literature review
pradeep333
Jones.DDissertation_Final04.29.21.pdf
Deconstructing Unconscious Bias in the Health Care Workforce: An Iterative Mixed Methods Approach
© 2021
Danielle D. Jones
M.P.H., University of West Florida, 2010
B.S., University of Missouri Kansas City, 2002
Submitted to the graduate degree program in Health Policy and Management and the
Graduate Faculty of the University of Kansas in partial fulfillment of the requirements
for the degree of Doctor of Philosophy.
_______________________________________ Dissertation Committee Chair: Tami Gurley, PhD
_______________________________________
Joanna V. Brooks, PhD, MBE
_______________________________________ Megha Ramaswamy, PhD, MPH
_______________________________________
Christopher Crenner, MD, PhD
_______________________________________ Jill Peltzer, PhD, APRN-CNS
Date Defended: 03/29/2021
ii
The dissertation committee for Danielle D. Jones certifies that
this is the approved version of the following dissertation:
Deconstructing Unconscious Bias in the Health Care Workforce: An Iterative Mixed
Methods Approach
____________________________________ Dissertation Committee Chair: Tami Gurley, PhD
____________________________________
Graduate Director: Tami Gurley, PhD
Date Approved: 04/13/2021
iii
Abstract
The prevalence of unconscious bias within the healthcare workforce is not well
understood. Likewise, not much is known about the potential impacts of unconscious
bias training interventions on the healthcare workforce as they have not been included
in studies evaluating effectiveness. This constrains any ability to evaluate the potential
for unconscious bias training as a means to reduce patient healthcare disparities. This
dissertation uses an iterative mixed methods approach to examine the prevalence of
unconscious bias, factors associated with individual mitigation activities, and the impact
on the healthcare workforce. Results demonstrate that the unconscious biases of
healthcare workers differ significantly from those of the general population and are
highly variable across geographic regions and provider types. Likewise, there is some
evidence to indicate that factors beyond that of the individual (i.e. type of practice and
community) may potentially influence physicians’ decisions to participate in unconscious
bias mitigation activities. Lastly, physicians have many reasons for wanting to address
unconscious bias, such as for their own personal and/or professional development.
However, there is a consensus that greater accountability on the part of organizations is
needed to address the upstream systemic issues that contribute to the formation and or
maintenance of unconscious bias.
iv
Acknowledgements
I’d like to acknowledge my three daughters, Daijah, Deanna and Danica Jones
for their patience and support during this process. Through all the long days and long
nights, they exhibited a great deal of responsibility, maturity, and independence by not
only tending to their own needs but mine as well. I’d like to acknowledge my parents,
Cecil, and Denise Morrison for providing consistent encouragement and reminding me
it’s not how you start but how you finish. I’d like to thank the members of my cohort Drs.
Joe Pacheco and Allen Solenberg for their comradery, support, and humor. Sincere
appreciation for my Chair Dr. Tami Gurley for her persistent and diligent commitment to
seeing me through the process and to Vice Chancellor for Diversity, Equity, and
Inclusion Dr. Jerrihlyn L. McGee for providing support and access for my research.
Also, I’d like to thank the members of my committee for seeing the significance of this
work as a contribution to the field. Lastly, I’d like to thank the staff and members of the
American Academy of Family Physicians for their confidence and trust which allowed for
the space and opportunity to lead them in this work. All individuals in a position to
control this content have indicated there are no relevant relationships, financial or
otherwise to disclose.
v
Table of Contents
Abstract ........................................................................................................................... iii
Acknowledgements .........................................................................................................iv
List of Figures ................................................................................................................. vii
List of Tables .................................................................................................................. vii
Chapter 1 – Introduction .................................................................................................. 1
Chapter 2 - Approach to the Literature ............................................................................ 5
Clinical Decisions ......................................................................................................... 6
Educational Training Interventions ............................................................................... 8
Dissertation Aims ......................................................................................................... 9
Chapter 3 - Differentiating the Unconscious Racial Biases and Attitudes of Physicians,
Nurses, and the Public: Implications for Future Healthcare Education and Practice ..... 12
Conceptual Framework ........................................................................................... 14
Methods .................................................................................................................. 15
Results .................................................................................................................... 20
Discussion .............................................................................................................. 26
Conclusion .............................................................................................................. 30
Chapter 4 - The Personal, Practice and Community Characteristics of Family Medicine
Physicians Engaged in Unconscious Bias Mitigation Activities ..................................... 32
Conceptual Framework ........................................................................................... 34
Methods .................................................................................................................. 37
Results .................................................................................................................... 39
Discussion .............................................................................................................. 43
Conclusion .............................................................................................................. 45
Chapter 5 - An Interpretive Phenomenological Analysis of Family Medicine Physicians’
Perspectives of and Experiences with Unconscious Bias and Unconscious Bias Training
...................................................................................................................................... 47
Methods .................................................................................................................. 49
Results .................................................................................................................... 53
Discussion .............................................................................................................. 58
Conclusion .............................................................................................................. 61
Chapter 6 – Implications ................................................................................................ 63
vi
Policy ......................................................................................................................... 65
Education ................................................................................................................... 66
Bibliography .................................................................................................................. 68
Appendices ................................................................................................................... 76
Appendix A: Overview of Studies Associating Unconscious Bias to Patient Outcomes
................................................................................................................................... 77
Appendix B: Summary of Unconscious Bias Intervention Studies ............................. 80
Appendix C: Email Solicitation ................................................................................... 81
Appendix D: AAFP Implicit Bias Survey ..................................................................... 82
Appendix E: AAFP Unconscious Bias Interview Guide .............................................. 83
Appendix F: Consent for Participation in a Research Study ...................................... 86
vii
List of Figures Figure 1 Unconscious bias theory of change framework for healthcare ........................ 15
Figure 2 Conceptual framework depicting how institutionalized racism reinforces the
biases, stereotypes and misbeliefs of clinicians ............................................................ 30
Figure 3 Conceptual framework depicting personal practice and community drivers of
unconscious bias assessment and training activities .................................................... 37
Figure 4 Stages of change towards unconscious bias self-mitigation ........................... 59
List of Tables Table 1 Summary of states and territories categorized by region ................................. 18
Table 2 Summary of Harvard RACE IAT measures ...................................................... 20
Table 3 Mean Harvard Race IAT D-Scores by occupation and region .......................... 23
Table 4 Unconscious bias attitudes by occupation and region ...................................... 23
Table 5 Correlation analysis of Harvard Race IAT explanatory and outcome variables 24
Table 6 Multivariate linear regression modeling effects of social identity on IAT scores 25
Table 7 Summary of statistics from the AAFP Implicit Bias Survey ............................... 40
Table 8 Correlations of variables from the AAFP Implicit Bias Survey .......................... 42
Table 9 Logistic regression model predicting implicit association testing among
physicians ..................................................................................................................... 42
Table 10 Logistic regression model predicting implicit association training among
physicians ..................................................................................................................... 43
1
Chapter 1 – Introduction
2
It’s been nearly twenty years since the Institute of Medicine’s (IOM) Committee
on Understanding and Eliminating Racial and Ethnic Disparities in Health Care first
reported its findings that providers’ biases may be contributing to racial and ethnic
disparities in healthcare [1]. In that report, they emphasized the need for research that
provided a greater understanding of a) the prevalence of unconscious bias and b) the
processes by which it impacts differential treatment. Despite this evidence gap, health
care systems and organizations have forged ahead with the implementation of
preventive measures to minimize the impact of unconscious bias [2-4]. However, fast
forward to 2020, a year in which a global pandemic has disproportionately impacted
racial minorities, and unconscious biases are still considered to be influencing clinicians’
COVID-19 diagnostic and treatment decisions [5, 6]. The impacts of unconscious bias
gained even greater attention in the aftermath of the death of Dr. Susan Moore, a
physician whose public pleas for more equitable treatment went grossly ignored and
heightened the sense of urgency and commitment to address healthcare inequities in
the post-pandemic era [7].
To be clear, bias is the attitudes, behaviors or tendencies that lead individuals to
prefer, favor or evaluate more positively one group relative to another. It may be
expressed consciously (explicit), where the individual is very clear in his or her feelings
and or intentions or unconsciously (implicit), operating without his or her awareness and
even in direct opposition to one’s espoused beliefs and values. Studies have shown
that despite the most egalitarian of viewpoints, bias is pervasive among all health care
professions and more specifically, that clinicians harbor unconscious racial biases at the
same rate or greater than the general population [8]. A clinicians’ ability to deliver a
3
differential diagnosis and treatment that is both equitable and optimal is often limited by
time, complexity and cognitive overload [1, 9]. However, the process may be further
constrained by lack of cultural competency and or unconscious biases, especially when
race is a factor, which has shown to increase racial health care disparities [10-13].
Clinicians’ biases have been associated with a number of diagnostic and treatment
recommendations, including pain, coronary artery disease, kidney dialysis,
contraception and prenatal care, as well as patient-provider communication, satisfaction
and adherence to treatment [14-17]. Interventions to address clinicians’ unconscious
bias often emphasize increasing awareness and teaching skills that mitigate its
influence in clinical practice.
Physician biases have been demonstrated to begin in medical school, throughout
residency training and reinforced by the health care system [18, 19]. For example, a
2017 study of first year medical students from 49 US medical schools found that faculty
role modeling discriminatory behavior towards LGBTQ patients significantly increased
students’ unconscious biases [20]. A more recent study found that medical school
curriculum, policies and culture increased students negative explicit racial attitudes,
resulting in a decreased intention to practice in underserved communities or with
minority populations [21]. Early evaluations of curricula designed to promote effective
dialogue on race and racism for medical students has shown some promise at reducing
these effects [22]. In addition, clinicians’ biases have been shown to also be moderated
by their personal identity (i.e. race, gender, etc.) [23, 24]. For example, in a study of
implicit and explicit racial bias among medical doctors, African Americans showed no
preference for either White Americans or Black Americans and females showed weaker
4
preference for White Americans than males. This association demonstrating the
potential for personal identity to moderate unconscious bias provides further support for
cultivating a diverse and inclusive healthcare workforce to reduce health care disparities
[25, 26].
This dissertation serves to make a contribution by providing an examination of
the prevalence of unconscious bias, influential factors beyond that of the individual and
conceptualize ways in which it develops within the healthcare workforce. As the
following review of the literature will demonstrate, there has been extensive study of the
association between unconscious bias and patient outcomes, as well as evaluations of
interventions. However, neither has produced definitive conclusions, which suggest
there is a need for research that examines some of the more fundamental and principle
aspects of unconscious bias in order to move this field of research forward.
5
Chapter 2 - Approach to the Literature
6
A scoping review of the literature occurred in two phases. First, with an
examination of studies associating unconscious bias with clinicians’ diagnostic and
treatment decisions and second, with an examination of studies evaluating the
effectiveness of educational training interventions with the aim of reducing unconscious
bias. For the purposes of both reviews, studies were only included if they were
conducted in the United States using practicing clinicians, not students or healthcare
trainees. The reason for excluding international studies is two-fold. First, because the
formation of unconscious bias relies on knowledge gained through social interactions
and experiences, there may exist unknown contributing factors in an international
context that are not applicable to the U.S. that may potentially alter findings. Second,
including only U.S. based studies is significant especially in the case of racial biases.
Because race is socially constructed, it can vary over time and from place to place,
indicating that racial bias between countries may not be comparable. Lastly, medical
students and other healthcare trainees are excluded as they are not yet in a position to
make clinical decisions relevant to patient care.
Clinical Decisions To date, there have been three systematic reviews of the literature analyzing the
association of unconscious bias to clinicians’ diagnostic and treatment decisions, which
includes a total of 58 studies [27-29]. After applying the inclusion criteria above to these
58 studies, that reduced the number of eligible studies to 28. As a follow up, an
additional search of the literature was conducted using the original combination of
MeSH keywords and criteria from each of the previously published systematic reviews
and by applying the additional criteria. The results returned only one new study
7
examining the unconscious biases of oncologists, which brought the total number of
studies to 29.
To summarize, the findings have produced mixed results, which indicates that
the impact of clinicians’ unconscious biases on their healthcare delivery decisions is
inconclusive. For example, some were able to demonstrate an effect on patient-
provider communication (n=3), diagnosis (n=7) and treatment (n=11), while nearly half
found no effects (n=13). Overall, the research was conducted using more than 7000
participants which included predominantly physicians from primary care specialties (i.e.
internal medicine, family medicine and pediatrics), although emergency medicine,
psychiatry, oncology, and surgical specialties were also represented as well as nurses.
The majority of study participants were recruited using convenience sampling
techniques due to their affiliation with a particular institution or organization and of those
studies using random sampling methods, participants were recruited via email, phone,
or mail. Each of the studies using convenience sampling were conducted at a single
site. All used cross-sectional study designs consisting mostly of hypothetical patient
care scenarios in assessing their outcomes. In these hypothetical situations, a
participant was presented with either a video vignette, case study or patient simulation
and asked to make a diagnosis or treatment recommendation to assess the impact of
bias on those decisions. Each study included an assessment of participants’ racial bias,
with a few also including a combination of gender, age, class, and socioeconomic status
bias. Three studies examined race from the perspective of medical compliance,
cooperation, and attitudes.
8
Over half of the studies reviewed used the Implicit Association Test (IAT)
developed by Harvard’s Project Implicit which has been described and validated
elsewhere in the literature [30, 31]. Of the remaining studies, the assumption method
was used repeatedly. This method measures differences across groups; the
assumption being that the majority of participants are explicitly motivated to disregard a
factor such as a patient’s race and therefore if a difference in diagnosis or treatment
does occur it can be inferred that the result is due to an implicit or unconscious process.
Differential diagnoses were most likely identified using the assumption method, whereas
studies using the IAT mostly identified differences in treatment and communication. All
of the studies included in this summary can be found in Appendix A.
Educational Training Interventions The second part of this review focuses on studies designed to demonstrate the
effects of unconscious bias educational training interventions. The same inclusion
criteria were applied to the literature as before, however, with one major exception.
Because no studies could be identified using practicing clinicians as the target study
population, this review includes those studies conducted using students in pre-health
professions (i.e. dietetic, medical, psychology, etc.), which might provide some
indication as to the appropriateness of these interventions for practicing clinicians in the
future. However, even with the expanded criteria, this still only resulted in fewer than
ten studies, which also included one systematic review, and as before, outcomes varied
[32-36]. For example, in one study an intervention was determined to be effective at
reducing unconscious racial biases among psychology students, but in another
completely separate study, the same intervention increased them in a different group of
psychology students [33, 34]. Other than demographics, not much if anything is known
9
about the participants in these studies. Having additional contextual factors beyond just
demographics may provide some indication as to why an intervention would work in one
group and not another.
The studies in this review were conducted much in the same way as those in the
previous review. They consisted mostly of cross-sectional designs, using convenience
samples for their study populations, and primarily examined effects on Black/white racial
bias, however some also looked at obesity bias. Effects were often measured using
pre/post analysis of implicit association test scores however, due to small sample sizes
and only minimal information provided regarding the analytical methods, even those that
demonstrated some positive effect at reducing unconscious bias were unable to
produce statistically significant effects to be considered reliable enough for practical
widespread use [32]. It is worth noting that use of the implicit association test as a
measurement tool to quantitative assess changes in individual’s unconscious biases or
the outcomes of curricular interventions is not recommended because unconscious bias
has been determined to be malleable and changes over time as social knowledge and
experiences change [37, 38]. A complete summary of the studies included in this
review can be found in Appendix B.
Dissertation Aims
As this review has demonstrated, the findings across these two branches of the
unconscious bias literature are highly variable and remain inconclusive. As with any
quality or performance improvement intervention in primary care, there exists a need for
research that examines the contextual factors (i.e. individual characteristics, practice
dynamics and or community/organizational culture) that potentially contribute to or can
10
be leveraged to disrupt unconscious bias [39, 40]. This research should be
comprehensive, using both qualitative and quantitative methods in an iterative process
as suggested by the literature, however, qualitative methods are largely underutilized in
examinations of unconscious bias [41-43]. While many opportunities exist for future
research to address some of the weaknesses and gaps previously highlighted by
others, this dissertation aims to prioritize the following while making a contribution to the
current literature.
