CSHS Nurse Leader Meeting 2/25/21 At the Intersection of Health and Education: Recognizing Bias for Better Outcomes Prior to attending the February 25th CSHS Nurse Leader Meeting, please do the following prework which will take about 30 minutes to complete. These materials will be discussed during the program: Read:
• Sadker, D. (n.d.) Seven Forms of Bias in Instructional Materials: Some Practical Ideas for Confronting Curricular Bias. The Myra Sadker Foundation.
• FitzGerald, C. & Hurst, S. (2017). Implicit bias in healthcare professionals: a systematic review. BMC Medical Ethics, (18) 19. doi: 10.1186/s12910-017-0179-8
Watch: Cracking the Codes: Joy DeGruy “A Trip to the Grocery Store” https://www.youtube.com/watch?app=desktop&v=Wf9QBnPK6Yg
1/21/2021 Seven Forms of Bias in Instructional Materials
https://www.sadker.org/curricularbias.html?back=https%3A%2F%2Fwww.google.com%2Fsearch%3Fclient%3Dsafari%26as_qdr%3Dall%26as_occt%… 1/2
Some Practical Ideas for Confronting Curricular Bias
Back in the 1970s and the 1980s, publishers and professional associations issued guidelinesfor non-racist and non-sexist books. As a result, texts of the last twenty years are muchimproved. Unfortunately, they are far from bias-free. The following seven forms of bias can befound not only in K-12 textbooks, but also in college texts, in the media – in fact, they are allaround us. Feel free to explore these definitions with your students, as well as try thestrategies suggested.
Seven Forms of Bias in Instructional Materials
Invisibility: What You Don’t See Makes a Lasting Impression.
The most fundamental and oldest form of bias in instructional materials is the complete orrelative exclusion of a group. Textbooks published prior to the 1960s largely omitted AfricanAmericans, Latinos, and Asian Americans from both the narrative and illustrations. Many oftoday’s textbooks are improved, but far from perfect. Women, those with disabilities, gays andhomosexuals continue to be missing from many of today’s texts.
Stereotyping: Shortcuts to Bigotry.
Perhaps the most familiar form of bias is the stereotype, which assigns a rigid set ofcharacteristics to all members of a group, at the cost of individual attributes and differences.While stereotypes can be positive, they are more often negative. Some typical stereotypesinclude:
Men portrayed as assertive and successful in their jobs, but rarely discussed ashusbands or fathers.Women as caregiversJews as rich
Imbalance and Selectivity: A Tale Half Told.
Curriculum may perpetuate bias by presenting only one interpretation of an issue, situation, orgroup of people. Such accounts simplify and distort complex issues by omitting differentperspectives.
A text reports that women were "given" the vote, but does not discuss the work,sacrifices, and even physical abuse suffered by the leaders of the suffrage movementthat "won" the vote.Literature is drawn primarily from western, male authors.Math and science courses typically reference European discoveries and formulas.
Unreality: Rose Colored Glasses.
Many researchers have noted the tendency of instructional materials to gloss over unpleasantfacts and events in our history. By ignoring prejudice, racism, discrimination, exploitation,oppression, sexism, and inter-group conflict, we deny students the information they need torecognize, understand, and perhaps some day conquer societal problems. Examples include:
Because of affirmative action programs, people of color and women now enjoyeconomic and political equality with (or superiority over) white males.The notion that technology will resolve persistent social problems.
Fragmentation and Isolation: The Parts Are Less than the Whole.
Did you ever notice a "special" chapter or insert appearing in a text? For example, a chapteron "Bootleggers, Suffragettes, and Other Diversions" or a box describing "Ten Black Achievers
David Sadker
The Myra Sadker Foundation
1/21/2021 Seven Forms of Bias in Instructional Materials
https://www.sadker.org/curricularbias.html?back=https%3A%2F%2Fwww.google.com%2Fsearch%3Fclient%3Dsafari%26as_qdr%3Dall%26as_occt%… 2/2
in Science." Fragmentation emerges when a group is physically or visually isolated in the text.Often, racial and ethnic group members are depicted as interacting only with persons likethemselves, isolated from other cultural communities. While this form of bias may be lessdamaging than omission or stereotypes, fragmentation and isolation present non-dominantgroups as peripheral members of society.
Linguistic Bias: Words Count.
Language can be a powerful conveyor of bias, in both blatant and subtle forms. Linguistic biascan impact race/ethnicity, gender, accents, age, (dis)ability and sexual orientation.
Native Americans described as "roaming," "wandering," or "roving" across the land.Such language implicitly justifies the seizure of Native lands by "more goal-directed"white Americans who "traveled" or "settled" their way westward.Such words as forefathers, mankind, and businessman serve to deny the contributions(even the existence) of females.The bias against non-English speakers.
Cosmetic Bias: "Shiny" covers.
The relatively new cosmetic bias suggests that a text is bias free, but beyond the attractivecovers, photos, or posters, bias persists. This "illusion of equity" is really a marketing strategyto give a favorable impression to potential purchasers who only flip the pages of books.
A science textbook that features a glossy pullout of female scientists but includesprecious little narrative of the scientific contributions of women.A music book with an eye-catching, multiethnic cover that projects a world of diversesongs and symphonies belies the traditional white male composers lurking behind thecover.
Investigative Strategies for Bias Detectives
Here are several strategies for teaching these concepts in K-12 and teacher educationclassrooms. Ask students to review school textbooks and identify each of these seven forms.Then ask them to suggest ways to remove the bias and create more equitable textbooks.
Extend this activity by asking students to identify these forms of bias in college leveltexts (academic areas as well as teacher education), or in magazines and televisionprogramming.While curriculum bias clearly impacts females and students of color, males can also bevictims as well. Using the 7 forms of bias as a framework, find examples that negativelyimpacts males, and suggest ways to overcome the bias.Ask students to identify how these seven forms emerge in interpersonal interactions.For example, teachers stereotype when males are asked to help with physicalclassroom tasks, or fragment by studying African Americans only during "Black HistoryMonth.
Share an example of curriculum bias or equity that you have identified.
FitzGerald and Hurst BMC Medical Ethics (2017) 18:19 DOI 10.1186/s12910-017-0179-8
RESEARCH ARTICLE Open Access
Implicit bias in healthcare professionals:a systematic review
Chloë FitzGerald* and Samia HurstAbstract
Background: Implicit biases involve associations outside conscious awareness that lead to a negative evaluation ofa person on the basis of irrelevant characteristics such as race or gender. This review examines the evidence thathealthcare professionals display implicit biases towards patients.
Methods: PubMed, PsychINFO, PsychARTICLE and CINAHL were searched for peer-reviewed articles publishedbetween 1st March 2003 and 31st March 2013. Two reviewers assessed the eligibility of the identified papers based onprecise content and quality criteria. The references of eligible papers were examined to identify further eligible studies.
Results: Forty two articles were identified as eligible. Seventeen used an implicit measure (Implicit Association Testin fifteen and subliminal priming in two), to test the biases of healthcare professionals. Twenty five articlesemployed a between-subjects design, using vignettes to examine the influence of patient characteristics onhealthcare professionals’ attitudes, diagnoses, and treatment decisions. The second method was includedalthough it does not isolate implicit attitudes because it is recognised by psychologists who specialise in implicitcognition as a way of detecting the possible presence of implicit bias. Twenty seven studies examined racial/ethnic biases; ten other biases were investigated, including gender, age and weight. Thirty five articles foundevidence of implicit bias in healthcare professionals; all the studies that investigated correlations found a significantpositive relationship between level of implicit bias and lower quality of care.
