Cultural Competency Training and Performance Reports to Improve Diabetes

advertisement
Cultural Competency Training and
Performance Reports to Improve Diabetes
Care for Black Patients
Thomas D. Sequist, MD MPH
Garret M. Fitzmaurice, ScD
Richard Marshall, MD
Shimon Shaykevich, MS
Amy Marston, BA
Dana G. Safran, ScD
John Z. Ayanian, MD MPP
Harvard Vanguard Medical Associates
Division of General Medicine, Brigham and Women’s Hospital
Department of Health Care Policy, Harvard Medical School
Funding Acknowledgement
Finding Answers: Disparities Research for Change
A National Program of the
Robert Wood Johnson Foundation
Background
• Racial disparities in diabetes care
– Persistence of disparities in outcomes
• Role of primary care clinicians
– Direct interaction with patient and system
– Influence of patient race on clinical decisions
– Limited physician awareness of racial disparities
Study Question
Can cultural competency training and performance
feedback significantly impact:
1) Clinician awareness of racial disparities
2) Achievement of clinical control targets for
black patients with diabetes
Study Setting
• Multispecialty group practice
• Integrated electronic health record
• Team-based primary care
– 2 to 3 physicians
– Nurse practitioner or physician assistant
– NP/PA central role in organizing diabetes care
Randomization Scheme
31 primary care teams
91 physicians
33 NP/PA
7,557 patients (≥ 18 years old)
Intervention
15 primary care teams
46 physicians
16 NP/PA
3,784 patients
Control
16 primary care teams
45 physicians
17 NP/PA
3,773 patients
Intervention Components
• Cultural competency training
– 1 to 2 day course (89% team attendance)
– Monthly educational materials
• Clinical performance feedback
– Monthly distribution
– Race-stratified (HbA1c<7%, LDL < 100, BP < 130/80)
– Provider level (benchmarked to practice)
Patient Survey Feedback
Sample Performance Report
Study Outcomes
• Clinician awareness of racial disparities
– “How often do you feel quality of care varies by race?”
• Across all health centers
• Within local health center
• Among patients personally treated
• Achieving clinical control among black patients
– HbA1c < 7%
– LDL cholesterol < 100 mg/dL
– BP < 130/80 mmHg
• Assessed at baseline and 12 months
Analytic Plan
• Multivariable logistic regression
• Clustering at level of primary care team
• Adjusted for baseline clinician awareness or
baseline clinical performance
Baseline Patient Characteristics
Intervention
(n = 3,784)
Control
(n = 3,773)
p value
62.5
62.3
0.24
Male, %
50
48
0.41
Insurance
Commercial
Medicare
Medicaid
Uninsured
58
35
4
3
57
36
5
3
0.66
Clinical targets, %
HbA1c<7%
LDL < 100 mg/dL
BP < 130/80 mmHg
45
50
26
43
52
31
0.35
0.58
0.04
Mean age, years
Baseline Racial Disparities
White
(n=4,858)
Black
(n=2,699)
p
value
Process measures, %
Annual HbA1c test
Annual LDL test
87
83
89
83
0.14
0.99
Outcomes measures, %
HbA1c < 7%
LDL < 100 mg/dL
BP < 130/80 mmHg
46
55
32
40
43
24
<0.001
<0.001
<0.001
Impact on Clinician Awareness
% Very or Somewhat Often
Do racial disparities in diabetes care exist in….
P=0.04
100
P=0.06
82
80
60
P=0.45
70
59
63
51
43
40
20
0
All health centers
Your health
center
Intervention
Your patients
Control
Clinician Support for Intervention
How effective are these strategies to reduce racial disparities?
% 'Very Effective'
50
P=0.01
40
P=0.02
34
33
30
20
17
11
10
0
Cultural Competency
Training
Intervention
Performance Feedback
Control
Diabetes Care Among Black Patients
Intervention,
%
HbA1c < 7%
49
LDL < 100 mg/dL
50
BP < 130/80 mmHg
23
Control,
%
47
50
24
P
value
0.36
0.68
0.73
Limitations
• Advanced care setting
– Electronic health record
– Team based care
• Only 50% of teams had both MD and
NP/PA attend cultural competency training
– NP/PA central role in diabetes care
Conclusions
• Intervention increased clinician awareness
of disparities
• Decreased clinician support for utility of
intervention
• No impact on clinical outcomes
Implications
• Increased clinician awareness may be
necessary, but not sufficient to improve
care for black patients
• Additional research needed to understand
how to facilitate behavior change for both
providers and patients
N
ov
-0
8
8
08
ar
-
7
Ju
l-0
M
N
ov
-0
7
07
ar
-
6
Ju
l-0
M
N
ov
-0
% Patients with Race/ Ethnicity Available
Race Collection Progress
100
80
60
40
20
0
Clinical Data Feedback
The “Not Me” Phenomenon: Primary Care
% Very or Somewhat Often
Do racial disparities in diabetes care exist in….
100
90
80
70
60
50
40
30
20
10
0
US Health System
HVMA
Your health center
Your patients
Sequist et al; JGIM 2008
Download