First, the unconscious biases of the health workforce are not well understood and
need to be examined to better assess their potential associations to patient care. As
implicit associations are known to be constructed based on social knowledge and
experiences, it warrants that variations may potentially exist. These variations may also
be contributing to the variable outcomes observed across the current literature. Once
evaluated, these findings could be leveraged in a way that leads to a more definitive
conclusion regarding their influence on patient outcomes and or more robust
interventions. Second, in addition to individual factors, there are also potentially
practice and community factors associated with individuals’ decisions to participate in
certain unconscious bias reduction activities that are not well known. If so, this
potentially presents an opportunity to develop interventions that disrupt unconscious
bias at the organizational and or community levels in addition to those focused on
modifying individual behaviors. Lastly, given the complexity of unconscious bias, it
warrants further examination and exploration outside of the two approaches discussed
in the review above. This necessitates applying qualitative methods to the study of
11
unconscious bias within the target population to generate hypothesis for future research
that moves the field closer to its aim of reducing disparities in healthcare delivery.
12
Chapter 3 - Differentiating the Unconscious Racial Biases and
Attitudes of Physicians, Nurses, and the Public: Implications for
Future Healthcare Education and Practice
13
Studies have demonstrated that the unconscious biases of healthcare
professionals are a contributor to racial healthcare disparities as they modify clinicians
decisions regarding care access and quality [27, 44, 45]. Much of the evidence used to
support this conclusion has been generated using mostly primary care physicians as
study participants (Appendix A) [27-29]. While some studies have included other types
of clinicians and or medical specialties, seldom if ever are those results stratified to
allow for comparisons between groups. However, a 2016 market survey by Medscape
found that physicians in primary care specialties reported fewer biases towards patients
than those in emergency medicine and psychiatry (62% and 48%, respectively) [46]. In
addition, it’s also been suggested that pediatricians may hold fewer biases towards
patients than any other specialty due characteristics associated with their training and
experiences working specifically with children [47].
Whereas the unconscious biases of physicians and providers as a whole have
been thoroughly examined, little is known about the unconscious biases of nurses
independent from other provider types. Wherein they are described in the literature the
focus is mostly didactic, only providing frameworks and strategies to mitigate the effects
of unconscious bias in nursing education and practice [48-53]. Advanced practice
nurses are increasingly providing holistic patient centered care that requires them to
make care decisions and treatment recommendations to prevent and manages complex
biopsychosocial issues independent of physician oversight [54-57]. These decisions
are also subject to influence from unconscious bias, which justifies the need to examine
nurses as thoroughly as physicians to infer their potential contribution to health care
disparities
14
This study aims to make a contribution to literature by examining and
distinguishing the implicit associations and attitudes of physicians and nurses in
reference to one another and the general public. According to the primary care
performance improvement literature, understanding the contextual factors of an
intervention, such as individuals’ attitudes towards it, are necessary as they are likely to
moderate behaviors associated with effectiveness [39, 40]. Previous studies comparing
the unconscious biases of primary care providers to the local community found no
substantial differences and suggested bias should be considered more of a societal
issue and less as a healthcare issue [8]. If so, that would then suggest that even when
stratified by type of provider, the unconscious biases of healthcare professionals are the
same as those of the general public and shaped by the same social knowledge and
experiences. However, different outcomes for healthcare professionals would indicate
that the unconscious biases of healthcare professionals are mediated by additional
differential knowledge and experiences encountered throughout medical education,
training, and practice, which may require alternative interventions.
Conceptual Framework
This study presents a theory of change framework conceptualizing how
unconscious bias results in disparate healthcare outcomes for patients and
opportunities to disrupt it (Figure 2). The academic medical literature includes
numerous studies examining interventions for disrupting the formation and effects of
unconscious bias in healthcare settings [37, 38, 58-61]. Type A interventions are
designed to disrupt the activation of stereotypes individuals form based on knowledge
and experiences gained from their environments in relation to their social identities.
15
However, once activated, Type B interventions aim to counteract these associations
and replace them with new more positive ones that reposition individual attitudes and
beliefs. Lastly, Type C interventions are intended to interrupt behaviors strongly
associated with unconsciously biased beliefs before they result in judgments and or
actions that result in disparate outcomes for certain groups of patients. The
effectiveness of Type A and B interventions are often measured quantitatively using pre
and post assessments examining changes in individuals’ IAT scores and qualitatively
using surveys that examine attitudes and beliefs [60, 62-64]. There are not yet any
studies examining the effectiveness of Type C interventions in healthcare settings with
clinicians and patients.
Figure 1 Unconscious bias theory of change framework for healthcare
Methods
Data from this study comes from Harvard’s Project Implicit, the most widely used
and well validated measure of implicit associations [65]. Through the Project Implicit
16
demonstration website IAT data has been collected from millions of voluntary
respondents across the globe for nearly two decades. Based upon data from the 2010
Census, Project Implicit respondents tend to be younger (median = 38.1 versus 23.0
years), female (49.8% versus 59.4%) and reflect the racial demographics of the regions
in which participants are located [66]. This study examines data from two samples of
respondents to the Race IAT from 2015 to 2019. Sample 1 includes respondents
categorized by age, geographical location, political affiliation, religious identity,
education, and income while Sample 2 is limited to a subset of respondents categorized
by occupation and geographic location only (Table 6).
Occupation data is available by 65 occupational categories, which includes five
categories for healthcare. As this study is specifically interested only in those
healthcare occupations that provide diagnostic and treatment recommendations, the
occupation variable was recoded to specify a) medical doctors, b) nurses and c) all
other occupations, which included for example occupational therapist, lab techs and
home health aides as part of the general public. It is important to note that a limitation of
this occupational data is that it does not specify the different types or levels of training
among medical (i.e. MD vs DO) and nursing (i.e., LPN, RN, etc.) respondents. There
are differences in scope of practice between registered nurses and licensed practical
nurses and education between BSN-prepared nurses and ADN-prepared nurses.
Likewise, the philosophy of care amongst Doctor of Osteopathic Medicine differ
significantly from those trained in allopathic medicine. It is unclear at this time the
potential impact these differences may have on their unconscious biases but may
present an opportunity for future research. Occupation information was only collected
17
for a subset of participants that were not also asked the personal identify questions
listed below. Analysis is conducted separately for these two groups. Age data capture
respondent’s year of birth which has been recoded to one of six generational
categories: The Greatest (1910-1924), the Silent (1925-1945), Boomers (1946-1964),
Generation X (1965-1979), Millennials (1980-1994) and Generation Z (1995-2012).
Geographical locations are captured by state. Previous research has demonstrated
that the unconscious biases of whites towards Blacks aggregated at the county and
state level are higher in the southeast and are also strongly correlated with disparities in
mortality, birth outcomes, police brutality and Medicaid spending and vary by region
[67]. As suggested, if unconscious bias is a societal issue then this regional variation
should also be consistent among healthcare professionals and reflected in their
perceptions overall. Regional variation of unconscious bias among healthcare
professionals has not been examined in the literature which presents an opportunity for
future research evaluating its correlation to health outcomes at a macro-level that may
be considered in regional disparity reduction initiatives should evidence continue to
suggest unconscious bias is indeed a healthcare issue. The geographical data
captured is recoded to one of six cultural regions based on aggregated attitudes and
beliefs: Caribbean, Frontier, Northeast, Midwest, Pacific and South (Table 5) [48].
Political Affiliation is measured using a seven-point scale ranging from strongly
conservative to strongly liberal and uses neutral as a reference. Religiosity is
measured using a four-point scale ranging from not at all (reference) to strongly
religious. Religious identify captures five of the major global faith traditions to include
Buddhist, Christian (Catholic and Protestant), Hindu, Jewish and Muslim using no
18
religion as a reference. Education captures 14 different types to include specific
degrees and uses high school as a reference. Finally, income is measured in
increments of $10,000 per year ranging up to more than $200k per year. As a
reference, $70,00 was used as the median annual U.S. income based on data from the
U.S. Census.
Table 1 Summary of states and territories categorized by region and occupation
Region States and Territories Physicians
(n=1128)
Nurses
(n=1462)
Caribbean American Samoa, Micronesia, Guam, Marshall Islands, Northern
Mariana Islands, Puerto Rico, Palau, Virgin Islands
0 1
Frontier Arizona, Colorado, Idaho, Kansas, Montana, Nebraska, Nevada,
New Mexico, North Dakota, Oklahoma, South Dakota, Texas,
Utah, Wyoming
112 164
Northeast Connecticut, Delaware, District of Columbia, Maine,
Massachusetts, Maryland, New Hampshire, New Jersey, New
York, Pennsylvania, Rhode Island, Vermont
258 241
Midwest Illinois, Indiana, Iowa, Michigan, Minnesota, Missouri, Ohio,
Wisconsin
307 613
Pacific Alaska, California, Hawaii, Oregon, Washington 192 249
South Alabama, Arkansas, Florida, Georgia, Kentucky, Louisiana,
Mississippi, North Carolina, South Carolina, Tennessee, Virginia,
West Virginia
259 194
The outcomes of interest in this study are the IAT D-score and attitudes. Implicit
associations are measured using a D score that has a theoretical range of -2 to +2 [68].
Respondents with a D-score equal to 0 (± .15) demonstrate no preference for either
19
white or Black individuals, whereas more positive scores suggest a ‘slight’ (.15<),
‘moderate’ (.35<) or ‘strong’ (.65<) preference for whites. To measure attitudes, survey
participants were asked to reflect on the exercise using three statements to indicate
their level of acceptance or disregard of their IAT results. It’s been suggested that
individuals who express agreement with and acceptance of these statements are able
to quickly process and understand their negative implicit associations and move
towards actions that dismantle them [46-48]. This may provide some insight as to
healthcare professionals intentions to take actions that address their unconscious
biases. The statements are measured using a four-point Likert scale from “strongly
disagree" (-2) to “strongly agree" (+2) which was recoded into a binary variable of
“disagree” or “agree”. The questions statements include:
a. My IAT score reflects the culture that I am exposed to, but not me, personally b. Whether I like my IAT score or not, it captures something important about me c. The IAT reflects something about my automatic thoughts and feelings concerning this topic
Analysis includes a summary of each of the described measures (Table 6)
followed by bivariate analysis (Tables 7 and 8) and multivariate linear regression (Table
10). A two-sample t-test was conducted to test the null hypotheses that there are no
differences in either overall IAT D-scores or attitudes towards unconscious bias
between a) physicians and nurses, b) physicians and the general public or c) nurses
and the general public. Each bivariate analysis also includes an examination by region
to detect geographical differences that may be compared to previous research. Lastly,
correlation analysis (Table 9) followed by multivariate linear regressions were
conducted using Sample 2 to examine the strength of the association between the
20
explanatory and dependent variables and to model the likelihood that an individuals’
social identify could predict their IAT score as described in the conceptual framework
provided in Figure 2.
Results
Study respondents were well distributed across each region except the
Caribbean, which did not include any physicians and therefore was excluded from
further comparative geographical analysis. The greatest proportion of respondents in
Sample 1 were individuals from the Midwest, those in the Millennials age group, those
who possessed bachelor’s degrees and those with an annual income between $20-
$30k. The majority of respondents held no specific political or religious beliefs which
reflects national trends [69]. While physicians and nurses make up less than 0.5% of
total survey respondents (n=678,196) they represent approximately 3% of the
respondents who identified their occupation. Overall, IAT D-scores indicate a slight
preference for whites among all respondents (M=.2817, SD =.44). Healthcare
professionals IAT scores were higher than the general public where nurses showed a
slightly greater preference for whites than physicians (.3331 and .3293, respectively).
The majority of respondents tend to agree that the IAT is more an indicator of
themselves as individuals, reflecting their automatic thoughts and feelings as opposed
to a reflection of their culture which may indicate acknowledgement and acceptance that
could lead some respondents to take further action to address their existing biases.
Table 2 Summary of Harvard RACE IAT measures
21
Demographics
Variable Category N %
Age Group Greatest 1,848 .07
Silent 10,754 .38
Boomer 174,230 6.19
Gen X 399,640 14.20
Millennials 1,146,896 40.76
Gen Z 1,080,113 38.39
Political Identity Strongly Conservative 90,130 3.10
Moderately Conservative 253,717 8.73
Slightly Conservative 237,347 8.17
Neutral 803,571 27.65
Slightly Liberal 316,735 10.90
Moderately Liberal 741,864 25.53
Strongly Liberal 462,564 15.92
Religiosity Not at all religious 1,070,433 36.55
Slightly Religious 855,654 29.21
Moderately Religious 698,562 23.85
Strongly Religious 304,226 10.39
Religion No religion 1,078,835 37.75
Buddhist 42,681 1.49
Catholic 646,679 22.63
Protestant 788,305 27.59
Hindu 33,117 1.16
Jewish 80,604 2.82
Muslim/Islamic 57,396 2.01
Other 129,966 4.55
Education High School 20,009 7.42
Some College 29,716 11.02
Associates 11,017 4.08
Bachelor’s 82,715 30.66
Some graduate school 11,389 4.22
Master’s 60,502 22.42
Juris Doctorate 11,921 4.42
Medical Doctor 4,822 1.79
Doctor of Philosophy 16,130 5.98
M.B.A. 7,238 2.68
Other advanced Degree 3,840 1.42
Occupation Physician 1,317 0.19
Nursing 1,621 0.24
General Public 73,127 10.78
Income More than $200k 14,483 4.60
<$200k 3,203 1.02
<$190k 1,202 .38
<$180k 2,059 .65
<$170k 1,801 .57
<$160k 2,709 .86
<$150k 5,185 1.65
<$140k 3,522 1.12
<$130k 4,808 1.53
<$120k 7,157 2.27
<$110k 8,379 2.66
<$100k 13,302 4.22
22
Question Disagree Agree
My IAT score reflects the culture that I am exposed to, but not me, personally.
191,438 (40.44)
281,981 (59.56)
Whether I like my IAT score or not, it captures something important about me.
107,021 (22.36)
371,578 (77.64)
The IAT reflects something about my automatic thoughts and feelings concerning this topic.
117,444 (24.54)
361,049 (75.46)
While analysis could find no difference between the overall IAT scores of
physicians (M=.329, SD=.45) and nurses (M=.333, SD=.43), each was greater than the
general public (M=.295, SD=.45), p< .005. When examined by region, results show that
physicians’ IAT scores were greater than each of the other groups in the South
(M=.3437, SD=.47) and greater than the general public in the Northeast (M=.3607,
SD=.45). Nurses IAT scores were greater than any other group in the Midwest
(M=.4039, SD=.40) and in the Frontier (M=.3756, SD=.42).
<$90k 11,783 3.74
<$80k 16,686 5.30
<$70k 20,438 6.49
<$60k 27,101 8.60
<$50k 32,419 10.29
<$40k 32,992 10.47
<$30k 36,251 11.51
<$20k 34,086 10.82
<$10k 35,421 11.25
Regions Pacific 406,309 19.55
Caribbean 1,342 0.06
Midwest 502,886 24.20
Northeast 452,206 21.76
South 406,219 19.55
Frontier 308,107 14.83
Overall IAT D- Scores
Occupation Mean (Std Dev) Min Max
Physicians .3293 (.45) -1.25 1.43
Nurses .3331 (.44) -1.29 1.47
General Public .2946 (.45) -1.76 1.64
Unconscious Bias Perceptions N (%)
23
Table 3 Mean Harvard Race IAT D-Scores by occupation and region
Nationwide Pacific Midwest Frontier South Northeast
Physician .3293^ .2763 .3333* .3257 .3437*^ .3607^
Nurses .3331^ .2538 .4039^ .3756^ .2656 .2976
Gen Public .2946 .2754 .3089 .3066 .2862 .2947
p-value < .05 *as compared to nurses and ^ as compared to the general public
Upon examining attitudes towards unconscious bias, results show that among nurses
there is greater agreement than among any other group that unconscious bias is a
reflection of one’s culture (M=.2658, SD=.96) and less an indication of individualistic or
automatic thoughts towards people of another race. Physicians’ attitudes (M=.4806,
SD=.87) were more similar to that of the general public (M=.5039, SD=.86) and
reflected the opposite perspective. This observation is consistent across each region
except for in the South where agreement among nurses is highest that unconscious
biases are more the result of individuals’ own thoughts and feelings (M=.5206, SD=.86)
than a reflection of the culture (M=.2727, SD=.96).