Discussion: The evidence indicates that healthcare professionals exhibit the same levels of implicit bias as the widerpopulation. The interactions between multiple patient characteristics and between healthcare professional and patientcharacteristics reveal the complexity of the phenomenon of implicit bias and its influence on clinician-patientinteraction. The most convincing studies from our review are those that combine the IAT and a method measuring thequality of treatment in the actual world. Correlational evidence indicates that biases are likely to influence diagnosisand treatment decisions and levels of care in some circumstances and need to be further investigated. Our review alsoindicates that there may sometimes be a gap between the norm of impartiality and the extent to which it is embracedby healthcare professionals for some of the tested characteristics.
Conclusions: Our findings highlight the need for the healthcare profession to address the role of implicit biases indisparities in healthcare. More research in actual care settings and a greater homogeneity in methods employed to testimplicit biases in healthcare is needed.
Keywords: Implicit bias, Prejudice, Stereotyping, Attitudes of health personnel, Healthcare disparities
* Correspondence: [email protected] for Ethics, History, and the Humanities, Faculty of MedicineUniversity of Geneva, Genève, Switzerland
© The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
FitzGerald and Hurst BMC Medical Ethics (2017) 18:19 Page 2 of 18
BackgroundA patient should not expect to receive a lower standardof care because of her race, age or any other irrelevantcharacteristic. However, implicit associations (uncon-scious, uncontrollable, or arational processes) may influ-ence our judgements resulting in bias. Implicit biasesoccur between a group or category attribute, such as beingblack, and a negative evaluation (implicit prejudice) oranother category attribute, such as being violent (implicitstereotype) [1].1 In addition to affecting judgements, im-plicit biases manifest in our non-verbal behaviour to-wards others, such as frequency of eye contact andphysical proximity. Implicit biases explain a potentialdissociation between what a person explicitly believesand wants to do (e.g. treat everyone equally) and thehidden influence of negative implicit associations onher thoughts and action (e.g. perceiving a black patientas less competent and thus deciding not to prescribethe patient a medication).The term ‘bias’ is typically used to refer to both implicit
stereotypes and prejudices and raises serious concerns inhealthcare. Psychologists often define bias broadly; such as‘the negative evaluation of one group and its membersrelative to another’ [2]. Another way to define bias is tostipulate that an implicit association represents a biasonly when likely to have a negative impact on analready disadvantaged group; e.g. if someone associatesyoung girls with dolls, this would count as a bias. It isnot itself a negative evaluation, but it supports an imageof femininity that may prevent girls from excelling in areastraditionally considered ‘masculine’ such as mathematics[3]. Another option is to stipulate that biases are notinherently bad, but only to be avoided when they inclineus away from the truth [4].In healthcare, we need to think carefully about exactly
what is meant by bias. To fulfil the goal of deliveringimpartial care, healthcare professionals should be waryof any kind of negative evaluation they make that islinked to membership of a group or to a particular char-acteristic. The psychologists’ definition of bias thus maybe adequate for the case of implicit prejudice; there areunlikely, in the context of healthcare, to be any justifiedreasons for negative evaluations related to group mem-bership. The case of implicit stereotypes differs slightlybecause stereotypes can be damaging even when theyare not negative per se. At least at a theoretical level,there is a difference between an implicit stereotype thatleads to a distorted judgement and a legitimate associ-ation that correctly tracks real world statistical informa-tion. Here, the other definitions of bias presented abovemay prove more useful.The majority of people tested from all over the world
and within a wide range of demographics show re-sponses to the most widely used test of implicit
attitudes, the Implicit Association Test (IAT), that indi-cate a level of implicit anti-black bias [5]. Other biasestested include gender, ethnicity, nationality and sexualorientation; there is evidence that these implicit attitudesare widespread among the population worldwide andinfluence behaviour outside the laboratory [6, 7]. Forinstance, one widely cited study found that simply chan-ging names from white-sounding ones to black-soundingones on CVs in the US had a negative effect on callbacks[8]. Implicit bias was suspected to be the culprit, and areplication of the study in Sweden, using Arab-soundingnames instead of Swedish-sounding names, did in factfind a correlation between the HR professionals whopreferred the CVs with Swedish-sounding names and ahigher level of implicit bias towards Arabs [9].We may consciously reject negative images and ideas
associated with disadvantaged groups (and may belongto these groups ourselves), but we have all beenimmersed in cultures where these groups are constantlydepicted in stereotyped and pejorative ways. Hence thedescription of ‘aversive racists’: those who explicitlyreject racist ideas, but who are found to have implicitrace bias when they take a race IAT [10]. Although thereis currently a lack of understanding of the exact mech-anism by which cultural immersion translates into impli-cit stereotypes and prejudices, the widespread presenceof these biases in egalitarian-minded individuals suggeststhat culture has more influence than many previouslythought.The implicit biases of concern to health care profes-
sionals are those that operate to the disadvantage ofthose who are already vulnerable. Examples includeminority ethnic populations, immigrants, the poor, lowhealth-literacy individuals, sexual minorities, children,women, the elderly, the mentally ill, the overweight andthe disabled, but anyone may be rendered vulnerablegiven a certain context [11]. The vulnerable in health-care are typically members of groups who are alreadydisadvantaged on many levels. Work in political philoso-phy, such as the De-Shalit and Wolff concept of ‘corro-sive disadvantage’, a disadvantage that is likely to lead tofurther disadvantages, is relevant here [12]. For instance,if a person is poor and constantly worried about makingends meet, this is a disadvantage in itself, but can be cor-rosive when it leads to further disadvantages. In a coun-try such as Switzerland, where private health insuranceis mandatory and yearly premiums can be lowered by in-creasing the deductible, a high deductible may lead sucha person to refrain from visiting a physician because ofthe potential cost incurred. This, in turn, could meanthat the diagnosis of a serious illness is delayed leadingto poorer health. In this case, being poor is a corrosivedisadvantage because it leads to a further disadvantageof poor health.
FitzGerald and Hurst BMC Medical Ethics (2017) 18:19 Page 3 of 18
The presence of implicit biases among healthcare pro-fessionals and the effect on quality of clinical care is acause for concern [13–15]. In the US, racial healthcaredisparities are widely documented and implicit race biasis one possible cause. Two excellent literature reviewson the issue of implicit bias in healthcare have recentlybeen published [16, 17]. One is a narrative review thatselects the most significant recent studies to provide ahelpful overall picture of the current state of the re-search in healthcare on implicit bias [16]. The other is asystematic review that focusses solely on racial bias andthus captures only studies conducted in the US, whererace is the most prominent issue [17]. Our review differsfrom the first because it poses a specific question, is sys-tematic in its collection of studies, and includes anexamination of studies solely employing the vignettemethod. Its systematic method lends weight to theevidence it provides and its inclusion of the vignettemethod enables it to compare two different literatureson bias in healthcare. It differs from the second becauseit includes all types of bias, not only racial; partly as aconsequence, it captures many studies conducted out-side the US. It is important to include studies conductedin non-US countries because race understood as white/black is not the source of the most potentially harmfulstereotypes and disparities in all cultural contexts. Forexample, a recent vignette study in Switzerland foundthat in the German-speaking part of the country, physi-cians displayed negative bias in treatment decisionstowards fictional Serbian patients (skin colour was un-specified, but it would typically be assumed to be white),but no significant negative bias towards fictional patientsfrom Ghana (skin colour would be assumed to be black)[18]. In the Swiss German context, the issue of skincolour may thus be less significant for potential biasthan that of country of origin.2
MethodsData sources and search strategyOur research question was: do trained healthcare profes-sionals display implicit biases towards certain types ofpatient? PubMed (Medline), PsychINFO, PsychARTICLEand CINAHL were searched for peer-reviewed articlespublished between 1st March 2003 and 31st March 2013.When we performed exploratory searches on PubMed be-fore conducting the final search, we noticed that in 2003there was a sharp increase in the number of articles onimplicit bias and so we decided to begin from this year.The final searches were conducted on the 31st March2013. We used a combination of subject headings and freetext terms that related to the attitudes of healthcareprofessionals (e.g. “physician-patient relations”, “atti-tude of health personnel”), implicit biases (e.g. “prejudice”,“stereotyping”, “unconscious bias”), particular kinds of
discrimination (e.g. “aversive racism”, anti-fat bias”,“women’s health”), and healthcare disparities (e.g. “healthstatus disparities”, “delivery of health care”) which werecombined with the Boolean operators “AND” and “OR”.
