
faster than black+good/white+bad pairings, then
the race IAT score differs from zero and is
positive – labeled implicit white preference
Greenwald, McGhee, Schwartz, 1998
Implicit preference for whites:
Response to these pairings is faster…
African
American & unpleasant
European
pleasant & American
pain
gentle
death
happy
stink
smile
grief
joy
agony
warmth
filth
pleasure
tragedy
paradise
vomit
rainbow
…than response to these pairings
European
American & unpleasant
African
pleasant & American
pain
gentle
death
happy
stink
smile
grief
joy
agony
warmth
filth
pleasure
tragedy
paradise
vomit
rainbow
Implicit association for European
American and compliant patient
Response to these pairings is faster…
European
&
American
Compliant
Patient
Reluctant
Patient
willing
doubting
cooperative
reluctant
compliant
hesitant
reliable
apathetic
adherent
resistant
helpful
lax
African
& American
…than response to these pairings
European
Reluctant
&
American
Patient
Compliant
Patient
doubting
willing
reluctant
cooperative
hesitant
compliant
apathetic
reliable
resistant
adherent
lax
helpful
African
& American
Methods, continued
• Main outcomes:
– Audiotaped Measures: Clinician and patient
communication behaviors measured by Roter
Interaction Analysis System (RIAS)
– Patient ratings of care: overall satisfaction,
trust in clinician, participation in decisionmaking, and quality of interpersonal care
measured by post-visit survey
• Analysis: determine whether clinicians’
implicit attitudes predict differences in
communication and patient ratings of care*
*Linear and logistic regression with generalized estimating
equations to account for clustering of patients by clinician
Measuring Clinical
Communication*
• Content (questions and information-giving)
– Biomedical talk
– Psychosocial talk
• Affect
– Emotional Talk
- Negative talk
– Positive talk
- Social talk
• Process
– Orientation (directions or instructions)
– Facilitation (includes partnership-building)
*Roter Interaction Analysis System (RIAS)
Roter D, Larson S. Patient Educ Couns 2002;46:243-51
Audiotape Ratings of Clinician
and Patient Emotional Tone
• Clinician behaviors
– Positive affect – average of 6 items each
rated on a 5-point scale: interest, warmth,
engagement, respect, and sympathy
– Negative affect – average of 2 items each
rated on a 5-point scale: dominance and
hurried/rushed
• Patient behaviors
– Positive affect – average of 5 items each
rated on a 5-point scale: interest, warmth,
engagement, sympathy, and respect
Patient Ratings of Clinician
• Overall satisfaction
– Overall, I was satisfied with this visit
– I would recommend this provider to a friend
• Quality of interpersonal care
– My provider has a great deal of respect for me
– My provider likes me
– I like this provider
• Participation in decision-making
– If there were a choice, this provider would ask me to
help make the decision
• Trust in provider
– I trust this provider to act in my best interests
Responses are on 5-point Likert scale from strongly agree to strongly disagree.
Interpersonal Care
Quality Measures
• Patient-centeredness ratio is high: more
psychosocial, emotional, and partnership
talk than biomedical and procedural talk
• Clinicians and patients exhibit more
positive emotional tone and less negative
emotional tone
• Patients report higher levels of trust,
respect, and satisfaction, and participation
in decision-making
Characteristics of Clinicians
Characteristic
Mean age, yrs (SD)
Practice experience, yrs (SD)
Female gender,%
Caucasian, %
African American,%
Asian, %
Liberal political idealogy, %
Internal medicine training, %
Board certified,%
(N=39)
44.1(8.2)
13.5 (7.4)
64
49
21
23
73
77
90
Characteristics of Patients
Characteristic
N=213
Mean age, yrs (SD)
High school graduate, %
Female gender, %
African American,%
Have health insurance,%
Annual income < $35,000, %
Poor/fair self-rated health status
Known by clinician (not first visit)
54.5 (13.3)
81
73
100
91
60
46
90
Clinician Responses to IAT(N=39)
Percent of respondents with each score
Strong preference for Whites
14%
Moderate preference for Whites
26%
Slight preference for Whites
26%
Little to no preference
10%
Slight preference for Blacks
14%
Moderate preference for Blacks
5%
Strong preference for Blacks
5%
66%
The IAT D (difference score) ranges from -2 to +2, with 0 indicating no relative preference
for blacks compared to whites, and positive scores indicating some degree of implicit bias
favoring Whites. [mean score for this sample is +0.24 (.49)]
Implicit Preference for White vs. Black People by
732,881 respondents on Project Implicit websites,
July 2000- May 2006
Percent of Harvard website respondents with each score
Strong preference for Whites
27%
Moderate preference for Whites
27%
Slight preference for Whites
16%
Little to no preference
17%
Slight preference for Blacks
6%
Moderate preference for Blacks
4%
Strong preference for Blacks
2%
70%
Association of Clinician
Implicit Racial Bias with
Communication Behaviors
Communication behavior
Patient-centeredness
β-coefficient
-0.67
P-value
0.29
Clinician positive affect
-0.28
0.21
Clinician negative affect
+0.23
0.03
Patient positive affect
-0.18
0.03
The beta coefficient means for each 1-point increase in the IAT score -indicating more pro-white bias among clinicians – clinician’s negative
affect was higher and African American patients’ positive affect was
lower . Adjusted for patient age, education, health status, clinician’s
gender, race, and the interaction of clinician race with implicit bias.