Table 4 Unconscious bias attitudes by occupation and region
My IAT score reflects the culture that I am exposed to, but not me, personally. Nationwide Pacific Midwest Frontier South Northeast
Physician .0428*^ -.2126*^ .1160* 0*^ .1839 0
Nurses .2658^ .2777^ .3209^ .3118 .2727 .1141
Gen Public
.1556 .1113 .1962 .1865 .1670 .1311
Whether I like my IAT score or not, it captures something important about me. Nationwide Pacific Midwest Frontier South Northeast
Physician .4806 .5905 .4696 .5211 .2686*^ .6058
Nurses .4554^ .5104 .4087^ .4408 .5372 .4666
Gen Public
.5039 .5363 .5297 .4771 .4711 .5061
The IAT reflects something about my automatic thoughts and feelings concerning this topic.
Nationwide Pacific Midwest Frontier South Northeast
24
Physician .4799* .5312 .6555*^ .3714 .1771*^ .6115*^
Nurses .3932^ .5 .3410^ .2688^ .5206 .4324
Gen Public
.4663 .4867 .4852 .4418 .4441 .4665
p-value < .05 *as compared to nurses and ^ as compared to the general public
Correlation analysis identifies weak yet statistically significant associations
between many of the explanatory and outcome variables. The highest correlations
were between IAT Scores and attitudes that unconscious biases reflect one’s culture
(r=.2477) followed by political identity and that biases reflect individuals (r=.1550).
Multivariate linear regression analysis was used to test if social identity significantly
predicted respondents IAT scores. The results indicated that the model explained 2%
of the variance (R2=.0215, F(59,107242)=39.77, p<.001) where individuals who
identified as strongly conservative showed a greater preference for whites
(β=.14,p<.001), as did those with annual incomes between $90-$100k (β=.0195,p=.01).
Across regions, the Midwest was most similar to the South (β=.0041, p=.332).
Preference for whites declines significantly among younger generations, those who
express more liberal political identities and stronger religious beliefs (p<.001).
Preferences for whites was also lowest among those of Muslim and Protestant faith
traditions (p<.001). Preference for whites declined with increasing levels of education
beyond high school, except for those with medical degrees (β= -.0072, p=.569).
Preference for whites increased significantly up to an annual salary of $30k (p<.001)
then varied.
Table 5 Correlation analysis of Harvard Race IAT explanatory and outcome variables
Variable IAT Score Culture Individual Thoughts
IAT Score .2477* -.0702* -.0717*
Region .0102* .0108* .0009 -.0023
25
Age -.0242* .0391* -.0104* -.0139*
Political -.0915* -.0345* .1550* .1380*
Religiosity -.0210* .0074* -.0251* -.0342*
Religion -.0200* -.0341* .0314* .0398*
Education -.0098* -.0506* .0486* .0523*
Income -.0390* .0471* .0015 -.0078*
Table 6 Multivariate linear regression modeling effects of social identity on IAT scores
R2=.0215, F(59,107242)=39.77, p<.001)
Category Variables Coefficient Std. Err p-value 95% Confidence Interval
Regions Pacific -.0198 .0043 <.001 -.0284 -.0113
Frontier -.0107 .0048 .027 .0203 -.0012
Northeast -.0266 .0046 <.001 -.0357 -.0174
Midwest .0041 .0042 .332 -.0042 .0125
South Ref
Age Group Greatest .1099 .1078 .308 -.1013508 .3211871
Silent .1334 .0127 <.001 .108415 .1584254
Boomers Ref
Gen X -.0423 .0041 <.001 -.050379 -.0343349
Millennials -.0172 .0042 <.001 -.0255649 .0088285
Gen Z -.1216 .0317 <.001 -.1838709 -.0593829
Political Id Strg Cons .147578 .01003 <.001 .1279193 .1672367
Mod Cons .1141 .0063 <.001 .1018036 .1264756
Slight Cons .0573 .0063 <.001 .0449904 .0696241
Neutral Ref
Slight Lib .0125 .0057 .03 .0012206 .0237032
Mod Lib -.0206 .0047 <.001 -.0298623 -.0114205
Strg Lib -.0646 .0050 <.001 -.0745034 -.0547046
Religiosity Not at all Ref
Slightly -.0215 .0045 <.001 -.0304032 -.0126438
Moderately -.0536 .0055 <.001 -.0645076 -.0428654
Strongly -.0860 .0065 <.001 -.0988633 -.0731821
Religion None Ref
Buddhist .0250 .0116 .031 .002341 .047703
Catholic .0322 .0054 <.001 .0215336 .0429083
Protestant -.0179 .0052 .001 -.0281809 -.0077803
Hindu .0124 .0178 .485 -.0225054 .0474234
Jewish .0801 .0077 <.001 .0649006 .0952219
Muslim -.1333 .0202 <.001 -.1729908 -.0937552
Other -.0359 .0080 <.001 -.0517429 -.0202298
Education High School Ref
Some college
-.0389 .0092 <.001 -.057014 -.0208677
Associates -.0171 .0107 .112 -.0381643 .0039939
Bachelor’s -.0237 .0085 .006 -.0405745 -.0069291
Some Grad -.0402 .0107 <.001 -.0611389 -.0191916
Master’s -.0398 .0088 <.001 -.0571643 -.0225284
JD -.0372 .0103 <.001 -.0575827 -.016874
MD -.0072 .0126 .569 -.0319166 .0175577
PhD -.0391 .0101 <.001 -.0589845 -.0192166
Other adv -.0293 .0147 .047 -.0581753 -.0004329
MBA -.0470 .0117 <.001 -.0699215 -.024144
26
Income More than $200k
.0235 .0075 .002 .0088 .03825
<$200k .0208 .0134 .120 -.0054 .0471
<$190k .0346 .0188 .065 -.0022 .0715
<$180k .0093 .0152 .540 -.0204 .0390
<$170k -.0298 .0155 .055 -.0603 .0006
<$160k -.0085 .0133 .523 -.0345 .0175
<$150k .0099 .0105 .342 -.0106 .0305
<$140k .0062 .0116 .593 -.0166 .0289
<$130k .0059 .0103 .562 -.0142 .0262
<$120k .0110 .0091 .226 -.0068 .0288
<$110k -.0039 .0087 .652 -.0210 .0131
<$100k .0195 .0076 .010 .0047 .0345
<$90k .0077 .0077 .318 -.0075 .0229
<$80k -.0169 .0071 .017 -.0308 -.0029
<$70k Ref
<$60k -.0019 .0064 .759 -.0145 .0105
<$50k -.0170 .0063 .007 -.0293 -.005
<$40k -.0117 .0066 .076 -.0246 .0012
<$30k -.0290 .0072 <.001 -.0432 -.0149
<$20k -.0337 .0084 <.001 -.0502 -.0172
<$10k -.0605 .0121 <.001 -.0842 -.0367
_cons .3665 .0109 <.001 .3450 .3880
Discussion
This study identified that the unconscious biases of physicians and nurses and
their attitudes towards them differ from the public and in some instances from one
another. Healthcare professionals were found to have a greater preference for whites
than the general public. This is contrary to previous work conducted in Colorado, a
Frontier state, which found no differences between primary care providers and the
general public. However, a limitation of that study was that it did not examine
differences by type of provider and as such was unable to detect the differences
identified by this study to support the conclusion made that unconscious bias is a
societal issue more so than a healthcare issue. A 2018 study conducted with 107 staff
members of the Alabama based Primary Care Research Coalition also examined
differences in implicit associations by healthcare occupation and race [70]. In this
study, medical doctors/registered nurses were compared to non-MD/RN staff (i.e.
27
receptionist, etc.) Unlike the findings reported in this dissertation, where physicians
(.3437) in this region had statistically significantly higher D-scores than either nurses
(.2656) or the general public (.2862), the Alabama study found that D-scores were
higher among non-MD/RN staff (0.51) as compared to MD/RN staff (0.4). Of the 107
subjects, only 22 possessed either an MD or an RN, indicating the study didn’t include a
large enough sample size which limited their ability to examine differences within the
clinical group. While no studies were found examining differences between occupations
in other fields (i.e. judges vs police or teachers vs administrators), the RACE IAT data
set includes both sets of occupations whereas further research could identify if
differences exist.
A review of the literature could find no studies evaluating changes in the
magnitude of the D-score or potential implications of such differences outside of the
theoretically significant effects of ‘slight’ (.15), ‘moderate’ (.35) and ‘strong’ (.65).
However, one study did determine that the confidence intervals for the D-score can be
very large, spanning a range of up to 0.76 points in some instances [71]. This suggest
that an individual with a D-score of 0.4 (moderate) would potentially have a confidence
interval ranging from -0.36 (moderate preference for Blacks) up to 1.16 (strong
preference for whites). For this reason, the use of the IAT as a tool to measure the
effectiveness of interventions is discouraged.
The findings of this study also demonstrate that in some regions, the
unconscious biases of nurses show a greater preference for whites which may be of
even greater concern than those of physicians, especially in areas where nurses have
full practice authority. For the most part, physicians and nurses responded equally
28
(Table1), except in the Midwest, where responses from nurses were nearly double that
of physicians (n=613 and n=307, respectively). Also, preferences towards whites were
highest among nurses in the Midwest region. This could be interpreted to suggest that
nurses outnumber physicians in this region, indicating that they should be prioritized in
research and interventions to address unconscious bias in an effort to reduce
healthcare disparities. However, this study found that nurses’ perceived their
unconscious biases as a reflection of the cultures to which they are exposed and less
as a reflection of their own explicit or implicit thoughts and feelings regarding race. This
may suggest that nurses are less inclined than physicians to participate in unconscious
bias interventions targeting individuals and are more likely to support those addressing
practice and workplace culture. Overall, these observed differences between
physicians and nurses may just reflect personal and professional character differences
(i.e. elitism, empathy, etc.) previously described in the literature. For example, in a
comparison study of empathy among female nurses and physicians, nurses scored
higher on 15 out of 20 indicators [64]. Increasing empathy has been described as one
method of mitigating unconscious bias which would explain why nurses would have
lower IAT scores than physicians [43, 65, 66].
This study also examined the magnitude to which measures of individuals’ social
identity were associated with their level of racial preference and found that it predicts
less than 2% of the variance in the analytical model, suggesting greater influence
comes from other and perhaps still unknown factors. Identifying these will be critical to
developing effective interventions. One area of investigation to start would be to
examine the contextual factors associated with our healthcare and medical education
29
systems that may either introduce or reinforce individuals’ existing unconscious bias.
An interesting finding from this analysis was that while white preference declined with
increasing educational attainment, among individuals with medical doctor degrees,
preference for whites remained similar to those with only a high school diploma. One
value of higher education is that it brings individuals into contact with culturally diverse
and inclusive learning communities in which the stereotypes and biases that contribute
to the formation of implicit associations can be counteracted in university settings. The
finding that those with medical degrees are no different than those with high school
diplomas suggest that somewhere along the medical education continuum, those
positive effects are either lost or other more negative influences are reinforced. This
conclusion has been supported by previous research demonstrating medical students
false beliefs about race and experiences with racism in medical school [14, 20-22].
Minimizing the influence of unconscious bias to produce disparate healthcare outcomes
necessitates moving beyond individual and interpersonal factors upstream to identify
and address systemic issues within education and practice. Emerging literature has
begun to describe how medical education and healthcare are rooted in systemic and
oppressive ideologies, such as white patriarchal supremacy, that introduce and or
reinforce students and practitioners explicit and implicit biases, stereotypes and
misbeliefs [21, 22, 72-74]. Some factors that have already been identified include the
poor modeling of patient interactions by faculty, the practice of inferring
biological/genetic racial differences in research and the use of unfounded race
correction factors in clinical guidelines. This concept is depicted in the framework below
(Figure 3). Currently, unconscious bias training targeting individuals has been the only
30
tool available to address the downstream effects of systematic racism however new
resources are emerging in the context of methodologies to address race-based
medicine and healthcare operations using principles of critical race theory [10, 11, 75-
78].
Figure 2 Conceptual framework depicting how institutionalized racism reinforces the biases, stereotypes, and misbeliefs of clinicians
Conclusion
Unconscious biases are determined only minimally by measures of individuals’
positioning with social hierarchies suggesting that other more influential factors need to
be identified and addressed. In healthcare, the unconscious biases of physicians and
nurses differ significantly from those of the general public and show regional variation in
areas where nurses have greater preference for whites. However, greater emphasis
needs to be placed on identifying and addressing factors associated with medical
education and or healthcare delivery that may introduce and or potentiate individuals’
31
unconscious biases. Interventions are emerging that go beyond addressing individual
attitudes and behaviors and refocus on system and institutional level interventions to
reduce healthcare disparities.
32
Chapter 4 - The Personal, Practice and Community Characteristics of
Family Medicine Physicians Engaged in Unconscious Bias Mitigation
Activities
33
Previous research has demonstrated that the unconscious biases of healthcare
professionals differ from those held by the general public and in some instances also
differ based on type of clinicians by geography [79]. The neuropsychological research
suggest that these implicit associations develop from individuals’ social knowledge and
experiences formed through the lens of their intersecting identities [80-82]. However,
when evaluating the contribution of measures of social identity, such as age, political
affiliation, religious identity, education, and income, as potential drivers of unconscious
biases, they are found to contribute only minimally, suggesting that alternative previous
unknown driving factors may exist, especially for healthcare professionals [79]. In
addition, little is known about what factors drive healthcare professionals to participate
in unconscious bias mitigation activities, such as self-assessments and training. To
date, participation in these activities have been primarily voluntary as there are no
mandated licensure or certification requirements, even though it may be “highly
encouraged” by employers, professional societies or others. This presents a potential
challenge to studies designed to examine the effectiveness of unconscious bias training
programs as it introduces a great deal of selection bias and or other limitations that may
significantly skew evaluation results.
The primary care quality or performance improvement literature emphasizes the
need to identify contextual factors to interventions. Contextual factors are those
characteristics and circumstances that are not part of an intervention but likely to
interact, influence, modify, facilitate or constrain an intervention which can determine its
effectiveness [40]. In primary care, these contextual factors are categorized into three
areas; organizational, team and individual [39]. For example, individual-level factors
34
include clinicians’ and administrative staffs’ beliefs regarding the intervention’s value,
motivations to adopt new behaviors as well as their own knowledge, skills, and self-
efficacy. Failure to acknowledge or address these factors during the development and
or implementation of an intervention potentially creates barriers to increasing
awareness, knowledge, and acceptance of it [39]. This study aims to examine
contextual factors associated with physician participation in unconscious bias reduction
activities. This makes a contribution to the literature by providing some insight into who
within the physician workforce is actually participating in unconscious bias training
which has implications for patient care. In addition, the findings have potential use in
the development of more effective interventions and organizational strategies. While
the design and frameworks used to guide the development and implementation of
current unconscious bias interventions in healthcare target mostly individual
characteristics (i.e. empathy, social identity, privilege, etc.) this study intends to also
examine factors associated with physician practices and communities, areas not yet
considered in the unconscious bias literature.
Conceptual Framework
A conceptual framework outlining personal, practice and community drivers
associated with unconscious bias activities, such as self-assessment and training is
presented below (Figure 1). First, this study examines the impact of personal factors
such as gender identity, age and years since residency which provides an opportunity to
demonstrate whether or not their association to unconscious bias mitigation activities
are consistent with previous findings examining their influence on unconscious bias.
35
However, unlike the previous study, due to data limitations, race/ethnicity data is
unavailable for this study.