Study selection3767 titles were retrieved and independently screened bythe two reviewers (SH and CF). The titles that wereagreed by both after discussion to be ineligible accordingto our inclusion criteria were discarded (3498) and theabstracts of the remaining articles (269) were independ-ently screened by both reviewers. Abstracts that wereagreed by both reviewers to be ineligible according toour inclusion criteria were discarded (241). When the in-eligible abstracts were discarded, the remaining 28 arti-cles were read and independently rated by us both. Outof these, 27 articles were agreed after discussion to meritinclusion in the final selection. One article was excludedat this stage because it did not fit our inclusion criteria(it did not employ the assumption method or an implicitmeasure). Additionally, the reference lists of these 27articles were manually scanned by CF, and the full textarticles resulting from this were independently read byboth reviewers, resulting in the inclusion of a further 11articles that both reviewers agreed fitted the inclusioncriteria. After a repeat process of scanning the referencelists of the 11 articles from the second round, the finalnumber of eligible articles was 42. All disagreementswere resolved through discussion.The inclusion criteria were:
1. Empirical study.2. A method identifying implicit rather than explicit
biases.3. Participants were physicians or nurses who had
completed their studies.4. Written in English or another language spoken by
CF or SH (CF: French, Italian, Spanish, Catalan; SH:French, Italian, German).
There is no clear consensus on the meaning of theterm ‘implicit’. The term is used in psychology to referto a feature or features of a mental process. We chose awide negative definition of implicit processes, assumingthat implicit social cognition is involved in the absenceof any of the four features that characterise explicit cog-nition: intention, conscious availability, controllability,and the need for mental resources. This absence doesnot rule out the involvement of explicit processes, butindicates the presence of implicit processes. Whilemost institutional policies against bias focus on explicitcognition, research on implicit bias shows that this ismistaken [6].
FitzGerald and Hurst BMC Medical Ethics (2017) 18:19 Page 4 of 18
There is broad agreement in psychology that methodsknown as ‘implicit measures’, including the affectivepriming task, the IAT and the affective Simon task,reveal implicit attitudes [19]. We included articles usingthese measures. We also included studies that employeda method popular in bioethics literature that we label‘the assumption method’. It involves measuring differ-ences across participants in response to clinical vignettes,identical except for one feature, such as the race, of thecharacter in the vignette. There is no direct measure ofthe implicitness or non-explicitness of the processes atwork in participants; instead, there is an assumption thatthe majority are explicitly motivated to disregard factorssuch as race. If there is a statistically significant differencein the diagnosis or treatment prescribed correlated with –for example- the race of the patient, the researchers inferthat it is partly a result of implicit processes in the phy-sicians’ decision-making. The assumption method ofmeasuring implicit bias has been used in a variety ofnaturalistic contexts where it is harder to bring subjectsinto the laboratory. It is recognised by psychologists whospecialise in implicit cognition as a way of detecting thepossible presence of implicit bias, if not as an implicitmeasure in itself [6].Studies that used self-report questionnaires were not
included because, although they can use subtle methodsto estimate a subject’s attitudes, they are typically usedin psychology as a measure of explicit mental processes.There are potential problems with the implicit/explicitdistinction as applied to psychological measures and itmay be preferable in future research to speak of ‘direct’and ‘indirect’ measures, but for the purposes of thereview we followed this convention in psychology. Theoriginal idea behind implicit measures was that theyattempted to measure something other than explicitmental processes, whereas self-report questionnaires aska subject direct questions and thus prompt a chain ofexplicit conscious reasoning in the subject.
Data extractionData were extracted by CF and reviewed by SH for ac-curacy and completeness. All disagreements with theinformation extracted were resolved through discussion.We contacted the corresponding author of an article toobtain information that was not available in the pub-lished manuscript that related to the nature of the pres-entation given to recruit participants, but received noresponse.
ResultsIdentified studiesThe eligible studies are described in Table 1 and theirmain characteristics are outlined in Table 2. The most fre-quently examined biases were racial/ethnic and gender,
but ten other biases were investigated (Table 2). Four ofthe assumption studies compared results from two ormore countries to explore effects of differences in health-care systems.The 14 assumption method studies examining mul-
tiple biases investigated interactions between biases.They recorded the socio-demographic characteristics ofthe participants to reveal complex interactions betweenphysician characteristics and the characteristics of theimaginary ‘patient’ in the vignette.All IAT studies measured implicit prejudice; five also
measured implicit stereotypes. When implicit prejudiceis measured, words or images from one category arematched with positive or negative words (e.g., blackfaces with ‘pleasant’). When implicit stereotypes aremeasured, words or images from one category arematched with words from a conceptual category (e.g. fe-male faces and ‘home’).Nine IAT studies combined the IAT with a measure
of physician behaviour or treatment decision to see ifthere were correlations between these and levels of im-plicit bias.The subliminal priming studies were dissimilar: one
was an exploratory study to see if certain diseases werestereotypically associated with African Americans, usingfaces as primes and reaction times to the names of dis-eases as the measure of implicit association; the otherstudy used race words as primes and tested the effect oftime pressure on responses to a clinical vignette.A variety of media were used for the clinical vignette
and the method of questioning participants within theassumption method. One unusual study used simula-tions of actual encounters with patients, hiring actorsand using a set for the physicians to role-play. Physi-cians’ treatment decisions were recorded by observers,and the physician recorded his own diagnosis, prognosisand perceptions after the encounter.
LimitationsOf specific studiesLimitations are detailed in Table 3. Some studies failed toreport response rates, or to provide full information onstatistical methods or participant characteristics. Somehad very small sample sizes and the majority did not men-tion calculating the power of their sample. Some authorsexplicitly informed participants of the purpose of thestudy, or gave participants questionnaires or other teststhat indicated the subject of the study before presentingthem with the vignette. For optimal results, participantsshould not be alerted to the particular patient charac-teristic(s) under study, particularly in an assumptionstudy where knowing the characteristic(s) may influ-ence the interpretation of the vignette. In IAT studies,
Table
1Stud
iesinclud
edin
thesystem
aticreview
Year
Firstauthor
Cou
ntry
Metho
dPo
pulatio
nRecruitm
ent
Respon
serate
Mainfinding
s(re
levant
tosystem
aticreview
)
Age 20
10Protière
[25]
France
Assum
ptionMetho
d388on
cologistsand
radiothe
rapists
Mail-o
utcoun
tryw
ide.
69%
Sign
ificant
negativedifferences
intreatm
ent
choice
forolde
rpatients.
