Do Physicians' Implicit Views of African Americans Affect Clinical Decision Making?

M. Norman Oliver, MD, MA; Kristen M. Wells, MPH, PhD; Jennifer A. Joy-Gaba, PhD; Carlee Beth Hawkins, MA; Brian A. Nosek, PhD

Disclosures

J Am Board Fam Med. 2014;27(2):177-188.

Conclusion

The conditions predicting when implicit biases will predict behavior or not are not yet fully understood. The lack of a significant correlation between our participants' implicit and explicit biases and their referral for TKR provides speculative opportunity to identify occasions where implicit biases do not affect clinical decision making or show their effect through other, less direct mechanisms. Studies of the possible role of implicit racial biases on clinical decision making should continue, and tools should be developed to help clinicians mitigate the effect of such biases on clinical practice.

1 2 3 4 5

Next Section

Funding
This work was completed under ARRA Stimulus Funding award no. ZC10017.

Conflict of interest
The senior author (BAN) is an officer of Project Implicit Inc., a nonprofit organization that includes in its mission "To develop and deliver methods for investigating and applying phenomena of implicit social cognition, including especially phenomena of implicit bias based on age, race, gender or other factors."

Table 1. Participants' Demographic Information
Characteristics	Patients, n (%)
Sex
Male	256 (47)
Female	272 (50)
Race/ethnicity
European American/white	371 (70)
African American/black	34 (6)
Asian/Pacific Islander	69 (13)
Hispanic/Latino	27 (5)
Mixed	14 (3)
Other/unknown	13 (2)
American Indian/Alaskan	2 (<1)
Specialty
Family medicine	163 (58)
Internal medicine	116 (42)
African American patients by clinic (%)
<10	124 (31)
10–19	51 (13)
20–29	65 (16)
30–39	60 (15)
40–49	40 (10)
≥50	64 (16)
Identification with racial/ethnic identity
Little identification	132 (31)
Moderate identification	209 (49)
Strong identification	86 (20)
Socioeconomic level growing up
Lower	27 (6)
Lower-middle	65 (15)
Middle	158 (37)
Upper-middle	141 (33)
Upper	35 (8)
Mean age (SD)	39.16 (11.52)
Mean years in practice (SD)	12.05 (10.26)

Percentages do not add up to 100 because of missing values and rounding.
SD, standard deviation.

Table 2. Implicit Association Test (IAT) Scores by Patient Race and Order of IAT
	IAT Completed First				Vignette Completed First
	Black Patients		White Patients		Black Patients		White Patients
Mean (SD)	No.	Mean (SD)	No.	Mean (SD)	No.	Mean (SD)	No.
Medical cooperativeness	0.19 (0.39)	112	0.30 (0.43)	107	0.21 (0.43)	104	0.23 (0.40)	103
Racial preference	0.50 (0.24)	19	0.50 (0.33)	14	0.38 (0.38)	27	0.36 (0.42)	12

Positive scores indicate stronger associations with white people and medically cooperative and white people and good.
SD, standard deviation.

Table 3. Predicting Total Knee Replacement (TKR) Recommendation from Patient Race, Implicit Association Test (IAT) Order, and Implicit Racial Attitude and Stereotype Moderators
	χ²	P Value
Cooperativeness IAT (n = 417)
Patient race main effect	0.15	.700
Cooperativeness IAT main effect	0.24	.624
Order main effect	2.30	.129
Interactions
Patient race + cooperativeness IAT	0.50	.479
Patient race + order	0.20	.652
Order + cooperativeness IAT	0.00	.983
Patient race + cooperativeness IAT + order	0.55	.458
Race preference IAT (n=72)
Patient race main effect	0.63	.428
Race preference IAT main effect	1.75	.185
Order main effect	3.11	.078
Interactions
Patient race + race preference IAT	0.21	.648
Patient race + order	0.07	.796
Order + race preference IAT	2.26	.133
Patient race + race preference IAT + order	0.32	.570

Table 4. Participants' Explicit Attitudes Toward Patients
Attitudes	No.	Overall	Black Patients	White Patients
Explicit racial preference*	427	0.28 (0.78)	0.28 (0.69)	0.28 (0.87)
Warmth toward black people^†	427	6.99 (1.80)	6.90 (1.73)	7.08 (1.88)
Warmth toward white people^†	427	7.31 (1.69)	7.17 (1.64)	7.45 (1.74)
Medically cooperative*	420	0.21 (0.75)	0.15 (0.70)	0.28 (0.79)
Perception of black patients' medical cooperativeness^†	425	6.22 (1.42)	6.21 (1.42)	6.24 (1.42)
Perception of white patients' medical cooperativeness^†	425	6.55 (1.38)	6.52 (1.38)	6.58 (1.42)
Aggressive care*	427	0.04 (0.41)	0.04 (0.38)	0.03 (0.44)
Subconscious biases affect my decisions^†	428	3.64 (1.12)	3.51 (1.17)	3.78 (1.05)
Treatment due to black patients' preferences^†	215	3.26 (1.25)	3.19 (1.31)	3.33 (1.20)
Treatment due to white patients' preferences^†	211	3.14 (1.26)	3.11 (1.30)	3.18 (1.21)

Data are mean (standard deviation) unless otherwise indicated.
*This scale is centered at zero, with positive values indicating more explicit preference for white people compared to black people, believing white people are more medically cooperative than black people, and providing more aggressive care for white people over black people.
^†Higher values indicate more warmth and more agreement that difference in treatment is due to patient preference.