Years since residency is of particular interest not as a measure of social identity
but because it may provide some evidence regarding the impact of changes to medical
education and residency training aimed at bringing increased awareness and integration
of unconscious bias into curricula to address racial health disparities. While there are
no requirements set forth by either the Liaison Committee on Medical Education
(LCME) for students or the Accreditation Council for Graduate Medical Education
(ACGME) for residents, unconscious bias training has become increasingly embraced
by these and other organizations such as the Association of American Medical Colleges
(AAMC), the American Academy of Family Physicians (AAFP) and others. As such,
curricula in recent years has also evolved to more frequently incorporate unconscious
bias than in previous years suggesting that younger physicians may be more familiar
with unconscious bias and participate in training interventions at proportions greater
than older more established physicians. However, there’s likely to be a strong
correlation between a physician’s age and their years since residency, with the
exception of non-traditional students, suggesting that either variable could be used to
make inferences regarding curricular impact on unconscious bias mitigation activities.
Practice factors to consider are the type of employer a physician works for and
community factors (urban/rural) describing where that practice is located, which
together can be used to infer other associations regarding patient demographics,
policies, care access, etc. In total, eight categories of practice are described including
public vs private, for-profit vs not-for-profit as well as academic and health system
36
settings. The patient demographics of an urban federally qualified health center
(FQHC) will differ drastically from that of a rural sole ownership practice. These
differences are likely to influence the types of patients physicians interact with and or
the standards set for employment (i.e. training) which may be motivating factors for a
physician to self-assess their own biases and or subsequently participate in an
unconscious bias training intervention. In addition, the type of practice to which a
physician belongs such as public vs private, for-profit vs non-profit, etc. may determine
both the availability of this type of training and whether or not it is required for
employment.
While the empirical analysis of this conceptual framework intends to focus on
direct associations between personal, practice and community factors and unconscious
bias mitigation activities, there is recognition that indirect associations, additional
factors, and alternative pathways may also exist. There is not yet evidence to support
that the factors presented here act independent of one another nor that self-assessment
is a required precursor to training. However, this study hypothesizes that implicit
association testing is strongly correlated with unconscious bias training attendance,
which would correspond to differences in attendance between those who take self-
assessment and those who don’t. Testing results, especially those contrary to an
individual’s explicit and or espoused beliefs, may potentially factor into decisions to
participate in training. As the research in this area expands in scope, the constructs
and pathways of these frameworks will become more refined, resulting in the
development of more evidence-based training.
37
Figure 3 Conceptual framework depicting personal practice and community drivers of unconscious bias assessment and training activities
Methods
To examine the personal characteristics associated with awareness of implicit
bias and training activities, this study uses data from a survey (Appendix C) conducted
by the American Academy of Family Physicians (AAFP). There are several rationales
for surveying this particular medical society. First, it is the largest specialty society
dedicated to primary care to which nearly 80% of family physicians belong suggesting a
representative sample of all family physicians can be obtained. Second, nearly half of
all office visits are to primary care physicians which increases the prevalence and
probability that issues of physician biases will emerge in the context of clinical care
more so than specialty care which may provide greater motivation to complete an IAT
and or attend a training [83]. Survey participants include individuals from a randomized
sample of 600 AAFP members who received an invitation to participate via online by
Implicit Association
Test
Practice
Community
Training
Personal
38
answering questions that explore their familiarity and engagement with unconscious
bias assessment and training (Appendix B). While this potentially introduces selection
bias as there are likely unknown factors associated with choice of medical specialty and
membership in a physician association that may also be associated with who
participates in unconscious bias mitigation activities, the sample should be fairly
generalizable to the broader membership and specialty as a whole. Demographic
variables collected on this sample include the following: gender (male/female), age
(categories), community (urban/rural), years since residency (categories) and primary
employer (multiple). Years since residency also includes a recoded binary variable to
distinguish new physicians, those in practice 7 years or less from more established
physicians. Outcome questions of interest include:
a) How familiar are you with the term “implicit bias”? (categories)
b) Have you ever participated in implicit bias training? (yes/no)
c) Have you ever taken an implicit bias test? (yes/no)
Analysis includes a summary of descriptive statistics (Table 1) followed by bivariate
analysis. A chi-squared test is used to identify if statistically significant differences exist
among the outcome variables (a, b, and c) by gender and setting (urban/rural) across
years since residency, new physician status and employer. After which a two-sample t-
test was conducted to test the following null hypotheses:
1. There is no difference in outcomes (b) (who takes the IAT) or (c) (who attends an
unconscious bias training) between:
i. Male and female physicians
39
ii. Physicians who practice in either urban or rural settings
iii. New or older physicians
2.) There is no difference in who attends an unconscious bias training based on
having taken the IAT.
Lastly, a correlation analysis between the explanatory and outcome variables followed
by logistic regression is conducted modeling a) the likelihood that personal, practice and
community factors can predict whether a physician will take the IAT and the b) if those
same factors can predict if they will attend an unconscious bias training. The strengths
of this study are that it examines new measures of unconscious bias in a well-defined
population that allows for broad inferences and generalizations to be made. However,
the lack of more detailed and specific characteristics (i.e. practice size, patient
demographics, etc.) presents a limitation to drawing conclusions that tie into more direct
associations. This presents an opportunity for further research using the findings here
to develop more meaningful measures. The reference individual for both models is a
female new physician, employed by an urban university, as findings from Harvard’s
Project Implicit indicate that IAT survey participants are mostly female and younger.
Analysis used Stata 15.1 and established statistical significance at 95%.
Results
In total there were 222 respondents to AAFP’s online survey which accounts for
approximately 37% percent of the randomized sample invited to participate.
Respondents to the survey were predominantly older (75% of AAFP membership)
female (46% of AAFP membership) physicians. A greater proportion practiced in urban
settings (86.80%) for a private non-profit hospital or health system (25.58%). More than
40
two-thirds indicated that they were “very familiar” and “have a clear understanding of the
term” implicit bias (64.57) however, less than 20% have ever taken the IAT and only a
third have ever participated in an unconscious bias training (31.39%). The remainder of
the descriptive statistics are provided in Table 1 below.
Table 7 Summary of statistics from the AAFP Implicit Bias Survey
Variable Label n %
Gender Female 115 51.80 Male 107 48.20
Years Since Residency
0 to 7 51 22.97 8 to 14 44 19.82 15 to 21 50 22.52 22 or more 77 34.68
Age 30-39 57 25.91 40-49 67 30.45 50-59 56 25.45 60-69 37 16.82 70 or older 3 1.36
New Physician Yes 51 22.97 No 171 77.03
Community Urban 171 86.80 Rural 26 13.20
Employer Federal, state, or local government, (not including universities)
19 8.84
Physicians group (single- or multi- specialty)
42 19.53
Self-employed (majority practice owner, independent contractor, etc.)
41 19.07
Private non-profit hospital or health system
55 25.58
University-owned (public or private) clinic or hospital
28 13.02
Private for-profit hospital or health system
12 5.58
Managed care organization or insurance company
6 2.79
Other 11 5.12
How familiar are you with the term “implicit bias”?
Not at all familiar, never heard the term before now
13 5.83
Somewhat familiar, only heard of the term but never had a clear understanding of the meaning
66 29.60
Very familiar, have clear understanding of the term
144 64.57
Have you ever participated in
No 153 68.61
Yes 70 31.39
41
implicit bias training?
Have you ever taken an implicit bias test?
No 159 81.54
Yes 36 18.46
Total Sample N 222
Chi-squared analysis identified a statistically significant difference in only one
area. Older male physicians participated in unconscious bias training at rates greater
than expected as compared to newer male physicians (38.64% and 11.11%,
respectively). While not statistically significant, physicians 15 to 21 years out of
residency took the IAT in greater proportions than expected than any other age group
(31.82% p=.109). Of the null hypotheses proposed, three were rejected in favor of the
alternate. First, urban physicians (M=.2027, SD=.40) were more likely than rural
physicians (M=.0416, SD=.20) to have taken the IAT (p=.03) and second older
physicians (M=.3491, SD=.48) were more likely than new physicians (M=.2157,
SD=.41) to have participated in an unconscious bias training (p=.04). Lastly, physicians
who had taken the IAT (M=.8611, SD=.35) were more likely to participate in an
unconscious bias training than those who had not (M=.1761, SD=.38) (p<.001).
Correlation analysis outlined in Table 2. identified a (positive) strong and statistically
significant association between years since residency and new physician status (r=.79
and p<.001). Alternatively, a (positive) weak yet statistically significant association was
identified between older physicians and participation in unconscious bias training (r=.15
and p=.04). The strongest (positive) correlation was found between taking the IAT and
participation in an unconscious bias training (r=.54 and p<.001). Lastly, male, or female
42
gender identity demonstrated a (positive) statistically significant weak association to
years since residency (r=.17 and p<.03).
Table 8 Correlations of variables from the AAFP Implicit Bias Survey
Participation Familiarity IAT Gender Years Since
Residency
New Physician
Setting
Familiarity -.0824
IAT .5455* -.0291
Gender (Male or Female)
.0667 -.0728 -.0708
Years Since
Residency
.1170 -.0671 .0622 .1699*
New Physician
.1599* -.0935 .0618 .1105 .7885*
Setting -.0069 -.0255 -.1472 -.0074 -.0162 .0130
Employer -.1338 .0191 -.0492 -.0510 .0196 .0399 -.2233*
Logistic regression analysis was used to test if personal, practice and community
factors can predict participation in implicit association testing (IAT) and training. Though
not significant, the following effects were identified. First, practicing in a private non-
profit hospital or health system (OR=.4117, p=.207, 95% CI:.1038,1.632) or a rural
setting (OR=.17, p=.095, 95% CI:.0217,1.359) decreased the probability that a
physician would take the IAT. Older physicians, specifically those with more than seven
years of practice experience showed higher probabilities of participating in unconscious
bias training (OR=2.745, p=.060, 95% CI:.9594, 7.853).
Table 9 Logistic regression model predicting implicit association testing among physicians
Variable ORs Std Err p-value 95% Confidence Interval
Male .5673 .2558 .209 .2344 1.372
Older Physicians 1.522 .8331 .442 .5208 4.449
Rural .1700 .1804 .095 .0217 1.359
Employer
Federal, state, or local government, (not including universities)
.5029 .4868 .478 .0754 3.353
43
Physicians group (single- or multi- specialty)
.6500 .4791 .559 .1532 2.756
Self-employed (majority practice owner, independent contractor)
1.285 .8937 .718 .3287 5.022
Private non-profit hospital or health system
.4117 .2893 .207 .1038 1.632
Private for-profit hospital or health system
1.400 1.373 .731 .2051 9.564
Managed care organization or insurance company
.6824 .7215 .521 .0683 2.326
Other .3223 .4239 .125 .0165 2.221
Years Since Residency
8 to 14 .9568 .6924 .951 .2317 3.952
15 to 21 2.154 1.397 .237 .6040 7.682
22+ 1.455 .9331 .558 .4142 5.113
Table 10 Logistic regression model predicting implicit association training among physicians
Variable ORs Std Err p-value 95% Confidence Interval
Male 1.159 .3876 .658 .6022 2.233
Older Physicians 2.284 1.005 .061 .9637 5.414
Rural .7485 .3489 .534 .3003 1.866
Employer
Federal, state, or local government, (not including universities)
1.400 .9398 .616 .3755 5.218
Physicians group (single- or multi- specialty)
1.054 .6155 .928 .3358 3.311
Self-employed (majority practice owner, independent contractor)
1.504 .8657 .479 .4864 4.647
Private non-profit hospital or health system
.8213 .4471 .718 .2825 2.387
Private for-profit hospital or health system
.2438 .2819 .222 .0253 2.351
Managed care organization or insurance company
.3847 .4590 .423 .0371 3.989
Other .2546 .2958 .239 .0261 2.483
Years Since Residency
8 to 14 2.745 1.472 .060 .9594 7.853
15 to 21 2.030 1.061 .175 .7289 5.653
22+ 2.233 1.079 .097 .8656 5.758
Discussion
This study aimed to examine the personal, practice and community
characteristics that are associated with the unconscious bias mitigation actions of family
medicine physicians. Overall, no gender differences were identified, a finding that is
consistent with previous work, which suggest that measures of identity have little impact
44
on individuals’ decision to engage in unconscious bias mitigation activities. In total,
participants in practice more than 15 years since residency represent at greater
proportion of the sample at 57.2% which would include individuals approximately age 46
and older. This estimation assumes a traditional pathway of four years of
undergraduate education, four years of medical education and three years of residency
training. This study considered that newer physicians would be more likely to
participate in unconscious bias training than older more established physicians because
of changes to medical education and training in recent years. However, this analysis
identified the opposite which may suggest older physicians may be provided with some
type of incentive to participate or organizations are prioritizing training for physicians
more so than for students and residents. However, the observed difference by age may
also reflect differing generational attitudes towards the concept of unconscious bias, an
area not yet examined in the research [84]. As expected, analysis indicated that IAT
rates were greater among urban physicians as compared to rural physicians which may
indicate that physicians practicing in urban settings experience a greater number of
interactions with more diverse patient populations that potentially trigger implicit
associations that warrant conducting self-assessments and training to address their
unconscious biases towards them. However, as national trends show, rural areas are
becoming increasingly diverse and as such will need a workforce that can address
unconscious bias in order to provide equitable care [85].
Correlation analysis identified a strong association between physician age by
category and years since residency, however results demonstrated that being a new
physician was strongly associated with all three outcomes and a better fit for the
45
regression model. In addition, a weak association was identified between gender and
years since residency suggesting that perhaps as physicians become more established,
the demographic becomes increasingly male, an indication that women may be
dropping out of the medical workforce at older ages [86]. As indicated by the logistic
regression model, older physicians with more than 8 years of practice experience was
the strongest predictor of whether a physician would participate in unconscious bias
training. This may indicate that for this particular demographic of physicians, their
participation in unconscious bias mitigation activities, which may be highly encouraged
by their organizations, could be associated with desires to obtain future leadership roles
that necessitate a commitment to advancing health equity and reducing health care
disparities.
While the logistic regression includes both qualitative and quantitative effects, for
the purposes of this analysis, quantitative interpretations are not overly emphasized due
to limitations with the data. For example, the odds that a physician practicing in a rural
area will participate in an unconscious bias training decreases by nearly 25 percent as
compared to a physician in an urban practice (OR=.745, p=.584). To provide a more
meaningful interpretation of the quantitative effects, more specified continuous
measures of rural would need to be included in the model, such as population size,
population density, etc., in order to demonstrate significant changes in effect size. The
lack of significance demonstrates that the current measures lack the necessary
precision needed to infer strong conclusions regarding their impact on outcomes.
However, it may provide a starting point for further research.
Conclusion
46
While unconscious bias education is highly encouraged within the specialty of
family medicine, physicians’ decisions to participate in mitigation activities are likely
influenced by their employer. Participation among older physicians, specifically those
with more than 8 years of practice experience may potentially be influenced by targeted
employer incentives, such as promotions, performance evaluations, etc. however
additional research is needed to understand further. Future research in this area should
aim to include more specific measures of personal, practice and community factors.
While the evidence demonstrating the effectiveness of unconscious bias training to
reduce healthcare delivery disparities is still unclear, unconscious bias training has
additional value at the organizational level by cultivating a more inclusive and equitable
workplace culture within healthcare with implications for clinical training, policies and
procedures, which have downstream implications for patient care [72, 73, 78, 87].
47
Chapter 5 - An Interpretive Phenomenological Analysis of Family
Medicine Physicians’ Perspectives of and Experiences with
Unconscious Bias and Unconscious Bias Training
48
It has been suggested that the unconscious biases of health care professionals
contribute to healthcare disparities [14, 15, 88-90], a conclusion which has prompted
many health care organizations and medical schools to implement training interventions
that serve to raise awareness of unconscious bias and teach skills that reduce its
influence on clinical decision-making and practice behaviors [4, 91]. However, despite
this widespread adoption of unconscious bias training, the evidence supporting its
effectiveness still remains unclear [32, 92]. Others have suggested possible legal
remedies to address providers’ unconscious biases, although there is no indication that
these will soon come to pass, however, the mere suggestion of it may be enough to
spur organizations to err on the side of caution in the absence of evidence rather than
suffer the consequences of inaction [93-95]. The lack of adequate evidence may
potentially explain why currently there are no mandates or requirements for
unconscious bias training associated with physician licensure, certification, or
accreditation, which indicate participation among physicians remains primarily voluntary.