AIDSpatients
2007
Li[61]
China
Assum
ptionMetho
d1101
(includ
ingjustover
50%
doctors,40%
nurses)
Rand
omselection
ofinstitu
tions
and
individu
als.
Less
than
8%refusal
rate
(includ
es10%
labtechnicians)
Attitu
desas
measuredacross
subjectswere
morene
gativetowards
AIDSpatientsthan
towards
hepatitisBpatients.
Braininjuredpatientswho
have
contrib
uted
totheirinjury
2011
Lind
en[58]
UK
Assum
ptionMetho
d69
nurses
Email.
24%
Morene
gativeattitud
esas
measuredacross
subjectswerefoun
dagainstindividu
alsseen
ascontrib
utingtowards
theirinjury.
2010
Redp
ath[59]
UK
Assum
ptionMetho
d155(94qu
alified
nurses,61
qualified
doctors)
Not
specified
.Not
specified
.Morene
gativeattitud
esas
measuredacross
subjectswerefoun
dagainstindividu
alsseen
ascontrib
utingtowards
theirinjury.A
ttitu
des
sign
ificantlyrelatedto
intend
edhe
lping
behaviou
r.
Disability
2012
Aaberg[62]
US
IAT(prejudice)
132nu
rseed
ucators
Email.
21%
Neg
ativeim
plicitbias
againstthedisabled
,strong
erthan
that
oftheaveragepo
pulatio
n.
Gen
der
2005
Abu
ful[63]
Israel
Assum
ptionMetho
d172ph
ysicians
(internists,
cardiologists,family
physicians,g
eneral
practitione
rs)
Con
tinuing
med
ical
education.
Not
specified
.Neg
ativebias
againstwom
enin
diagno
sisof
riskandprescriptio
nof
lipid-lo
wering
med
ications
andaspirin
.
Intraven
ousDrugUsers(ID
Us)
2007
Bren
er[49]
Australia
IAT(prejudice)
60he
alth
care
workers
(HCW)from
drug
and
alcoho
lfacilitiesandliver
clinics:21
physicians,37
nurses,twomed
ical
stud
ents
Differen
tfacilitiesand
GPs
iden
tifiedthroug
hne
tworking
.
Not
specified
.HCW
hadpo
sitiveexplicitattitud
esand
negativeim
plicitattitud
estowards
HCV
positiveIDUs.Con
tact
(asestim
ated
byHCW)
pred
ictedexplicit(positive)andim
plicit
attitud
es(neg
ative)
towards
IDUbe
yond
the
effect
ofconservatism.
2008
vonHippe
l[48]
Australia
IAT(prejudice)
44DrugandAlcoh
ol(D&A
)nu
rses
Selectionof
Drug&
Alcoh
oltreatm
ent
facilities,ne
edleand
syrin
geexchange
prog
rams,andprim
ary-
care
facilities.
Not
specified
.Challeng
ingbe
haviou
rsby
IDUspred
icted
self-repo
rted
stress
ofnu
rses,w
hich,inturn,
pred
ictedintentionto
change
jobs.The
relatio
nbe
tweenstress
andintentionto
change
jobs
sign
ificantlymed
iatedby
the
nurses’implicitprejud
ice,no
texplicitprejud
ice.
Men
tally
ill
2007
Cho
w[64]
Hon
gKo
ngAssum
ptionMetho
d433(107
physicians,322
nurses
andfour
who
didn
'tstate)
Rand
omdistrib
utionvia
wardmanagers(nurses)
andem
ail(ph
ysicians).
36.1%
Morene
gativeattitud
esas
measuredacross
subjectsfoun
dtowards
psychiatric
patients
than
tono
n-psychiatric
patients.
FitzGerald and Hurst BMC Medical Ethics (2017) 18:19 Page 5 of 18
Table
1Stud
iesinclud
edin
thesystem
aticreview
(Con
tinued)
2005
Mackay[26]
UK
Assum
ptionMetho
d89
A&E
med
ical
andnu
rsingstaff
4A&E
Dep
artm
entsin
Greater
Manchester.
49%
Thegreatertheattributions
ofcontrollability
toself-harm
ingpatients,thegreaterthe
negativeaffect
ofstafftowards
thepatient
asmeasuredacross
subjects,and
theless
the
prop
ensity
tohe
lp.
2012
Neaup
ort[65]
France
Assum
ptionMetho
d322med
icalreside
ntsof
all
specialties
inon
eho
spital
Email.
47.4%
Thoseassign
edthevign
ette
that
includ
edthe
psychiatric
illne
sslabe
lsaidthat
they
were
less
likelyto
wantto
treattheindividu
aland
beinvolved
with
her/him
invarious
ways.
2008
Peris
[42]
Worldwide
IAT(prejudice
and
stereo
type
)682men
talh
ealth
profession
als(clinical
psycho
logists,social
workers,cou
nsellors,
psychiatristsandothe
rs)
andclinicalgraduate
stud
ents.
ProjectIm
plicitweb
site
and110clinicians
and
graduate
stud
ents
recruiteddirectly
throug
hlistserves.
81%
(after
rand
omassign
men
tto
stud
yviaProjectIm
plicit
andinclud
ing747
non-men
talh
ealth
profession
als)
Overall,explicitandim
plicitview
swereno
tne
gativetowards
individu
alswith
men
tal
illne
ss.Tho
sewith
men
talh
ealth
training
displayedless
implicitandexplicitprejud
ices.
Theire
xplicit(but
notimplicit)biases
predicted
morenegativepatient
prog
noses,bu
timplicit
(and
notexplicit)biases
predictedover-diagn
osis.
Multip
lebiases
2004
Arber
a[27]
US/UK
Assum
ptionMetho
dBiases:A
ge,G
ende
r,Racial/ethnicandSocio-
Econ
omicStatus
(SES)
256prim
arycare
physicians
intheUSandtheUK
Screen
ingteleph
one
calls.
65%
intheUSand
60%
intheUK.
Gen
derandageinfluen
cedthedo
ctors’
questio
ning
ofpatientspresen
tingwith
coronary
heartdisease(CHD)in
both
coun
tries.Men
wereaskedmorequ
estio
nsoveralland
middle-aged
patientswereasked
morelifestylequ
estio
ns.
2006
Arber
a[32]
UKandUS
Assum
ptionMetho
dBiases:A
ge,G
ende
r,Racial/ethnicandSES.
256prim
arycare
physicians
intheUSandtheUK
Screen
ingteleph
one
calls.
65%
intheUSand
60%
intheUK.
Thege
nder
ofthepatient
sign
ificantly
influen
ceddo
ctors’diagno
sticand
managem
entactivities.M
idlifewom
enwere
askedfewestqu
estio
nsandprescribed
least
med
icationapprop
riate
forCHD.
2006
Barnhart[66]
US
Assum
ptionMetho
dBiases:Racial/e
thnic,
Gen
der,andSocial
Circum
stances.
544family
med
icine
physicians,internists,
cardiologists,and
cardiothoracicsurgeo
ns
Mail-o
ut.
70%
Thepatient’srace
andge
nder
didno
tsign
ificantlyaffect
theph
ysicians’treatmen
tpreferen
ces.How
ever,significantdifferences
werefoun
daccordingto
socialcircum
stance.
2008
Böntea
[31]
US,UKand
Germany
Assum
ptionMetho
dBiases:A
ge,G
ende
r,Racial/ethnicandSES.
384ph
ysicians
(internistsor
family
practitione
rsin
the
USandGermanyor
gene
ral
practitione
rsin
theUK)
Screen
ingteleph
one
calls.