Association of Clinician
Race/Medical Compliance Bias
with Communication Behaviors
Communication behavior
Patient-centeredness
β-coefficient
-3.12
P-value
0.004
Clinician positive affect
-0.14
0.38
Clinician negative affect
+0.02
0.95
Patient positive affect
-0.04
0.81
The beta coefficient means for each 1-point increase in the IAT score -indicating more pro-white bias among clinicians – the communication in
the visit was less patient-centered. Adjusted for patient age, education,
health status, clinician’s gender, race, and the interaction of clinician
race with implicit bias.
Clinician Racial Bias and
Patient Reports of Care
0.63
I was satisfied with this visit
0.32
I would recommend this doctor
to a friend
0.24
This doctor respects me
0.48
This doctor asks me to help
decide my treatments
0.22
I like this doctor
0.47
I trust this doctor
0
0.5
1.0
1.5
2.0
4.0
6.0
8.0
10.0
Odds Ratio
As the implicit bias score increases the patient has lower odds of strongly agreeing
Clinician Race/ Medical Compliance
Bias and Patient Reports of Care
0.49
I was satisfied with this visit
0.57
I would recommend this doctor
to a friend
0.48
This doctor respects me
0.20
This doctor asks me to help
decide about my treatments
0.89
I like this doctor
0.55
0
I trust this doctor
0.5
1.0
1.5
2.0
4.0
Odds Ratio
6.0
8.0
10.0
As the implicit bias score increases the patient has lower odds of strongly agreeing
Summary
• This is the first study to explore links among
implicit bias, clinician behaviors, and patient
ratings in actual patient encounters
• Primary care clinicians in this sample display
implicit attitudes about race that are similar to
those measured in large samples of society
• Implicit bias favoring whites and the association of
white race with medical compliance predicts:
–
–
–
–
less patient-centered communication
more negative clinician emotional tone
less positive patient emotional tone
poorer ratings of care by African-American patients
Implications
• Research – Examine links among clinician
attitudes, behaviors, and health outcomes
• Health Professional Education - employ patientcentered communication skills programs that
emphasize rapport building and affective
dimensions and enhance awareness of bias
and intercultural skills
• Clinical Practice - implement patient activation
programs; improve scheduling, increase time to
build rapport and develop continuity of care
• Policy - increase numbers of underrepresented
ethnic minorities among health professionals
Minority Health Policy Timeline
1972
Tuskegee
Syphilis Study
becomes public
1970
1980
Health Revitalization
Act of 1993 establishes
the Office of Research
on Minority Health
Minority Health and Health
Disparities Research and
Education Act of 2000
1985
DHHS Heckler
Report
1990
2000
2003 IOM Report “Unequal
Treatment” and first
National Healthcare
Disparities Report published
2008
Evolution of Research
on Health Disparities
1980
1990
2000
Describing the problem
Understanding mechanisms
Designing interventions
Evaluating outcomes
Patient-Physician
Partnership to Improve
HBP Adherence
• Design: Randomized controlled trial
• Population: 50 primary care MDs and 500
patients (60% AA) with high blood pressure
• Setting: 15 urban, community-based clinics in
East and West Baltimore
• Interventions: Communication skills training on
interactive CD-ROM for MDs; Patient activation
by community health worker
• Main Outcomes: patient adherence, BP control
Supported by the National Heart, Lung, and Blood Institute
R01HL69403, 09/01/01-08/31/07
Blacks Receiving Interventions for
Depression and Gaining Empowerment
• Design: Randomized controlled trial
• Population: 30 primary care clinicians and 250
African American patients with depression
• Setting: Urban, community-based clinics in
Delaware and Maryland
• Interventions: standard quality improvement vs.
patient-centered, culturally tailored program
• Main Outcomes: Depression remission,
depression level, guideline-concordant care
Supported by the Agency for Healthcare Research and Quality
R01 HS13645-01, 09/30/03-09/29/09
Funding Sources
• National Heart, Lung, and Blood Institute
– R01HL69403 and K24HL083113
• Agency for Healthcare Research and Quality
– R01HS013645
• National Center for Minority Health and Health
Disparities (P60MD000214)
• Robert Wood Johnson Foundation Amos
Medical Faculty Development Program
• The Commonwealth Fund
• Fetzer Foundation
Acknowledgments
•
•
•
•
•
Debra Roter
Neil Powe
Daniel Ford
Rachel Johnson
Don Steinwachs
•
•
•
•
•
Mary Catherine Beach
Thomas Inui
Anthony Greenwald
Janice Sabin
Kathryn Carson