Voluntary participation is a limitation of outcome effectiveness studies on
unconscious bias training programs as there’s no way to identify or control for the
number of potential selection and or confirmation biases that are associated with who
participates in these trainings and why. However, previous research has provided some
indication that older physicians (those with more than 8 years of practice experience)
and the type of employer a physician works for may be associated with their
participation in unconscious bias mitigation activities [96]. For example, physicians
employed by a federally qualified health center were more likely to indicate they had
participated in an unconscious bias training than those in a private for-profit hospital or
49
health system. It is contextual factors like these (i.e. population demographics, practice
settings, participants attitudes, etc.) that have been shown to moderate the effects and
implementation of an intervention when not well understood [39, 40]. This gap exists
primarily because interventions have been developed and evaluated extensively using
subjects from the general population or sometimes students, rather than the populations
they are intended for (i.e. practicing physicians).
This study aims to identify and examine contextual factors associated with
physicians’ perspectives of and participation in unconscious bias training which
contributes to the literature in the following ways. First, it examines potential outcomes
from unconscious bias training among the population unconscious bias training is
intended for, practicing physicians. This is absent from the existing literature. Second,
it provides some further indication as to who within the physician workforce is
participating in unconscious bias training and additionally why which has implications for
the potential impact on patients. Lastly, the findings may potentially lead to the
generation of new hypotheses and approaches for future research that result in more
innovative interventions beyond those aimed simply at moderating the behaviors of
individual physicians to reduce health care disparities.
Methods
Data for this study was collected from focus group interviews conducted with
members of the American Academy of Family Physicians (AAFP) and analyzed using a
phenomenological approach to identify key themes. The AAFP is the largest medical
specialty society dedicated to primary care to which nearly 80% of family physicians
belong and primary care physicians provide more than 200 million patient visits annually
50
[97, 98]. Given their policies and educational offerings on unconscious bias, AAFP
members provide a well-defined and knowledgeable target population for this study to
which generalizations to other primary care specialties can be inferred. All aspects of
the study were approved by the Institutional Review Board of the AAFP (Appendix F).
The benefit of using focus groups as the primary qualitative data source is that it
allows the opportunity to observe interactions between participants, which is conducive
to the generation of new knowledge [43]. As described by Merton et. al. (1990) focus
groups bring several different perspectives into contact; for some, until they’ve
interacted with others on the topic, they are unaware of their own perspectives. Focus
groups create the opportunity for this type of interaction which is difficult to obtain using
other methods (i.e. individual interviews or participant observation). Also, studies have
shown that focus group settings are more likely to generate knowledge regarding
sensitive and personal information than individual interviews because of the support and
trust provided by peers [99]. Lastly, focus groups have been described elsewhere in the
literature as ideal for family medicine physicians because unlike surveys and other
questionnaires, focus groups are less time consuming and create an informal
atmosphere that rewards them with stimulating debate and discussion [100].
Participants were selected using a purposive sampling strategy based on their
responses to an online survey and self-reported demographics (Appendix C and D).
Selected participants were sorted into two distinct groups; those who have participated
in unconscious bias training within the last year and those who have not to obtain
broadly diverse viewpoints from both perspectives to meet study aims. Having separate
groups creates a homogeneity that allows for a freer flowing within group discussion
51
that can be more easily analyzed to identify key differences in perspectives between the
two groups. A total of 24 participants were invited for this study to accommodate a
potential no-show rate of 20% which resulted in a final group size of 9 participants in
each group. Focus groups were conducted in conjunction with the AAFP’s annual
meeting and participants received incentives in the form of $150 Visa gift cards.
The focus group guide was developed using a funnel strategy approach which
began with a less structured free discussion with open-ended questions such as, “tell us
about yourself, where you live, and where you work?” and “ When you hear the term ‘implicit
or unconscious bias’ what comes to mind first?” to more structured discussion questions
such as “What did you value most from the training?” and “How important do think is it that
implicit bias training be included in the medical school training?” (Appendix D). The major
domains and elements of the interview guide were developed primarily from the
literature on contextual factors of interventions in primary care settings and refined
based on the screening survey findings [39, 40]. Both focus groups concluded with
asking participants to provide a final summary statement suggesting actions the AAFP
should take to communicate the aim and potential outcomes of unconscious bias
education and training interventions to members and strategies to motivate
participation.
Focus group transcripts were analyzed using an interpretive phenomenological
approach (IPA) which is a method that has been identified as ideal for health services
research studies using small groups of participants like focus groups [101, 102]. While
the purpose of IPA is not to generate new theories regarding a phenomenon, like
unconscious bias training, it allows for the identification of key themes that are reflective
52
of physicians perspectives on and experiences with unconscious bias and unconscious
bias trainings that can accessed in the development of further research, which is an aim
of this particular study. In accordance with established IPA guidelines developed by
Smith et. al, audio and written transcripts were listened to and read multiple times to
ensure as the researcher, the content and the context of each participants’ account was
accurate and a true reflection of their experience [103]. The first analysis served to gain
familiarity with the materials, the second to identify and describe themes and the last to
verify the accuracy of the second stage of review. Subsequently, themes were then
sorted and organized into major and subordinate categories.
The ensure that the analysis met a certain stand of rigor the following steps were
taken. First, findings were validated by triangulation to a set of data collected from a
previous study examining the personal, practice and community factors associated with
physicians participation in unconscious bias mitigation activities [96, 104]. This previous
study used a quantitative survey method which also included open-ended responses to
which the findings of this qualitative study could be validated for accuracy. In addition,
the focus group participants for this study were selected based on their responses to the
aforementioned study, indicating that these findings possess a high degree of validity
and credibility. Second, to ensure reliability of the findings, a detailed record of
decisions made at each stage of analysis was maintained to include personal
reflections. Because IPA has been described as “an interpretative process between the
researcher and the researched” the researcher must ensure that neither personal
biases nor vested interests influence any stage of the research process [102, 105].
Journaling personal reflections during the process allows the researcher to challenge
53
how their own perceptions and interpretations may influence the findings. No peer
reviewers were used during study analysis, a limitation which may have some effect on
results. Lastly, this study acknowledges that generalizability of findings is not feasible in
IPA studies as participants are selected based on their individualized experiences (i.e.
with unconscious bias training) rather than for their ability to represent the perceptions
of a larger population [102, 103].
Results
While no self-identified demographic data was collected, study participants were
asked to describe themselves, where they live and practice. Urban metropolitan areas
were well represented and included cities such as Atlanta, San Francisco, and Chicago.
Only one participant described their location, Salem, OH, as “semi-rural/suburban”. The
majority of respondents were employed physicians serving in academic roles such as
residency faculty (4) or in direct patient care roles such as medical directors (4). Of the
two participants who indicated either sole or partial ownership of their practice, one was
a direct primary care (DPC) provider. Among those in direct patient care roles, their
practice settings included federally qualified heath centers, community hospitals (5),
government and outpatient clinics. Where indicated, scope of practice included
references to sports medicine (1) and women’s health (2), including labor and delivery.
Lastly, two participants indicated they were either active or recently retired military
physicians and a third indicated they were currently serving as a chief resident.
Analysis identified five major themes. First, personal resistance to the
insinuation of unconscious bias. Participants in both groups frequently indicated that
just discussing the topic initially triggered feelings of defensiveness, discomfort and
54
vulnerability that led them to reject the notion that they themselves were biased and or
question the validity of assessment tools such as the implicit association test. For
example, upon first learning about unconscious bias, one participant responded that,
“when I first heard about it and when I asked people about it…..I became very defensive
until you start learning about it”. They also continued to further elaborate on this feeling
of defensiveness in discussing it with other individuals. Another described “multiple
levels of vulnerability” associated with feeling as though as physicians “we’re supposed
to have the insight” however “I’m admitting that I’m unable to do anything about it at this
point”. Lastly, discomfort was acknowledged as an important part of the learning
process, where in one participant stated, “because it’s hard to talk about, you have to
sort of agree we’re going to be uncomfortable with the topic. We’re all going to have to
just assume good intent and be able to say what we want to say”. These findings are
consistent with previous studies examining reactions to unconscious bias among
medical students, which indicates a need to create a learning environment that is safe
and inclusive for learners at all stages of the medical education continuum prior to
engaging in potentially sensitive and emotionally charged discussions which may
quickly derail the goals of unconscious bias training [59, 60, 64, 106]. Ultimately, for the
group of participants who had never participated in an unconscious bias training, it was
for these reasons cited, in addition to the fact that for some, this was their first time
encountering the concept.
Despite these initial reactions, there was a consensus between the participants in
each group that indicated a desire to be responsive and proactive to the issue of
unconscious bias, primarily because of personal accounts from those having
55
experienced unconscious bias firsthand. This responsiveness is often characterized
with an acknowledgment of its influence on patients but followed up with a diverse set of
perspectives regarding how it should be addressed. First, those who had participated in
an unconscious bias training expressed overwhelmingly that unconscious bias training
should be required on an annual basis starting in medical school and into continuing
medical education. Across both groups there were those who suggested that individual
training was insufficient, and some cited the need for interactions engaging members of
the community. For example, one participant indicated that to minimize resistance in
their residency program, “we bring in people from the community that talk about what
it’s like to be them” and similarly another indicated that a portion of their curriculum
included “instruction by volunteers from the community”. This perspective is consistent
with literature suggesting that unconscious bias training should incorporate social
interactions with what are known as “counter stereotypical exemplars” to increase
empathy and obtain a greater awareness from the perspective of others affected by it
[32, 37, 107]. Lastly, both groups expressed a desire for greater organizational
accountability. Feelings of ineptitude and questions of effectiveness with regards to
unconscious bias training were countered in some instances with statements such as
“…individual action is one part of it, but I think the organizations and institutions, that
action is more important”. Likewise, another participant emphasized, “….it has to be in
the culture. It has to be mandatory. You know, lectures are not going to help”. Both
statements suggest that interventions designed to target workplace cultures could
potentially play a more significant role in reducing individuals’ unconscious biases. The
concept of unconscious bias in an organizational context warrants further examination
56
to understand how systems, policies and procedures potentially form or reinforce
individual biases.
Next, participants who had indicated they had attended an unconscious bias
training within the last 12 months engaged in a robust conversation discussing their
reasons for participating. Four subthemes quickly emerged; personal development,
curricular requirements for medical education or training, professional development and
employer mandated. First, those who indicated reasons associated with personal
development often referred to a desire for self-improvement or to gain a greater level of
understanding. For example, one participant remarked, “to better understand what I
could do to better recognize implicit frames that may impact my understanding and
relationship with others”. Those who were in academic faculty roles indicated they did
so not only for their own “personal enrichment” but also to apply to educating students
and training residents. Several indicated they were required to participate in
unconscious bias training as a standard part of their medical education and or residency
training. The majority, nearly half of respondents, indicated their participation was
associated with their professional development goals. For example, references to
continuing medical education credits, additional graduate degree programs (i.e.
healthcare administration, public health, etc.) and leadership development courses.
Lastly, those who indicated their participation was the result of an employer mandate
sometimes referred to it as “corporate policy” or a prerequisite to participating in certain
activities such as clinical case reviews or interviewing applicants for medical school
admissions. Future research should examine ways in which these reasons for
57
participation are influenced by personal factors (i.e. gender, age, stage of career, etc.)
as well as organizational factors (i.e. policies, promotions, etc.).
In addition, this group discussed extensively the features of unconscious bias
training that resonate with physicians. These included the normalization of bias, the
neuropsychological research explaining the formation of bias and the use of medically
relevant case studies demonstrating how it shows up in clinical practice. For
participants, normalizing bias reduced the defensiveness, shame, and vulnerability
around it. Statements such as “…it's normal human behavior, not a flaw” and being
reminded of “how pervasive” it is were associated with reduced resistance to additional
training constructs. Understanding the neuropsychological science, such as how bias is
formed and triggered in the brain, helped to legitimize it, and providing case examples
made it relatable to their previous medical education and training. For example,
statements such as providing “case studies and scenarios with practice and feedback”
and “case examples of how bias affects us” were deemed the most relevant part of
training. Trainings in which these three features were absent were frequently described
as effective at “raising awareness” but provided no “concrete tools” or “strategies” to
address them. Understanding what works and what doesn’t may lead to the
development of some core standards for unconscious bias training, for which currently
there are none and may be a factor contributing to the mixed findings regarding its
effectiveness.
Lastly, this group of participants shared how they were applying what they
learned to practice what they learned after participating in training. The results of
training can be sorted into two categories, personal and practice outcomes. Personal
58
outcomes, such as references to gaining increased “cultural humility”, “self-awareness”
and “objectivity” reflect participants’ intentions for personal development. Likewise, an
increase in patient centeredness was reflected by one participant who indicated, “I strive
to put the patient’s goals for their health first. That seems to bring down any bias”.
Others cited more deliberate actions related to changes in practice behaviors. Practice
outcomes included an increased use of in-person interpreters to provide cultural
background, developing customized individual treatment plans and increased
communication with patients. Future research should consider the use of observational
methods (i.e. ethnography, simulations, etc.) to examine post training behaviors
between physicians and patients to determine its potential impact.
Discussion Though small, the focus group participants represent a rather comprehensive
cross-section of the broader AAFP membership [98]. Findings demonstrate that the
reactions of practicing physicians towards unconscious bias training are rather
consistent with other populations. There seems to be a great deal of variability in
perspectives regarding how to respond effectively to unconscious bias and participants
reasons for participating in trainings that warrant additional research that may lead to
more effective interventions at levels beyond just targeting individuals. These findings
have demonstrated that physicians have a clear indication of what works and what
doesn’t from their perspective as the target audience and those factors should be taken
into consideration when developing unconscious bias training interventions.
In addition, these findings closely resemble constructs proposed by the
Transtheoretical (Stages of Change) Model in which participants cycle through decision
59
making check-points that ultimately result in maintenance of a change in behavior [108].
In this case the behavior is mitigation of unconscious bias. This model could potentially
be applied by 1) developing self-assessment alternatives to the IAT that more
objectively indicate which stage individuals enter into the process (i.e. awareness,
resistance, responsiveness, etc.), 2) assigning individuals into peer learning cohorts that
reduce the negative feelings associated with their unconscious biases (i.e.
defensiveness, shame, etc.) and 3) establishing learning objectives that move
individuals from one stage to the next until they reach “Maintenance” or sustained
behaviors that ultimately reduce the impact of unconscious bias (Figure 4).
Figure 4 Stages of change towards unconscious bias self-mitigation
For physicians, mitigating unconscious bias should be considered as a
deliberate, ongoing process that requires self-awareness and self-regulation where
individuals check in with themselves on a regular basis to ensure that they are acting
and making decisions based on a rational assessment of clinical situations rather than
on stereotypes and prejudices. Likewise, organizations should begin to consider ways
in which established cultural norms are associated with the unconscious biases of
Awareness
Resistance
Responsiveness
Action
Maintenance
60
individuals and identify opportunities to develop more inclusive environments which may
have implications for patient health. As suggested little attention has been given to
identifying and or implementing interventions at this level. Equal if not greater priority
should be given to developing organizational assessments of policies, systems and
procedures that may potentially be reinforcing unconscious bias and develop
sustainable actions that can be used to mitigate it.
While the strength of the methods used in this study have been outlined, several
limitations remain. First, selection bias for targeting members of the AAFP must be
acknowledged. Potentially, there are unknown factors associated with choice of
medical specialty, membership in a physician association and attendance at annual
meetings that may also be associated with the values, perspectives and behaviors
associated with this topic. Future studies should broaden the scope of study
participants across various medical specialties and health care professions to limit the
influence of selection bias on outcomes.