64.9%
intheUS,
59.6%
intheUK,
and65%
inGermany.
Results
show
edge
nder
differences
inthe
diagno
sticstrategies
ofthedo
ctors.
2010
Deh
lend
orf[39]
US
Assum
ptionMetho
dBiases:Racial/e
thnic
andSES.
524he
alth
care
providers
(96%
MD/DO,4%
Nurse
Practitione
r/Ph
ysician
Assistant)
Con
venien
cesample
from
meetin
gsof
profession
alsocieties.
Not
specified
.Low
SESwhiteswereless
likelyto
have
intrauterin
econtraceptionrecommen
ded
than
high
SESwhites.By
race/ethnicity,low
SESLatin
asandblacks
weremorelikelyto
have
intrauterin
econtraception
recommen
dedthan
low
SESwhites,with
noeffect
ofrace/ethnicity
forhigh
SESpatients.
Low
SESpatientswerejudg
edto
besign
ificantlymorelikelythan
high
SESpatients
FitzGerald and Hurst BMC Medical Ethics (2017) 18:19 Page 6 of 18
Table
1Stud
iesinclud
edin
thesystem
aticreview
(Con
tinued)
tohave
anSTIand
anun
intend
edpreg
nancy,
andwerealso
judg
edto
beless
know
ledg
eable.
2005a
Kales[37]
US
Assum
ptionMetho
dBiases:Racial/e
thnic
andGen
der.
321psychiatrists
Atten
dees
atthe2002
annu
almeetin
gof
the
American
Psychiatric
Associatio
n.
Not
specified
.Patients’race
andge
nder
was
notassociated
with
sign
ificant
differences
inthediagno
sesof
major
depression
.How
ever,w
hite
patients
wereratedas
beingof
sign
ificantlyhigh
erSES
than
blackpatients.Asign
ificant
relatio
nship
was
foun
dbe
tweenratin
gof
SESand
estim
ates
ofpatient
demeano
ur(lower
SES
associated
with
moreho
stile
demeano
ur).
2005b
Kales[38]
US
Assum
ptionMetho
dBiases:Racial/e
thnic
andGen
der.
178Prim
aryCare
Physicians
(PCPs)
Atten
dees
atthe2002
American
Acade
myof
Family
Physicians
Ann
ual
Meetin
g.
Not
specified
.Patients’race
andge
nder
was
notassociated
with
sign
ificant
differences
inthediagno
sesof
major
depression
.How
ever,w
hite
patients
wereratedas
beingof
sign
ificantlyhigh
erSESthan
blackpatients.
2009a
Lutfey
a[29]
US,UKand
Germany
Assum
ptionMetho
dBiases:A
ge,G
ende
r,Racial/ethnicandSES.
384ph
ysicians
(internistsor
family
practitione
rsin
the
USandGermanyand
gene
ralp
ractition
ersin
theUK)
Screen
ingteleph
one
calls.
64.9%
intheUS,
59.6%
intheUK,
and65.0%
inGermany
Physicians
wereleastcertainof
CHD
diagno
seswhe
npatientswereyoun
gerand
female.Certainty
was
positivelycorrelated
with
severalclinicalactio
ns,including
test
orde
ring,
prescriptio
ns,referralsto
specialists,
andtim
eto
follow-up.
2009b
Lutfey
a[35]
US
Assum
ptionMetho
dBiases:A
ge,G
ende
r,Racial/ethnicandSES
128ge
neralistph
ysicians
Screen
ingteleph
one
calls.
64.9%
Physicians
wereleastcertainof
theirCHD
diagno
sesforblackpatientsandforyoun
ger
wom
en.Physiciansrespon
deddifferentially
todiagno
sticcertaintyin
term
sof
theirclinical
therapeutic
actio
nssuch
astestorde
ringand
writingprescriptio
ns.
2010
Lutfey
[33]
US
Assum
ptionMetho
dBiases:A
ge,G
ende
r,Racial/ethnicandSES.
256internistsor
family/
gene
ralp
ractition
ers
Screen
ingteleph
one
calls.
Not
specified
.Ph
ysicians
prim
edwith
CHDweremorelikely
toorde
rCHD-related
testsandprescriptio
ns.
Maineffectsforpatient
gend
erandage
remaine
d,sugg
estin
gthat
physicians
treated
thesede
mog
raph
icvariables
asdiagno
stic
features
indicatin
glower
riskof
CHDfor
thesepatients.
2009a
Maserejiana
[34]
US
Assum
ptionMetho
dBiases:A
ge,G
ende
r,Racial/ethnicandSES.
128ph
ysicians
(internist
orfamily
practitione
r)Screen
ingteleph
one
calls.
Not
specified
.Ph
ysicians
weresign
ificantlyless
certainof
the
unde
rlyingcauseof
symptom
sam
ongfemale
patientsregardless
ofage,bu
ton
lyam
ong
middle-aged
wom
enwerethey
sign
ificantly
less
certainof
theCHDdiagno
sis.
2009b
Maserejian[36]
US
Assum
ptionMetho
dBiases:A
ge,G
ende
r,Racial/ethnicandSES.
256internistsor
family/
gene
ralp
ractition
ers
Screen
ingteleph
one
calls.
Not
specified
.48%
ofph
ysicians
wereinconsistent
intheir
popu
latio
n-leveland
individu
al-levelC
HD
assessmen
ts.Physicians’assessmen
tsof
CHD
prevalen
cedidno
tattenu
atetheob
served
gend
ereffect
indiagno
sticcertaintyforthe
individu
alpatient.
FitzGerald and Hurst BMC Medical Ethics (2017) 18:19 Page 7 of 18
Table
1Stud
iesinclud
edin
thesystem
aticreview
(Con
tinued)
2006
McKinlaya
[28]
US/UK
Assum
ptionMetho
dBiases:A
ge,G
ende
r,Race
andSES.
256prim
arycare
physicians
intheUSandtheUK
Screen
ingteleph
one
calls.
64.9%
intheUS
and59.6%
intheUK.
Age
,race,ge
nder,b
utno
tSES,influen
ced
decision
-makingforthecond
ition
sstud
ied
inbo
thcoun
tries.Differen
ceswerealso
foun
dbe
tweentreatm
entde
cision
sin
theUKand
theUS.
2007
McKinlaya
[30]
US
Assum
ptionMetho
dBiases:A
ge,G
ende
r,Racial/ethnicandSES.
128internistsandfamily
physicians
Screen
ingteleph
one
calls.
64.9%
Femalepatientswereless
likelythan
males
toreceive4of
5type
sof
physicalexam
ination;
olde
rpatientswereless
likelyto
beadvisedto
stop
smoking.
Race
andSESof
patientshad
noeffect
onprovider
adhe
renceto
guidelines.
2003
Tamayo-Sarver
[40]
US
Assum
ptionMetho
dBiases:Racial/e
thnic
2872
emerge
ncyph
ysicians
Mail-o
ut.
53%
Therace/ethnicity
ofpatientsin
thevign
ettes
hadno
effect
onph
ysicianprescriptio
nof
opioids.Makingsociallyde
sirableinform
ation
explicitincreasedtheprescribingratesby
4%forthemigrainevign
ette
and6%
fortheback
pain
vign
ette.
Racial/ethnic
2011
Barnato[41]
US
Assum
ptionMetho
d33
hospital-b
ased
physicians
(emergencyph
ysicians,
hospitalists,intensivists)
Prob
ability
sampling
(15)
andconven
ience
sampling(18).
Repo
rted
as‘low’.