Table 5. Predicting Total Knee Replacement Recommendation from Patient Race, Order, and Explicit Racial Attitude Moderators
	χ²	P Value
Patient race main effect	0.04	0.85
Aggregated explicit attitudes main effect	0.09	0.77
Order main effect	2.50	0.11
Interactions
Patient race + aggregated explicit attitudes	0.00	0.95
Patient race + order	0.17	0.68
Order + aggregated explicit attitudes	1.68	0.20
Patient race + order + aggregated explicit attitudes	0.22	0.64

Table 6. Predicting Total Knee Replacement Recommendation from Patient Race, Order, and Aggregated Patient Belief Moderators
	χ²	P Value
Patient race main effect	1.07	.30
Aggregated patient beliefs main effect	6.01	.01
Order main effect	0.09	.77
Interactions
Patient race + aggregated patient beliefs	1.20	.27
Patient race + order	3.51	.06
Order + aggregated patient beliefs	0.76	.38
Patient race + order + aggregated patient beliefs	4.22	.04

Table 7. Participants' Beliefs About the Study
Beliefs	No.	Overall	Black Patients	White Patients
Subconscious biases influence my decisions	380	0.72 (1.07)	0.62 (1.13)	0.83 (1.01)
Learning about biases can improve patient care	378	1.23 (0.89)	1.15 (0.95)	1.32 (0.83)
IAT experience is worthwhile	381	0.99 (1.00)	0.91 (1.03)	1.08 (0.96)
Knowledge about what the study	379	−1.37 (1.47)	−1.43 (1.48)	−1.31 (1.46)
The IAT results reflect my attitudes	381	0.02 (1.07)	−1.98 (1.10)	0.05 (1.04)
Would not have participated had I known about the study's contents	379	−2.45 (0.87)	−2.44 (0.88)	−2.46 (0.85)

Data are mean (standard deviation) unless otherwise indicated. Higher values indicate stronger agreement.
IAT, implicit association test.

Appendix 2

Overall Correlations

Correlations	Race IAT	Medical Cooperativeness IAT	Aggregated Beliefs about Patient	Aggregated Explicit Beliefs	TKR Recommendation
Overall
Race
IAT	1.00	0.33*	0.03	−0.07	−0.07
No.	72	67	71	64	72
Medical cooperativeness
IAT	1.00	0.00	0.09^†	−0.02
No.	426	414	401	417
Aggregated beliefs about patient
IAT		1.00	−0.08	−0.13*
No.		455	412	451
Aggregated explicit beliefs
IAT			1.00	0.03
No.				416	413
Intervention condition
Race
IAT	1.00	0.18	0.25	−0.34^†	−0.29^†
No.	33	32	33	32	33
Medical cooperativeness
IAT		1.00	−0.06	0.12^†	−0.03
No.		219	209	202	211
Aggregated beliefs about patient
IAT			1.00	−0.06	−0.10
No.			215	206	213
Aggregated explicit beliefs
IAT				1.00	0.09
No.				208	206
Control condition
Race
IAT	1.00	0.42*	−0.15	0.19	0.04
No.	39	35	38	32	39
Medical cooperativeness
IAT		1.00	0.09	0.04	−0.02
No.		207	205	199	206
Aggregated beliefs about patient
IAT			1.00	−0.11^†	−0.16*
No.			240	206	238
Aggregated explicit beliefs
IAT				1.00	−0.04
No.				208	207
When participants view a white patient
Race
IAT	1.00	0.19	−0.13	0.05	−0.08
No.	26	41	46	40	26
Medical cooperativeness
IAT		1.00	−0.01	0.09	−0.05
No.		210	212	205	204
Aggregated beliefs about patient
IAT			1.00	−0.11^†	−0.19*
No.			215	200	213
Aggregated explicit beliefs
IAT				1.00	0.03
No.				203	201
When participants view a black patient
Race
IAT	1.00	0.19	−0.13	0.05	−0.06
No.	46	41	46	40	46
Medical cooperativeness
IAT		1.00	−0.01	0.09	0.01
No.		216	212	205	213
Aggregated beliefs about patient
IAT			1.00	−0.04	−0.07
No.			240	212	238
Aggregated explicit beliefs
IAT				1.00	0.02
No.				213	212
*P ≤ .05. ^† P ≤ .10. IAT, implicit association test; TKR, total knee replacement.

Comments

processing....

Do Physicians' Implicit Views of African Americans Affect Clinical Decision Making?

Conclusion

Comments

Most Popular Articles

What to Read Next on Medscape

About

Membership

Apps

WebMD Network

Editions

Do Physicians' Implicit Views of African Americans Affect Clinical Decision Making?

Conclusion

Tables

Tables

Tables

Tables

Tables

Tables

Tables

References

Authors and Disclosures

Authors and Disclosures

Sidebar

Appendix 1

Clinical Vignette

Sidebar

Appendix 2

Overall Correlations

Comments

Most Popular Articles

What to Read Next on Medscape

Recommended Reading

Drugs Related to null

Most Popular Articles

Email This

Find Us On

About

Membership

Apps

WebMD Network

Editions