A second limitation considers risk to validity given that this study relies on recall
and self-reported attitudes and behaviors from focus groups which may produce results
that differ from the natural observation of participants. The cognitive overload health
care professionals experience in the workplace as a result of things like stress, burnout
and compassion fatigue may alter their behaviors in the practice setting and their
perceptions of those behaviors, thus negating their self-reports [58]. One way to
address this limitation is by conducting future studies that utilize an ethnographic
approach to observe demonstrations of behaviors associated with unconscious bias in
health care culture.
61
Lastly, the segmentation of the study into two focus groups in which one group
has participated in an unconscious bias intervention and the other has not essentially
only provides one set of data for each segment. A general rule of thumb in conducting
qualitative research using focus groups is that three to five groups are adequate to
reach saturation – the point at which additional data collection no longer generates new
knowledge – and increasing the number of segments and or variability of the
participants within and across groups requires more groups. Due to budget constraints
($2650 per focus group), size of the research team (principle investigator, moderator
and recorder) and the limited availability of participants (annually), the feasibility of
conducting more than two focus groups at this time and in this location is just not
plausible. However, the findings of this study should sufficiently justify the need for
future research in this area and guide the development and implementation of future
methods.
Conclusion
Though physicians are often the target of unconscious bias interventions, there’s
little evidence examining experiences with it from their perspectives, which potentially
has implications for patient care. While some might outright reject the validity of
unconscious bias, for others, there is a desire to be responsive and take actions that
mitigate its effects. There are both personal and professional factors driving
participation in unconscious bias training interventions, with some indication as to the
influence of organizations. Organizations have a responsibility to their employees to
examine ways in which workplace culture could be reinforcing unconscious bias and
identify relevant interventions. Additional research is needed to further examine both
62
individual and organizational constructs to which more effective interventions to address
unconscious bias can be developed to reduce health care disparities.
63
Chapter 6 – Implications
64
As indicated, this dissertation set out with the aim to examine the prevalence of
unconscious bias within the healthcare workforce and conceptualize ways in which it
may potentially result in disparate health outcomes. As these findings suggest, racial
unconscious bias among the healthcare workforce differs significantly from that of the
general public and can vary by geographic and provider type. This may provide some
indication as to why unconscious bias educational training interventions have not yet
demonstrated their effectiveness as they are incapable of hitting a moving target. In
addition, these findings suggest that individual factors like social identity may have less
of an association to unconscious biases as previously thought and suggest that there is
potentially a greater need to examine external factors associated with the workplace
and community. This perspective of greater organizational responsibility and
accountability is shared by physicians however additional research is needed to
examine unconscious bias in an organizational context and its potential impact on both
physicians and patients.
More broadly, these findings may also be considered within the context of
debates within the social sciences regarding the primacy of individual agency versus
structure [109] . As it relates to healthcare, the question here is, “are the racially
disparate clinical decisions of providers the result of their own individual autonomy,
unconscious or otherwise, or socialization within a system of healthcare with norms,
customs, policies, etc. designed from its inception to marginalize and minimize the
healthcare needs of racial minorities”. These findings suggest the latter, and further
justify the need to develop and implement interventions that focus on healthcare
65
systems and culture instead of individuals to reduce healthcare disparities effectively
and sustainably. Furthermore, the following implications should be considered.
Policy Several State and Federal policy actions supporting widespread implementation
of unconscious bias education and training interventions have been proposed. For
example, the MOMMAs Act (S.3776/H.R.5977) was introduced in 2018 and called for
the establishment of regional centers of excellence to address unconscious bias and
cultural competency in patient-provider interactions. These centers are intended to
improve how health care professionals are educated on unconscious bias and the
delivery of culturally competent health care. In addition, the Maternal Care Access, and
Reducing Emergencies (CARE) Act (S.3363/H.R.6698) focuses specifically on
institutional racism by providing funding for the implementation of unconscious bias
training programs for clinicians and evidence-based culturally proficient support
programs and services for pregnant women. In 2019, California was the first state to
approve a bill (AB-241) that requires unconscious bias training for health care
professionals, law enforcement and judicial employees. Several other states have
proposed similar legislation.
These findings have demonstrated that within the healthcare workforce
unconscious bias is highly variable and may also be influenced by other external
factors, for which educational training interventions are not yet designed to address. As
such, future legislative actions should consider placing a greater emphasis on continued
research as opposed to mandates for an intervention that has not yet been proven
efficacious nor meets the necessary standards to be considered evidence-based.
Furthermore, continuing medical education mandates are often controversial, do not
66
always result in practice changes and in some instances can create barriers to licensure
and certification [110, 111]. However, there still exist the need to create an overall
sense of awareness and acknowledgement of the potential impact of unconscious bias
to influence disparate outcomes, which is the responsibility of those organizations who
oversee medical education and training. As these findings have demonstrated,
organizations may have some influence on physicians’ decisions to address their
unconscious biases.
Education As mentioned, regardless of efficacy, unconscious bias educational training
interventions are being implemented and should be done so as just one approach of a
larger strategy to address health care disparities. These training serve to raise
awareness and an acknowledgement of the potential impact on patients. However, as
these findings indicate, those should be implemented in a way that takes into
consideration the variability of bias, influential factors on the individual learner, potential
reactions and responses and training elements that are most applicable to those
responsible for patient care. For faculty who design and deliver these programs there is
often little guidance to prepare them to effectively manage these issues which is
problematic for two reasons. First, unconscious bias educational trainings often involve
emotionally charged discussions involving race, systematic oppression and social
inequities which can quickly disrupt the learning environment and derail course
objectives [106]. Second, it suggests a lack of standards in how training should be
conducted which allows for a significant amount of variability when it comes to
objectives, formats and style which presents a challenge to examining their
67
effectiveness. There is a need for a pedagogical approach to teaching unconscious
bias to health care professionals that develops both skilled facilitators and learners.
68
Bibliography
69
1. Institute of Medicine Committee on, U., R. Eliminating, and C. Ethnic Disparities in Health, in Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care, B.D. Smedley, A.Y. Stith, and A.R. Nelson, Editors. 2003, National Academies Press (US) Copyright 2002 by the National Academy of Sciences. All rights reserved.: Washington (DC).
2. Miller, D., Tech companies spend big money on bias training-but it hasn’t improved diversity numbers. The Conversation: US Pilot, 2015.3. Zelevansky, N., The Big Business of Unconscious Bias. The New York Times, 2019.
4. Crawford, C., From the American Academy of Family Physicians: THE EVERYONE PROJECT UNVEILS IMPLICIT BIAS TRAINING GUIDE. Annals of Family Medicine, 2020. 18(2): p. 182.
5. Milam, A.J., et al., Are Clinicians Contributing to Excess African American COVID-19 Deaths? Unbeknownst to Them, They May Be. Health Equity, 2020. 4(1): p. 139-141.
6. Williams, D.R. and L.A. Cooper, COVID-19 and Health Equity—A New Kind of “Herd Immunity”. JAMA, 2020. 323(24): p. 2478-2480.
7. Eligon, J., Black doctor dies of Covid-19 after complaining of racist treatment. New York Times, 2020.
8. Blair, I.V., et al., Assessment of biases against Latinos and African Americans among primary care providers and community members. Am J Public Health, 2013. 103(1): p. 92-8.
9. West, C.P., L.N. Dyrbye, and T.D. Shanafelt, Physician burnout: contributors, consequences and solutions. J Intern Med, 2018. 283(6): p. 516-529.
10. Roberts, D.E., Is race-based medicine good for us?: African American approaches to race, biomedicine, and equality. 2008, SAGE Publications Sage CA: Los Angeles, CA.
11. Tsai, J. and A. Crawford-Roberts, A Call for Critical Race Theory in Medical Education. Acad Med, 2017. 92(8): p. 1072-1073.
12. Yu, J.H., S. Goering, and S.M. Fullerton, Race-based medicine and justice as recognition: exploring the phenomenon of BiDil. Camb Q Healthc Ethics, 2009. 18(1): p. 57-67.
13. Kumar, R., et al., Cultural competence in family practice and primary care setting. J Family Med Prim Care, 2019. 8(1): p. 1-4.
14. Hoffman, K.M., et al., Racial bias in pain assessment and treatment recommendations, and false beliefs about biological differences between blacks and whites. Proc Natl Acad Sci U S A, 2016. 113(16): p. 4296-301.
15. Daugherty, S.L., et al., Implicit Gender Bias and the Use of Cardiovascular Tests Among Cardiologists. J Am Heart Assoc, 2017. 6(12).
16. Kogan, M.D., et al., Racial disparities in reported prenatal care advice from health care providers. Am J Public Health, 1994. 84(1): p. 82-8.
17. Peek, M.E., et al., Development of a conceptual framework for understanding shared decision making among African-American LGBT patients and their clinicians. Journal of general internal medicine, 2016. 31(6): p. 677-687.
18. White-Means, S., et al., Cultural competency, race, and skin tone bias among pharmacy, nursing, and medical students: implications for addressing health disparities. Med Care Res Rev, 2009. 66(4): p. 436-55.
70
19. Chapman, E.N., A. Kaatz, and M. Carnes, Physicians and implicit bias: how doctors may unwittingly perpetuate health care disparities. J Gen Intern Med, 2013. 28(11): p. 1504-10.
20. Phelan, S.M., et al., Medical School Factors Associated with Changes in Implicit and Explicit Bias Against Gay and Lesbian People among 3492 Graduating Medical Students. J Gen Intern Med, 2017. 32(11): p. 1193-1201.
21. Phelan, S.M., et al., The Effects of Racism in Medical Education on Students' Decisions to Practice in Underserved or Minority Communities. Acad Med, 2019. 94(8): p. 1178-1189.
22. Hardeman, R.R., et al., Developing a Medical School Curriculum on Racism: Multidisciplinary, Multiracial Conversations Informed by Public Health Critical Race Praxis (PHCRP). Ethn Dis, 2018. 28(Suppl 1): p. 271-278.
23. Sabin, J., et al., Physicians' implicit and explicit attitudes about race by MD race, ethnicity, and gender. J Health Care Poor Underserved, 2009. 20(3): p. 896-913.
24. Sabin, J.A., M. Marini, and B.A. Nosek, Implicit and explicit anti-fat bias among a large sample of medical doctors by BMI, race/ethnicity and gender. PLoS One, 2012. 7(11): p. e48448.
25. Cooper, L.A.a.P., N.R., Disparities in Patient Experiences, Health Care Processes, and Outcomes: The Role of Patient-Provider Racial, Ethnic, and Language Concordance. The Common Wealth Fund, 2004.
26. Poma, P.A., Race/Ethnicity Concordance Between Patients and Physicians. J Natl Med Assoc, 2017. 109(1): p. 6-8.
27. FitzGerald, C. and S. Hurst, Implicit bias in healthcare professionals: a systematic review. BMC Med Ethics, 2017. 18(1): p. 19.
28. Hall, W.J., et al., Implicit Racial/Ethnic Bias Among Health Care Professionals and Its Influence on Health Care Outcomes: A Systematic Review. Am J Public Health, 2015. 105(12): p. e60-76.
29. Dehon, E., et al., A Systematic Review of the Impact of Physician Implicit Racial Bias on Clinical Decision Making. Acad Emerg Med, 2017. 24(8): p. 895-904.
30. Greenwald, A.G. and B.A. Nosek, Health of the Implicit Association Test at age 3. Z Exp Psychol, 2001. 48(2): p. 85-93.
31. Greenwald, A.G., et al., Understanding and using the Implicit Association Test: III. Meta-analysis of predictive validity. J Pers Soc Psychol, 2009. 97(1): p. 17-41.
32. FitzGerald, C., et al., Interventions designed to reduce implicit prejudices and implicit stereotypes in real world contexts: a systematic review. BMC Psychol, 2019. 7(1): p. 29.
33. Woodcock, A. and M.J. Monteith, Forging links with the self to combat implicit bias. Group Processes & Intergroup Relations, 2013. 16(4): p. 445-461.
34. Devine, P.G., et al., Long-term reduction in implicit race bias: A prejudice habit- breaking intervention. J Exp Soc Psychol, 2012. 48(6): p. 1267-1278.
35. Wallaert, M., A. Ward, and T. Mann, Explicit control of implicit responses: Simple directives can alter IAT performance. Social Psychology, 2010. 41(3): p. 152- 157.
36. Swift, J.A., et al., Are anti-stigma films a useful strategy for reducing weight bias among trainee healthcare professionals? Results of a pilot randomized control trial. Obes Facts, 2013. 6(1): p. 91-102.
71
37. Sukhera, J. and C. Watling, A Framework for Integrating Implicit Bias Recognition Into Health Professions Education. Acad Med, 2018. 93(1): p. 35-40.
38. Boscardin, C.K., Reducing Implicit Bias Through Curricular Interventions. J Gen Intern Med, 2015. 30(12): p. 1726-8.
39. Shea, C.M., et al., Contextual factors that influence quality improvement implementation in primary care: The role of organizations, teams, and individuals. Health Care Manage Rev, 2018. 43(3): p. 261-269.
40. Coles, E., et al., The influence of contextual factors on healthcare quality improvement initiatives: what works, for whom and in what setting? Protocol for a realist review. Syst Rev, 2017. 6(1): p. 168.
41. Shelton, R.C., D.M. Griffith, and M.C. Kegler, The Promise of Qualitative Research to Inform Theory to Address Health Equity. Health Educ Behav, 2017. 44(5): p. 815-819.
42. Griffith, D.M., R.C. Shelton, and M. Kegler, Advancing the Science of Qualitative Research to Promote Health Equity. Health Educ Behav, 2017. 44(5): p. 673- 676.
43. Morgan, D.L., Practical strategies for combining qualitative and quantitative methods: applications to health research. Qual Health Res, 1998. 8(3): p. 362- 76.
44. Institute of Medicine Committee on Understanding Premature, B. and O. Assuring Healthy, The National Academies Collection: Reports funded by National Institutes of Health, in Preterm Birth: Causes, Consequences, and Prevention, R.E. Behrman and A.S. Butler, Editors. 2007, National Academies Press (US) National Academy of Sciences.: Washington (DC).
45. Satcher, D., et al., What if we were equal? A comparison of the black-white mortality gap in 1960 and 2000. Health Affairs, 2005. 24(2): p. 459-464.
46. Peckham, C., Medscape lifestyle report 2016: bias and burnout. New York, NY: Medscape, 2016.
47. Sabin, J.A., F.P. Rivara, and A.G. Greenwald, Physician implicit attitudes and stereotypes about race and quality of medical care. Med Care, 2008. 46(7): p. 678-85.
48. Gatewood, E., et al., Making the invisible visible: Implementing an implicit bias activity in nursing education. J Prof Nurs, 2019. 35(6): p. 447-451.
49. Bucknor-Ferron, P. and L. Zagaja, Five strategies to combat unconscious bias. Nursing, 2016. 46(11): p. 61-62.
50. Sadarangani, T.R., The nurse's role in promoting health equity and improving racial justice in older adults through elimination of unconscious bias. Geriatr Nurs, 2020.
51. Schultz, P.L. and J. Baker, Teaching Strategies to Increase Nursing Student Acceptance and Management of Unconscious Bias. J Nurs Educ, 2017. 56(11): p. 692-696.
52. Narayan, M.C., CE: Addressing Implicit Bias in Nursing: A Review. Am J Nurs, 2019. 119(7): p. 36-43.
53. Persaud, S., Addressing Unconscious Bias: A Nurse Leader's Role. Nurs Adm Q, 2019. 43(2): p. 130-137.
72
54. Bosse, J., et al., Position statement: Full practice authority for advanced practice registered nurses is necessary to transform primary care. Nursing outlook, 2017. 65(6): p. 761-765.
55. Xue, Y., et al., Impact of state nurse practitioner scope-of-practice regulation on health care delivery: Systematic review. Nurs Outlook, 2016. 64(1): p. 71-85.