Physicians
didno
tmakedifferent
treatm
ent
decision
sforblackandwhite
patients,de
spite
believing
that
blackpatientsweremorelikely
toprefer
intensive,life-sustaining
treatm
ent.
2013a
Blair[44]
US
IAT(prejudice)a
ndinterpersonalinteractio
nmeasures
134clinicians
Dataforclinicians
collected
intheBlair[44]
stud
y.Prim
arydata
from
patientsin
abroade
rstud
yon
hype
rten
sion
.
60%
Clinicians
with
greaterim
plicitbias
against
blacks
wereratedlower
inpatient-cen
tred
care
bytheirblackpatientsas
comparedwith
areferencegrou
pof
white
patients.
2013b
Blair[22]
US
IAT(prejudice)
210expe
rienced
prim
ary
care
providers(PCPs)
Invitatio
nlaun
ched
with
presen
tatio
ns.
60%
Substantialimplicitbias
againstLatin
osand
AfricanAmericansin
PCPs
2008
Burgess[43]
US
Assum
ptionMetho
d382ge
neralinternal
med
icineph
ysicians
Mail-o
ut.
40%
Therewas
nosign
ificant
effect
ofpatient
race
alon
e.Amon
gblackpatients,ph
ysicians
were
sign
ificantlymorelikelyto
statethat
they
wou
ldsw
itchto
ahigh
erdo
seor
strong
erop
ioid
forpatientsexhibitin
g“challeng
ing”
behaviou
rscomparedto
thoseexhibitin
g“non
-challeng
ing”
behaviou
rs.
2012
Coo
per[23]
US
IAT(prejudice
and
stereo
type
),audiotape
measuresof
visit
commun
icationand
patient
ratin
gs.
40prim
arycare
clinicians
(36ph
ysicians,fou
rnu
rses)
inurbancommun
ity-based
practices.
Second
aryda
tafrom
twoprevious
stud
ies,
whe
repa
tientsand
providerspa
rticipated
inrand
omised
clinical
trialsof
interven
tions
toen
hance
commun
ication.
63%
Clinicianim
plicitrace
bias
andrace
and
compliancestereo
typing
wereassociated
with
markersof
poor
visitcommun
ication
andpo
orratin
gsof
care,p
articularlyam
ong
blackpatients.
FitzGerald and Hurst BMC Medical Ethics (2017) 18:19 Page 8 of 18
Table
1Stud
iesinclud
edin
thesystem
aticreview
(Con
tinued)
2007
Green
[46]
US
IAT(prejudice
and
stereotype)and
vign
ettes
287ph
ysicians
Email.
50.6%
All3IATs
show
edsign
ificant
race
bias.
Physicians
weremorelikelyto
diagno
seblack
patientsthan
white
patientswith
CAD.
How
ever,p
hysicianswereeq
ually
likelyto
give
thrombo
lysisforblackandwhite
patients.Therewas
thus
aracialdisparity
inthrombo
lysisrelativeto
CADdiagno
sis.
2012
Moskowitz
[20]
US
Sublim
inalprim
ing
(faces).
Stud
y1:16
physicians.
Stud
y2:11
physicians
Con
venien
cesample.
Not
specified
.Whe
nprim
edwith
ablackface,p
hysicians
reactedmorequ
icklyforstereo
typical
diseases,ind
icatingan
implicitassociationof
certaindiseases
with
blackpatients.These
comprised
noton
lydiseases
that
black
patientsarege
neticallypred
ispo
sedto,b
utalso
cond
ition
sandsocialbe
haviou
rswith
nobiolog
icalassociation(e.g.obe
sity,d
ruguse).
2010
Penn
er[24]
US
IAT(prejudice)a
ndinteractionmeasures
15resid
entph
ysicians
(3white
and12
self-identified
asIndian,PakistaniorA
sian)
Patientsrecruited
consecutivelyand
physicians
invited.
83%
physicians
Overall,ph
ysicians
didno
tdisplayim
plicitrace
bias.H
owever,b
lack
patientshadless
positive
reactio
nsto
med
icalinteractions
with
physicians
relativelylow
inexplicitbu
trelativelyhigh
inim
plicitbias
than
tointeractions
with
physicians
who
wereeither
(a)low
inbo
thexplicitandim
plicitbias,or(b)
high
inbo
thexplicitandim
plicitbias.
2008
Sabinb
[47]
US
IAT(prejudice
and
stereo
type
)and
vign
ettes
86academ
icpaed
iatricians
from
onede
partmen
tInvitedallfaculty,
reside
ntsandfellowsat
alarge,urbanresearch
university
toparticipate.
58%
Paed
iatricians
held
implicitrace
bias,b
utit
was
weakerthan
that
ofothe
rMDsand
othe
rsin
society.
2009
Sabin[67]
Worldwide
IAT(prejudice)
2535
MDs
ProjectIm
plicitweb
site.
Not
applicable
Med
icaldo
ctors,liketherestof
thesample,
show
edastrong
implicitpreferen
cefor
whitesover
blacks.
2012a
Sabinb
[45]
US
IAT(prejudice
and
stereo
type
)and
vign
ettes
86academ
icpaed
iatricians
from
onede
partmen
t.Invitedallfaculty,
reside
ntsandfellowsat
alarge,urbanresearch
university
toparticipate.
58%
Paed
iatricians’implicitbias
was
associated
with
treatm
entrecommen
datio
ns.A
spaed
iatricians’implicitpro-white
bias
increased,
prescribingnarcoticmed
ication
decreasedforblackpatients,bu
tno
tfor
white
patients.
2012
Step
anikova[21]
US
Sublim
inalprim
ing
(words)andvign
ettes
81family
physicians
and
gene
ralinternists
Email.
2%Und
erhigh
ertim
epressure,b
utno
tlower,
implicitbiases
againstblacks
andHispanics
ledto
less
serio
usdiagno
sis.Und
erhigh
ertim
epressure,implicitbias
againstblacks
led
tolower
rate
ofreferralto
specialist.
Weigh
t
2012b
Sabin[60]
Worldwide
IAT(prejudice)
2284
med
icaldo
ctors(M
Ds)
ProjectIm
plicitweb
site.
Not
applicable.
MDs,likethewider
popu
latio
ntested
,had
astrong
implicitanti-fatbias
andastrong
explicitanti-fatbias.
FitzGerald and Hurst BMC Medical Ethics (2017) 18:19 Page 9 of 18
Table
1Stud
iesinclud
edin
thesystem
aticreview
(Con
tinued)
2003
Schw
artz[50]
Canada
IAT(prejudice
and
stereo
type
)389(122
physicians,12
psycho
logists,fivenu
rses,
18othe
rob
esity
clinicians)
Atten
dees
oftheAnn
ual
Meetin
gof
theNorth
American
Associatio
nfor
theStud
yof
Obe
sity.
Not
specified
.Therewas
asign
ificant
implicitanti-fat
prejud
iceandstereo
type
foun
d.
2007
Vallis[68]
Canada
IAT(prejudice
and
stereo
type
)78
(14.3%
physicians,
15.4%
nurses)
Atten
dees
ofon
obesity
conferen
ce.
86%
oftotal
attend
ees.
Strong
eviden
ceforanti-fatprejud
ice
andstereo
type
.a W
hatap
pear
tobe
thesameda
tafrom
theUS,theUKan
dGerman
yareselectivelyan
alysed
indifferen
twaysin
theseeigh
tstud
ies
bWha
tap
pear
tobe
thesameda
taarean
alysed
indifferen
twaysin
thesetw
ostud
ies
FitzGerald and Hurst BMC Medical Ethics (2017) 18:19 Page 10 of 18
Table 2 Main characteristics of studies
Studies (N = 42)
Method
Assumption method 25 [25–41, 43, 58, 59, 61, 63–66]
Implicit measure 17 [20–24, 42, 44–50, 60, 62, 67, 68]
Implicit measure - IAT 15 [22–24, 42, 44–50, 60, 62, 67, 68]
(of which: IAT combined with behaviouror decision)
9 [23, 24, 42, 44–48, 67]
Implicit measure - Subliminal priming 2 [20, 21]
Setting and type of test
IAT – implicit prejudice 15 [22–24, 42, 44–50, 60, 62, 67, 68]
IAT – implicit stereotype 5 [23, 45–47, 50]
IAT – standard 13 [22–24, 42, 44–47, 50, 60, 62, 67, 68]
IAT – Single Category 2 [48, 49]
IAT – uncontrolled setting 10 [22, 23, 42, 45–47, 50, 60, 62, 67]
IAT - controlled laboratory setting 3 [48, 49, 68]
IAT – setting unspecified 2 [24, 44]
Assumption method – video vignette with oral questions 10 [27–36]
Assumption method – written texts 11 [25, 26, 40, 43, 58, 59, 61, 63–66]
(of which: photos in addition) 1 [43]
Assumption method – video vignette with written questions 3 [37–39]
Assumption method – simulations of encounters with patients and role-play 1 [41]
Assumption method – controlled setting 16 [27–39, 41, 58, 63]
Assumption method – uncontrolled setting 8 [25, 26, 40, 43, 61, 64–66]
Assumption method – setting unspecified 1 [59]
Bias(es) studied
Racial/ethnic 27 [20–24, 27–41, 43–47, 67]
Multiple 14 [27–39, 66]
Gender 14 [27–38, 63, 66]
Socio-economic status (SES) 11 [27–36, 39]
Age 11 [25, 27–36]
Mental illness 4 [26, 42, 64, 65]
Weight 3 [50, 60, 68]
Brain-injured patients perceived to have contributed to their injury 2 [58, 59]
Intravenous drug users 2 [48, 49]
Disability 1 [62]
AIDS patients 1 [61]
Social circumstances (desiring an active lifestyle, having a demandingcareer, having family demands)
1 [66]
Country of study
US 27 [20–24, 27–41, 43–47, 62, 66]
UK 8 [26–29, 31, 32, 58, 59]
Compared countries (US, UK and Germany) 4 [27–29, 31]
Worldwide 3 [42, 60, 67]
France 2 [25, 65]
Australia 2 [48, 49]
Germany 2 [29, 31]
FitzGerald and Hurst BMC Medical Ethics (2017) 18:19 Page 11 of 18
Table 2 Main characteristics of studies (Continued)
Canada 2 [50, 68]
Israel 1 [63]
Hong Kong 1 [64]
China 1 [61]
Participants (N = 15148)
Profession of participants
Physicians 12156
Nurses 740
Either physicians or nurses 1404
‘Clinicians’, or ‘mental health professionals’ (at least some of whomwere nurses and physicians)
834
Psychologists 12
Medical Students 2
FitzGerald and Hurst BMC Medical Ethics (2017) 18:19 Page 12 of 18
this is less worrying because IAT effects are to someextent uncontrollable.
Of the fieldImplicit bias in healthcare is an emerging field of researchwith no established methodology. This is to be expectedand is not a problem in itself, but it does present an obs-tacle when conducting a review of this kind. The range ofmethods used and the variety of journals with differingstandards and protocols for describing experiments madeit difficult to compare the results. In addition, authors fo-cusing on a particular bias (e.g. gender), often in combin-ation with a particular health issue (e.g. heart disease),frequently did not appear to be familiar with one another’sresearch. This lack of familiarity meant that often used dif-ferent terms to describe the same phenomenon, whichalso made conducting the review more difficult.
Table 3 Limitations of specific studies
Recruitment method not reported
Failed to report response rate
Response rate reported as ‘low’
Response rate less than 40%
Explicitly informed participants of the purpose of the study
Gave participants tests or questionnaires that indicatedpatient characteristic under scrutiny prior to vignette
Did not specify whether they informed participants aboutthe purpose of the study
Small sample size
Failed to report calculating power when designing study
Full information on statistical methods used not provided
Few of the existing results can be described as ‘realworld’ treatment outcomes. The two priming studiesinvolved very small samples and were more exploratorythan result-seeking [20, 21]. The IAT and assumptionstudies were conducted under laboratory conditions.The only three studies conducted in naturalistic settingscombined the IAT with measures of physician-patientinteraction [22–24]. However, many of the assumptionstudies attempted to make their vignettes as realistic aspossible by having them validated by clinicians [25–41]and also by having participants view/read the vignettesas part of a normal day at work [27–36, 39, 41].Because the studies of interest used psychological
techniques, but were mainly to be found in a medicaldatabase (PubMed), the classification of the studies wasnot always optimal. There is no heading in Medline for‘implicit bias’ and studies using similar methods weresometimes categorized under different subject headings,
1 [59]
12 [20, 33, 34, 36–39, 48–50, 59, 63]
1 [41]
7 [21, 26, 43, 58, 62, 64, 65]
7 [25, 27, 32, 42, 58, 60, 67]
2 [48, 49]
16 [22, 24, 28–31, 34–36, 39, 44, 59, 61, 62, 64, 66]
3 [20, 21, 48]
All studies failed except the 15 referenced here thatdid [27–36, 39, 40, 42, 43, 59]
4 [32, 49, 61, 63]
FitzGerald and Hurst BMC Medical Ethics (2017) 18:19 Page 13 of 18
some of which were introduced during the last ten years,which increased the risk of missing eligible studies.
Existence of implicit biases/stereotypes in healthcareprofessionals and influence on quality of careHealthcare professionals have implicit biasesAlmost all studies found evidence for implicit biasesamong physicians and nurses. Based on the available evi-dence, physicians and nurses manifest implicit biases toa similar degree as the general population. The followingcharacteristics are at issue: race/ethnicity, gender, socio-economic status (SES), age, mental illness, weight, havingAIDS, brain injured patients perceived to have contributedto their injury,3 intravenous drug users, disability, andsocial circumstances.Of the seven studies that did not find evidence of bias,
one compared the mentally ill with another potentiallyunfavourable category, welfare recipients; this study didfind a positive correlation between levels of implicit biasand over-diagnosis of the mentally ill patient in thevignette [42]. Another used simulated interactions withactors, which may result in participants being on ‘bestbehaviour’ in the role-play [41]. The two studies that re-ported no evidence of bias in diagnosis of depressionfound that physicians’ estimates of SES were influencedby race (lower SES estimated for black patients); [37, 38]one reported that estimates of SES in turn were signifi-cantly related to estimates of patient demeanour (lowerSES associated with hostile patient demeanour) [37]. Afurther study failed to find differences due to patientrace in the prescription of opioids, but found an inter-action whereby black patients who exhibited ‘challen-ging’ behaviour (such as belligerence and asking for aspecific opioid) were more likely to be prescribed opioidsthan those who did not, an effect possibly due to a racialstereotype [43]. Another study that failed to find implicitrace bias suggested that this was due to the setting ofthe study in an inner-city clinic with high levels of blackpatients and the fact that many physicians were bornoutside the US [24]. Finally, one study that found no evi-dence of racial bias in prescription of opioid analgesicspresented each participant with three vignettes depictingpatients of three different ethnicities, thus probablyalerting them to the objective of the study [40].The interaction effects between different patient char-
acteristics in assumption studies are varied and a few aresurprising. The authors of one study expected that phy-sicians would be less likely to prescribe a higher dose ofopioids to black patients who exhibited challenging be-haviours; in fact, physicians were more likely to pre-scribe higher doses of opioids to challenging blackpatients, yet slightly less likely to do so to white patientsexhibiting the same behaviour. Sometimes significant
effects on the responses to the vignette of a patient char-acteristic, e.g. race, are only found when the interactionbetween gender and race or SES and race is examined.For example, physicians in one study were less certain ofthe diagnosis of coronary heart disease for middle-agedwomen, who were thus twice as likely to receive a men-tal health diagnosis than their male counterparts [34]. Inanother, low SES Latinas and blacks were more likely tohave intrauterine contraception recommended than lowSES whites, but there was no effect of race for high SESpatients [39].