56. MacLeod, M.L.P., et al., Rural and Remote Registered Nurses' Perceptions of Working Beyond Their Legislated Scope of Practice. Nurs Leadersh (Tor Ont), 2019. 32(1): p. 20-29.
57. Poghosyan, L., A.A. Norful, and M.J. Laugesen, Removing restrictions on nurse practitioners' scope of practice in New York State: Physicians' and nurse practitioners' perspectives. J Am Assoc Nurse Pract, 2018. 30(6): p. 354-360.
58. Burgess, D.J., M.C. Beach, and S. Saha, Mindfulness practice: A promising approach to reducing the effects of clinician implicit bias on patients. Patient Educ Couns, 2017. 100(2): p. 372-376.
59. Gonzalez, C.M., et al., How to Make or Break Implicit Bias Instruction: Implications for Curriculum Development. Acad Med, 2018. 93(11S Association of American Medical Colleges Learn Serve Lead: Proceedings of the 57th Annual Research in Medical Education Sessions): p. S74-s81.
60. Gonzalez, C.M., et al., A Qualitative Study of New York Medical Student Views on Implicit Bias Instruction: Implications for Curriculum Development. J Gen Intern Med, 2019. 34(5): p. 692-698.
61. Byrne, A. and A. Tanesini, Instilling new habits: addressing implicit bias in healthcare professionals. Adv Health Sci Educ Theory Pract, 2015. 20(5): p. 1255-62.
62. Stone, J., et al., Testing active learning workshops for reducing implicit stereotyping of Hispanics by majority and minority group medical students. Stigma and Health, 2019: p. No Pagination Specified-No Pagination Specified.
63. Salm Ward, T.C., et al., "You learn to go last": perceptions of prenatal care experiences among African-American women with limited incomes. Matern Child Health J, 2013. 17(10): p. 1753-9.
64. Gonzalez, C.M., M.Y. Kim, and P.R. Marantz, Implicit bias and its relation to health disparities: a teaching program and survey of medical students. Teach Learn Med, 2014. 26(1): p. 64-71.
65. Greenwald, A.G., D.E. McGhee, and J.L. Schwartz, Measuring individual differences in implicit cognition: the implicit association test. J Pers Soc Psychol, 1998. 74(6): p. 1464-80.
66. Xu, K., B. Nosek, and A. Greenwald, Psychology data from the race implicit association test on the project implicit demo website. Journal of Open Psychology Data, 2014. 2(1).
67. Hehman, E., et al., Establishing construct validity evidence for regional measures of explicit and implicit racial bias. J Exp Psychol Gen, 2019. 148(6): p. 1022- 1040.
68. Nosek, B.A., et al., Understanding and using the brief Implicit Association Test: recommended scoring procedures. PloS one, 2014. 9(12): p. e110938-e110938.
69. Lipka, M., 5 key findings about the changing US religious landscape. Pew Research Center. May, 2015. 12.
73
70. Tajeu, G.S., et al., Exploring the Association of Healthcare Worker Race and Occupation with Implicit and Explicit Racial Bias. J Natl Med Assoc, 2018. 110(5): p. 464-472.
71. Klein, C., Confidence intervals on Implicit Association test scores are really rather large. 2020.
72. Feagin, J. and Z. Bennefield, Systemic racism and U.S. health care. Social Science & Medicine, 2014. 103: p. 7-14.
73. Griffith, D.M., et al., Dismantling institutional racism: theory and action. Am J Community Psychol, 2007. 39(3-4): p. 381-92.
74. Jones, C.P., Levels of racism: a theoretic framework and a gardener's tale. Am J Public Health, 2000. 90(8): p. 1212-5.
75. Bloche, M.G., Race-based therapeutics. New England Journal of Medicine, 2004. 351(20): p. 2035-2037.
76. Braun, L., Theorizing Race and Racism: Preliminary Reflections on the Medical Curriculum. Am J Law Med, 2017. 43(2-3): p. 239-256.
77. Ford, C.L. and C.O. Airhihenbuwa, The public health critical race methodology: praxis for antiracism research. Soc Sci Med, 2010. 71(8): p. 1390-8.
78. Bourke, C.J., H. Marrie, and A. Marrie, Transforming institutional racism at an Australian hospital. Aust Health Rev, 2018.
79. Jones, D.D., Differentiating the Unconscious Racial Biases and Attitudes of Physicians, Nurses, and the Public: Implications for Future Healthcare Education and Practice. Unpublished Manuscript 2021.
80. Hart, A.J., et al., Differential response in the human amygdala to racial outgroup vs ingroup face stimuli. Neuroreport, 2000. 11(11): p. 2351-2354.
81. Davy, Z., The promise of intersectionality theory in primary care. Qual Prim Care, 2011. 19(5): p. 279-81.
82. Bauer, G.R., Incorporating intersectionality theory into population health research methodology: Challenges and the potential to advance health equity. Social Science & Medicine, 2014. 110: p. 10-17.
83. Petterson, S.M., et al., Projecting US primary care physician workforce needs: 2010-2025. Ann Fam Med, 2012. 10(6): p. 503-9.
84. Chopik, W.J. and H.L. Giasson, Age Differences in Explicit and Implicit Age Attitudes Across the Life Span. The Gerontologist, 2017. 57(suppl_2): p. S169- S177.
85. Lichter, D.T., Immigration and the new racial diversity in rural America. Rural Sociology, 2012. 77(1): p. 3-35.
86. Eden, A.R., et al., Burnout among family physicians by gender and age. The Journal of the American Board of Family Medicine, 2020. 33(3): p. 355-356.
87. Adkins-Jackson, P.B., R.K. Legha, and K.A. Jones, How to Measure Racism in Academic Health Centers. AMA Journal of Ethics, 2021. 23(2): p. 140-145.
88. Orchard, J. and J. Price, County-level racial prejudice and the black-white gap in infant health outcomes. Soc Sci Med, 2017. 181: p. 191-198.
89. Haider, A.H., et al., Unconscious race and class bias: its association with decision making by trauma and acute care surgeons. J Trauma Acute Care Surg, 2014. 77(3): p. 409-16.
74
90. Haider, A.H., et al., Unconscious race and social class bias among acute care surgical clinicians and clinical treatment decisions. JAMA Surg, 2015. 150(5): p. 457-64.
91. Raphael, J.L. and S.O. Oyeku, Implicit Bias in Pediatrics: An Emerging Focus in Health Equity Research. Pediatrics, 2020. 145(5).
92. Forscher, P.S., et al., A Meta-Analysis of Procedures to Change Implicit Measures. 2019.
93. Matthew, D.B., A new strategy to combat racial inequality in American health care delivery. DePaul J Health Care Law, 2005. 9(1): p. 793-853.
94. Bowen Matthew, D., Toward a Structural Theory of Implicit Racial and Ethnic Bias in Health Care. Health Matrix Clevel, 2015. 25: p. 61-85.
95. Matthew, D.B., Just medicine: A cure for racial inequality in American health care. 2018: NYU Press.
96. Jones, D.D., The Personal, Practice and Community Characteristics of Family Medicine Physicians Engaged in Unconscious Bias Mitigation Activities Unpublished Manuscript, 2021.
97. Rui, P. and T. Okeyode, National ambulatory medical care survey: 2016 national summary tables. Centers for Disease Control and Prevention, Atlanta, GA: https://www. cdc. ov/nchs/data/ahcd/namcs_summary/2016_namcs_web_tables. pdf.[Google Scholar], 2016.
98. Physicians, A.A.o.F., Facts about: family practice. 2019: American Academy of Family Physicians.
99. Guest, G., et al., Comparing focus groups and individual interviews: findings from a randomized study. International Journal of Social Research Methodology, 2017. 20(6): p. 693-708.
100. Wood, M.L., Focus Group Interview in Family Practice Research: Implementing a qualitative research method. Can Fam Physician, 1992. 38: p. 2821-7.
101. Priest, H., An approach to the phenomenological analysis of data. Nurse Researcher (through 2013), 2002. 10(2): p. 50-63.
102. Peat, G., A. Rodriguez, and J. Smith, Interpretive phenomenological analysis applied to healthcare research. Evidence - Based Nursing, 2019. 22(1): p. 7.
103. Smith, J.A., M. Jarman, and M. Osborn, Doing interpretative phenomenological analysis. Qualitative health psychology: Theories and methods, 1999: p. 218- 240.
104. Carter, N., et al., The use of triangulation in qualitative research. Oncol Nurs Forum, 2014. 41(5): p. 545-7.
105. Delaney, N. and E. McCann, A phenomenological exploration of transgender people's experiences of mental health services in Ireland. J Nurs Manag, 2021. 29(1): p. 68-74.
106. Wear, D. and J.M. Aultman, The limits of narrative: medical student resistance to confronting inequality and oppression in literature and beyond. Med Educ, 2005. 39(10): p. 1056-65.
107. Lai, C.K., et al., Reducing implicit racial preferences: I. A comparative investigation of 17 interventions. J Exp Psychol Gen, 2014. 143(4): p. 1765-85.
108. Prochaska, J.O., S. Johnson, and P. Lee, The transtheoretical model of behavior change. 2009.
75
109. Barker, C., Cultural studies: Theory and practice. 2003: Sage. 110. Miller, S.H., et al., Continuing medical education, professional development, and
requirements for medical licensure: a white paper of the Conjoint Committee on Continuing Medical Education. J Contin Educ Health Prof, 2008. 28(2): p. 95-8.
111. Donen, N., No to mandatory continuing medical education, yes to mandatory practice auditing and professional educational development. Cmaj, 1998. 158(8): p. 1044-6.
76
Appendices
77
Appendix A: Overview of Studies Associating Unconscious Bias to Patient Outcomes
A p
p e
n d
ix A
O
v e rv
ie w
o f
S tu
d ie
s E
x a
m in
in g
t h
e A
s s
o c
ia ti
o n
o f
U .S
. H
e a
lt h
c a
re P
ro fe
s s
io n
a ls
' Im
p li
c it
B ia
s a
n d
P a
ti e
n t
H e
a lt
h O
u tc
o m
e s
S
a m
p le
P
a ti
e n
t T
y p
e
M e th
o d
R e
v ie
w
A u
th o
r Y
e a
r W
h o
S
iz e
M
e th
o d
H
y p
o th
e ti
c a
l o
r R
e a
l S
tu d
y
D e
s ig
n
IB
M e a
s u
re
IB T
y p
e
A s
s o
c ia
ti o
n
w /H
e a lt
h
O u
tc o
m e
s
H A
L L
B
la ir
2 0 1 3
P
ri m
a ry
C a
re
P h
y s ic
ia n
s
2 1 0
c o
n v e
n ie
n c e
H
y p
o th
e ti c a
l
C ro
s s -
s e
c ti o
n a
l IA
T
R a
c e
N
A
H A
L L
B
la ir
2 0 1 3
P
ri m
a ry
C a
re
P h
y s ic
ia n
s
1 3 8
C
o n
v e
n ie
n c e
R
e a
l C
ro s s -
s e
c ti o
n a
l IA
T
R a
c e
N
o n
e
H A
L L
B
la ir
2 0 1 4
P
ri m
a ry
C a
re
P ro
v id
e rs
1 3 8
C
o n
v e
n ie
n c e
R
e a
l C
ro s s -
s e
c ti o
n a
l IA
T
R a
c e
N
o n
e
H A
L L
C
o o
p e
r 2 0 1 2
P
ri m
a ry
C a
re
P ro
v id
e rs
4 0
C
o n
v e
n ie
n c e
R
e a
l C
ro s s -
s e
c ti o
n a
l IA
T
R a
c e
a n
d
M e d
ic a
l C
o m
p lia
n c e
C o
m m
u n
ic a
ti o
n
H A
L L
H
a g
iw a
ra
2 0 1 3
P
ri m
a ry
C a
re
P h
y s ic
ia n
s
1 4
C
o n
v e
n ie
n c e
R
e a
l C
ro s s -
s e
c ti o
n a
l IA
T
R a
c e
C
o m
m u
n ic
a ti o
n
H A
L L
H
a u
s m
a n
n
2 0 1 5
P
h y s ic
ia n
s
1 4
C
o n
v e
n ie
n c e
R
e a
l C
ro s s -
s e
c ti o
n a
l IA
T
R a
c e
S
p in
a l C
o rd
I n ju
ry
o u tc
o m
e s
H A
L L
P
e n
n e r
2 0 1 0
P
ri m
a ry
C a
re
P h
y s ic
ia n
s
1 5
C
o n
v e
n ie
n c e
R
e a
l C
ro s s -
s e
c ti o
n a
l IA
T
R a
c e
In
te ra
c ti o
n s
H A
L L
S
a b
in
2 0 0 8
P
e d
ia tr
ic ia
n s
9 5
C
o n
v e
n ie
n c e
H
y p
o th
e ti c a
l C
ro s s -
s e
c ti o
n a
l IA
T
R a
c e
A
tt it u
d e
, C
o m
p lia
n t
P a
ti e
n t,
Q
u a lit
y
n o n e
H A
L L
S
a b
in
2 0 1 2
P
e d
ia tr
ic ia
n s
8 6
C
o n
v e
n ie
n c e
H
y p
o th
e ti c a
l C
ro s s -
s e
c ti o
n a
l IA
T
R a
c e
, M
e d
ic a
l C
o m
p lia
n c e
, Q
u a lit
y
T re
a tm
e n
t
78
A p
p e
n d
ix A
O
v e rv
ie w
o f
S tu
d ie
s E
x a
m in
in g
t h
e A
s s
o c
ia ti
o n
o f
U .S
. H
e a
lt h
c a
re P
ro fe
s s
io n
a ls
' Im
p li
c it
B ia
s a
n d
P a
ti e
n t
H e
a lt
h O
u tc
o m
e s
S
a m
p le
P
a ti
e n
t T
y p
e
M e th
o d
R e
v ie
w
A u
th o
r Y
e a
r W
h o
S
iz e
M
e th
o d
H
y p
o th
e ti
c a
l o
r R
e a
l S
tu d
y
D e
s ig
n
IB
M e a
s u
re
IB T
y p
e
A s
s o
c ia
ti o
n
w /H
e a lt
h
O u
tc o
m e
s
F it z g
e ra
ld
B a
rn h
a rt
2 0 0 6
P
h y s ic
ia n
s
5 4 4
R
a n
d o
m
H y p
o th
e ti c a
l C
ro s s -
s e
c ti o
n a
l A
s s u
m p
ti o
n
R a
c e
, G
e n d
e r
N o
n e
F it z g
e ra
ld
K a
le s
2 0 0 5
P
s y c h
ia tr
is ts
3 2 1
C
o n
v e
n ie
n c e
H
y p
o th
e ti c a
l C
ro s s -
s e
c ti o
n a
l A
s s u
m p
ti o
n
R a
c e
a n
d
G e
n d
e r
N o
n e
F it z g
e ra
ld
K a
le s
2 0 0 5
P
ri m
a ry
C a
re
P h
y s ic
ia n
s
1 7 8
C
o n
v e
n ie
n c e
H
y p
o th
e ti c a
l C
ro s s -
s e
c ti o
n a
l A
s s u
m p
ti o
n
R a
c e
a n
d
G e
n d
e r
n o n e
F it z g
e ra
ld
L u tf e y
2 0 0 9
G
e n e
ra lis
t 1 2 8
R
a n
d o
m
H y p
o th
e ti c a
l C
ro s s -
s e
c ti o
n a
l A
s s u
m p
ti o
n
R a
c e
, G
e n d
e r,
A
g e ,
S E
S
D ia
g n
o s is
F it z g
e ra
ld
L u tf e y
2 0 1 0
P
ri m
a ry
C a
re
P h
y s ic
ia n
s
2 5 6
R
a n
d o
m
H y p
o th
e ti c a
l C
ro s s -
s e
c ti o
n a
l A
s s u
m p
ti o
n
R a
c e
, G
e n d
e r,
A
g e ,
S E
S
D ia
g n
o s is
F it z g
e ra
ld
M a s e
re jia
n
2 0 0 9
P
ri m
a ry
C a
re
P h
y s ic
ia n
s
1 2 8
R
a n
d o
m
H y p
o th
e ti c a
l C
ro s s -
s e
c ti o
n a
l A
s s u
m p
ti o
n
R a
c e
, G
e n d
e r,
A
g e ,
S E
S
D ia
g n
o s is
F it z g
e ra
ld
M a s e
re jia
n
2 0 0 9
P
ri m
a ry
C a
re
P h
y s ic
ia n
s
2 5 6
R
a n
d o
m
H y p
o th
e ti c a
l C
ro s s -
s e
c ti o
n a
l A
s s u
m p
ti o
n
R a
c e
, G
e n d
e r,
A
g e ,
S E
S
D ia
g n
o s is
F it z g
e ra
ld
M c K
in la
y
2 0 0 7
P
ri m
a ry
C a
re
P h
y s ic
ia n
s
1 2 8
R
a n
d o
m
H y p
o th
e ti c a
l C
ro s s -
s e
c ti o
n a
l A
s s u
m p
ti o
n
R a
c e
, G
e n d
e r,
A
g e ,
S E
S
T re
a tm
e n
t a
n d
D ia
g n
o s ti c
F it z g
e ra
ld
T a
m a
y o
- S
a rv
e r
2 0 0 3
E
R P
h y s ic
ia n
s
2 8 7 2
R
a n
d o
m
H y p
o th
e ti c a
l C
ro s s -
s e
c ti o
n a
l A
s s u
m p
ti o
n
R a
c e
n o n e
F it z g
e ra
ld
B a
rn a
to
2 0 1 1
H
o s p
it a
l P
h y s ic
ia n
s
3 3
C
o n
v e
n ie
n c e
a n d
P ro
b a
b ili
ty
S im
u la
ti o
n
C ro
s s -
s e
c ti o
n a
l A
s s u
m p
ti o
n
R a
c e
N
o n
e
F it z g
e ra
ld
B u
rg e
s s
2 0 0 8
P
ri m
a ry
C a
re
P h
y s ic
ia n
s
3 8 2
R
a n
d o
m
H y p
o th
e ti c a
l C
ro s s -
s e
c ti o
n a
l A
s s u
m p
ti o
n
R a
c e
n o n e
F it z g
e ra
ld
M o s k o
w it z
2 0 1 2
P
h y s ic
ia n
s
2 7
C
o n
v e
n ie
n c e
H
y p
o th
e ti c a
l C
ro s s -
s e
c ti o
n a
l S
u b
lim in
a l
P ri m
in g
R
a c e
D
ia g
n o s is
F it z g
e ra
ld
S te
p a
n ik
o v a
2 0 1 2
P
ri m
a ry
C a
re
P h
y s ic
ia n
s
8 1
R
a n
d o
m
H y p
o th
e ti c a
l C
ro s s -
s e
c ti o
n a
l S
u b
lim in
a l
P ri m
in g
R
a c e
D
ia g
n o s is
79
A p
p e n
d ix
A
O v e rv
ie w
o f
S tu
d ie
s E
x a
m in
in g
t h
e A
s s o
c ia
ti o
n o
f U
.S .
H e a lt
h c
a re
P ro
fe s s
io n
a ls
' Im
p li c
it B
ia s a
n d
P a ti
e n
t H
e a lt
h O
u tc
o m
e s
S
a m
p le
P
a ti
e n
t T
y p
e
M e th
o d
R e v ie
w
A u
th o
r Y
e a
r W
h o
S
iz e
M
e th
o d
H
y p
o th
e ti
c a l
o r
R e a
l S
tu d
y
D e s ig
n
IB
M e a s u
re
IB T
y p
e
A s s o
c ia
ti o
n
w /H
e a lt
h
O u
tc o
m e s
D e h o
n
H a id
e r
2 0 1 4
S
u rg
e o n s
2 4 8
C
o n v e n
ie n c e
H
y p o th
e ti c a
l P
ro s p e c ti v e
IA
T
R a c e a
n d
C la
s s
n o n e
D e h o
n
H a id
e r
2 0 1 5
S
u rg
e o n s
2 1 5
C
o n v e n
ie n c e
H
y p o th
e ti c a
l P
ro s p e c ti v e
IA
T
R a c e a
n d
c la
s s
n o n e
D e h o
n
O liv
e r
2 0 1 4
P
ri m
a ry
C a re
P
h y s ic
ia n s
5 4 3
C
o n v e n
ie n c e
H
y p o th
e ti c a
l C
ro s s -
s e c ti o n a
l IA
T
R a c e
n o n e
D e h o
n
P u u
m a
la
2 0 1 6
E
R P
h y s ic
ia n s
4 8
C
o n v e n
ie n c e
H
y p o th
e ti c a
l C
ro s s -
s e c ti o n a
l O
th e r
R a c e
n o n e
P
e n
n e r
2 0 1 6
O
n c o lo
g is
ts
1 3 0
C
o n v e n
ie n c e
H
y p o th
e ti c a
l C
ro s s -
s e c ti o n a
l IA
T
R a c e
C
o m
m u n
ic a ti o n
T
o ta
l 7 3 7 5
80
Appendix B: Summary of Unconscious Bias Intervention Studies Appendix B
Summary of Studies Evaluating the Effectiveness of Unconscious Bias Interventions in Pre- Health Professionals
Author Type of Bias
Intervention Method
Effectiv e
Study Population
Method s
Analysis Contr ol
O’Brien (2010)
Obesity Tutorial on uncontrollable reasons for obesity (genes/environm ent)
Yes health promotion/pu blic health bachelor’s degree seeking students
pre- post
ANOVA No
Rukavin a (2010)
Obesity stereotype: fat/lazy versus thin/motivat ed
Classroom & service-learning components, including perspective taking
No kinesiology pre- professionals
Pre- Post
ANOVA; MANOVA
Yes
Woodco ck (2013)
Race: black/white
Ex. 1: Conditioning links between self and black
No psychology students
posttest only
ANOVA Yes
Devine (2012)
Race: black/white
Multi-faceted prejudice habit- breaking intervention including perspective taking
Yes psychology students
Pre- Post
general linear models
Yes
Wallaert (2010)
Race: black/white
Ex. 1Told to avoid stereotyping on IAT
Yes psychology students
No
Wallaert (2010)
Ex. 2: Conditioning links between self and black (replication and extension)
Yes psychology students
No
Stone (2019)
Racial Bias Active Learning Workshop
Yes Medical students
Pre- Post
Bivariate Correlatio ns and Descriptiv e Statistics
No
81
Appendix C: Email Solicitation
82
Appendix D: AAFP Implicit Bias Survey
For purpose of this study we are defining Implicit (or Unconscious) Bias as the following: Implicit bias is the unconscious collection of stereotypes and attitudes that we develop toward certain groups of people, which can affect our patient relationship and care decisions.
1. How familiar are you with the term “implicit bias” as described above?
o Very familiar, have clear understanding of the term
o Somewhat familiar, only heard of the term but never had a clear understanding of
the meaning
o Not at all familiar, never heard the term before now
2. Have you ever participated in implicit bias training?
o Yes
o No
3. How important do you think it is that implicit bias training should be included in the
following?
V e
ry I m
p o
rt a
n t
Im p
o rt
a n t
N e u
tr a
l
N o t
Im p
o rt
a n
t
N o t
a t a
ll
im p
o rt
a n
t
Medical School Education
CME activities (Live or self-study)
Residency training
4. Have you ever taken an implicit bias test? The most common example is the implicit
association test (IAI).
o Yes
o No
5. Please add any other comments or observations for educating family physicians
about implicit bias and strategies to address it to support culturally appropriate, patient-
centered care, and reduce health disparities.
83
Appendix E: AAFP Unconscious Bias Interview Guide
Location: Penn Convention Center; Room 3018, Philadelphia, PA Group #1: Wednesday, Sept. 25; 1:00 to 2:30 PM – Criteria - Completed implicit bias training Group #2: Thursday, Sept. 26; 11:00 to 12:30 PM – Criteria - No implicit bias training I. Welcome/Introductions (5 minutes)
• My name is (Staff Name). The primary purpose of this discussion is to hear your thoughts on implicit bias training.
Before we get started, here are some ground rules and points of information: o Discuss housekeeping rules
REINFORCE OJBECTIVITY: Not looking for consensus, negative comments won’t hurt, honest opinions are most helpful.
CONFIDENTIALITY. Everything you say here will be kept strictly confidential. Nothing said in this group will ever be addressed with any individual by name. We would ask also that you similarly maintain the confidentiality of what is said in the group.
AUDIOTAPING: This session is being audio taped so we can write an accurate report and we don’t have to be taking notes throughout the discussion.
• Tell us about yourself, where you live, and where you work? What is your practice setting?
II. General Perceptions of Implicit Bias
• When you hear the term “implicit or unconscious bias” what comes to mind first? How would you explain this term to someone that wasn’t familiar with it?
• [SHOW DEFINITION ON THE SCREEN] Implicit bias is the unconscious collection of stereotypes and attitudes that we develop toward certain group of people, which can affect our patient relationships and care decisions.
• Based on this definition described on the screen, what are some examples of how implicit bias might be played out in a medical practice/setting? Describe scenarios in detail. PROBE:
84
o Differential treatment of patients by race, gender, weight, age, language, religion, ethnic background, income, specific diseases, or insurance status, etc.
• How does implicit bias contribute to health disparities? How might it influence clinical care? What is the clinical outcome of implicit bias? PROBE:
o Patient behavior and decisions – higher treatment dropout, lower participation in screening, delays in seeking help, lack of filling prescriptions, etc.
• How serious of issue do you believe implicit bias is impacting the health of patients? If yes, how?
• In order to address the impact of implicit bias on clinical care decisions, what has your organization or practice implemented? Do you work as a group/team to uncover implicit biases and develop strategies to address them?
• What actions can you as a family physician do to combat implicit bias? III. General Perceptions of Implicit Bias Training IF PARTICIPATED IN TRAINING ASK:
• [SHOW OF HANDS] How many have participated in implicit bias training? How many years ago? What were the drivers of participating in the training?
• What did you value most from the training? o Increase awareness o Mindfulness o Change behavior
• Did you put in a practice any of the strategies you learned during the training? If yes, which ones? If no, why not?
• How confident are you in your ability to recall knowledge learned from the training? If yes, please explain? If no, why not?
• [SHOW POTENTIAL OUTCOMES ON THE SCREEN] Have you experienced any changes in any of the following as a result of the training? If yes, please explain how? Give examples.
o Job satisfaction o Cultural competence o Patient-centered care/relationships o Communication skills (body language and verbal cues o Perspective-taking o Learned to slow down
85
o Learn/educate your blind spots
• What are the shortcomings of implicit bias training? Barriers?
• How important do think is it that implicit bias training be included in the medical school training? Residency training? CME activities?
IF NO TRAINING ASK:
• [SHOW OF HANDS] How many have participated in implicit bias training? Have you had any opportunities to participate in training? If yes, when, and where? Why didn’t you participate?
• If you had the option to participate in implicit bias training, what would be the key drivers for participation? What would be the expected benefits? Value proposition?
• What type of training would you find most valuable?
• How important do think is it that implicit bias training be included in the medical school training? Residency training? CME activities?
IV. AAFP’s Resources on Implicit Bias
• How should the AAFP communicate the value of implicit training to family physicians? Where should we focus our messaging? What would be compelling? Probe:
o Facts/numbers/statistics o Illustrative examples o Narratives o Testimonials o Theory to action
• How should the AAFP motivate family physicians and their teams to adopt
strategies for controlling implicit bias?
• SHOW CONCEPT OF AAFP’s IMPLICIT BIAS WEBSITE? CONCEPT?
QUALITY IMPROVEMENT? AS PART OF A WIDER PROGRAM? CME
ACTIVITY
• Thanks for your time.
86
Appendix F: Consent for Participation in a Research Study Title Protocol # 19-358 The American Academy of Family Physicians (AAFP) Center for Diversity and Health Equity: Identifying Family Physicians’ Values, Knowledge and Barriers Regarding Implicit Bias Training
Investigators Danielle D. Jones Invitation to Participate As a member of the AAFP you are being asked to participate in a focus group discussing perceptions and knowledge of implicit (unconscious) bias and its effects on patient health outcomes. The main purpose of this study is to create new knowledge for the benefit of informing the development and implementation of future implicit bias training and education curricula. Research studies may or may not benefit the people who participate. Research is voluntary, and you may change your mind at any time. There will be no penalty to you if you decide not to participate, or if you start the study and decide to stop early. This consent form explains what you should do if you are in the study. It also describes the possible risks and benefits. Please read the form carefully and ask as many questions as you need to, before deciding about this research. You can ask questions now or anytime during the study. The researchers will tell you if they receive any new information that might cause you to change your mind about participating. This research study will take place as part of the American Academy of Family Physicians Family Medicine Experience (FMX). Who will Participate 24 members of the AAFP have been invited to participate in one of two focus groups during FMX. Eligible participants were identified based on a short survey disseminated to the Member Insight Exchange asking about participation in implicit bias training. Purpose The purpose of this study is to understand physicians’ knowledge, perceptions, behaviors, and skills associated with implicit bias education and training. Participants will be asked about their behaviors following an implicit bias training as well as barriers to participating in training. Procedures Participation in this study consist of a group interview that will last approximately 1.5 hours. The study takes place at the Penn Convention Center; Room 3018 on Wed. Sept. 25; 1:00 to 2:30 PM and again on Thursday, Sept. 26; 11:00 to 12:30 PM in Philadelphia, PA. Both sessions will be audio-recorded so that the researchers have an accurate record of the interview and notes will stored securely on a password protected network in accordance with the AAFP’s record management requirements.
87
Voluntary Participation Research studies may or may not benefit the people who participate. Research is voluntary, and you may change your mind at any time. There will be no penalty to you if you decide not to participate, or if you start the study and decide to stop early. This consent form explains what you should do if you are in the study. It also describes the possible risks and benefits. Please read the form carefully and ask as many questions as you need to, before deciding about this research. You can ask questions now or anytime during the study. The researchers will tell you if they receive any new information that might cause you to change your mind about participating. Fees and Expenses There is no monetary cost to the participants. Payments for Participation Members will receive a $150 gift card for their participation. Risks and Inconveniences The interview questions may be personal. Some of the questions might be embarrassing or uncomfortable. You are free not to answer any questions. The risk for someone outside of the research study to learn of your participation or responses is low. Your name will not be used in any publication or presentation about this research. There may be other risks of the study that are not yet known. Benefits Researchers hope that the information collected from this study may be useful in understanding physician educational and training needs and improve the quality of care delivered to patients. Alternatives to Study Participation Participation in this study is voluntary. Deciding not to participate will have no effect on your membership in the AAFP. Confidentiality Interviews be audio-recorded and transcribed so that the researchers have an accurate record. All audio and notes will be stored securely on a password protected network in accordance with the AAFP’s record management requirements. While every effort will be made to keep confidential all of the information you complete and share, it cannot be absolutely guaranteed. Individuals from the American Academy of Family Physician’s
88
Institutional Review Board (a committee that reviews and approves research studies) and Federal regulatory agencies may look at records related to this study for quality improvement and regulatory functions. Future Use Subject's information will not be used or distributed for future research. In Case of Injury Although it is not the AAFP’s policy to compensate or provide medical treatment for persons who participate in studies, if you think you have been injured as a result of participating in this study, please contact Jennifer Farris, AAFP IRB Assistant, at 913- 906-6134 or [email protected]” Questions If you have any questions about the study that you are participating in you are encouraged to contact Danielle Jones, the investigator, at 913-906-6319 or [email protected]. If you have any questions about your rights as a research subject, you are encouraged to contact Jennifer Farris, AAFP IRB Assistant, at 913-906-6134 or [email protected] . Signing here means that you have read the information provided in this Informed Consent Form and have had your questions answered to your satisfaction, and voluntarily agree to participate in this study. This consent or a copy of this consent will be kept ______________________________________________________________________ ______ Printed Name (Participant) Signature Date ______________________________________________________________________ ______ Printed Name (Investigator) Signature Date