Implicit bias affects clinical judgement and behaviourThree studies found a significant correlation betweenhigh levels of physicians’ implicit bias against blacks onIAT scores and interaction that was negatively rated byblack patients [23, 24, 44] and, in one study, also nega-tively rated by external observers [23]. Four studiesexamining the correlation between IAT scores andresponses to clinical vignettes found a significant correl-ation between high levels of pro-white implicit bias andtreatment responses that favoured patients specified aswhite [42, 45–47]. In one study, implicit prejudice ofnurses towards injecting drug users significantly mediatedthe relationship between job stress and their intention tochange jobs [48].Twenty out of 25 assumption studies found that some
kind of bias was evident either in the diagnosis, thetreatment recommendations, the number of questionsasked of the patient, the number of tests ordered, orother responses indicating bias against the characteristicof the patient under examination.
Determinants of biasSocio-demographic characteristics of physicians andnurses (e.g. gender, race, type of healthcare setting, yearsof experience, country where medical training received)are correlated with level of bias. In one study, male staffwere significantly less sympathetic and more frustratedthan female staff with self-harming patients presentingin A&E [26]. Black patients in the US –but not the UK-were significantly more likely to be questioned aboutsmoking than white [28]. In another study, internationalmedical graduates rated the African-American malepatient in the vignette as being of significantly lowerSES than did US graduates [38]. One study found thatpaediatricians held less implicit race bias comparedwith other MDs [47].Correlations between explicit and implicit attitudes
varied depending on the type of bias and on the kind ofexplicit questions asked. For instance, implicit anti-fatbias tends to correlate more with an explicit anti-fat biasthan racial bias, where explicit and implicit attitudesoften diverge significantly. Because physicians’ and nurses’
FitzGerald and Hurst BMC Medical Ethics (2017) 18:19 Page 14 of 18
implicit attitudes diverged frequently from their expli-cit attitudes, explicit measures cannot be used aloneto measure the presence of bias among healthcareprofessionals.
DiscussionA variety of studies, conducted in various countries, usingdifferent methods, and testing different patient character-istics, found evidence of implicit biases among healthcareprofessionals and a negative correlation exists betweenlevel of implicit bias and indicators of quality of care. Thetwo most common methods employed were the assump-tion method and the IAT, the latter sometimes combinedwith another measure to test for correlations with the be-haviour of healthcare professionals.Our study has several limitations. Four studies in-
cluded participants who were not trained physicians ornurses and failed to report separate results for thesecategories of participants [42, 44, 49, 50]. Since eitherthe majority of participants were qualified physiciansand nurses, or were other health care professionals in-volved in patient care, we included these studies despitethis limitation. Excluding them would not have changedthe conclusions of this paper. In addition, we initiallycentred our research on studies employing implicit mea-sures recognised in psychology, but the majority of theincluded studies in the final review used the assumptionmethod. However, the limitations imposed by the lack ofconsistency in keywords and categorization of articlesactually worked in our favour here, enabling us to cap-ture a variety of methods and thus to consider includingthe assumption method. Scanning the references of thearticles that were initially retained and repeating thisprocess until there were no new articles helped us tocapture further pertinent articles. From the degree ofcross-referencing we are confident that we succeeded inidentifying most of the relevant articles using theassumption method.Publication bias could limit the availability of results
that reveal little or no implicit bias among healthcareprofessionals. Moreover, eight articles appeared to referto the same data collected in a single cross-countrycomparison study [27–32, 34, 35] and a further twoarticles analysed the same data [45, 47]. The sum of 42articles thus can give the impression that more researchhas been carried out on more participants than is thecase. The solidity of data revealing high levels of implicitbias among the general population suggest that this isunlikely to have invalidated the conclusion that implicitbias is present in healthcare professionals [6, 7].However, our decision to exclude studies that involved
students rather than fully-trained healthcare profes-sionals meant that we did not include a study conductedon medical students that showed no significant association
between implicit bias and clinical assessments [51].Several studies post 2013 (thus after our cut-off date)have also indicated a null relationship between levels ofimplicit bias and clinical decision-making [52–54]. Thescientific community working in this area agrees thatthe relationship between levels of implicit bias inhealthcare professionals and clinical decision-making iscomplex and that there is currently a lack of good evi-dence for a direct negative influence of biases [16, 17].As our review shows, there is clearer evidence for arelationship between implicit bias and negative effectson clinical interaction [23, 24, 44]. While this may notalways translate into negative treatment outcomes, therelationship between a healthcare professional and herpatient is essential to providing good treatment, thus itseems likely that the more negative the clinical inter-action, the worse the eventual treatment outcome (notto mention the likelihood that the patient will consulthealthcare services for future worries or problems).This is where the bulk of future research should beconcentrated.The interactions between multiple patient characteris-
tics and between healthcare professional and patientcharacteristics reveal the complexity of the phenomenonof implicit bias and its influence on clinician-patientinteraction. They also highlight the pertinence of workin feminist theory on ‘intersectionality’, a term for thedistinctive issues that arise when a person belongs tomultiple identity categories that bring disadvantage, suchas being both black and female [55]. For instance, onestudy only found evidence of bias against low SES Latinapatients, not against high SES Latinas, illustrating howbelonging to more than one category (here, both lowSES and Latina) can have negative effects that are notpresent if membership of one category is eliminated(here, low SES) [39]. Class may trump race in some cir-cumstances so that being high SES is more salient thanbeing non-white. One criticism of mainstream feminismby theorists who work on intersectionality is that per-tinent issues are unexplored because of the dominanceof high SES white women in feminist theory. Using ourexample from the review, high SES Latina women maynot experience the same prejudice as low SES Latinawomen and thus may falsely assume that there is noprejudice against Latina women tout court in this con-text. This could be frustrating for low SES Latinawomen who have unrecognized lived experiences ofprejudice in a clinical setting.In some studies, the attitudes of patients towards
healthcare professionals were recorded and used toevaluate clinical interaction [23, 24, 44]. It is importantto remember that patients also may come to a clinicalinteraction with biases. In these cases, the biases of oneparticipant may trigger the biases of the other, magnifying
FitzGerald and Hurst BMC Medical Ethics (2017) 18:19 Page 15 of 18
the first participant’s biased responses and leading to asnowball effects [56]. Past experience of discriminationmay mean that a patient may come to an interaction withnegative expectations [57].Our findings in the review suggest that the relation-
ship between training and experience and levels of im-plicit bias is mixed. In one study, increased contact withpatients with Hepatitis C virus was associated with morefavourable explicit attitudes, yet more negative implicitattitudes towards intravenous drug users [49]. Anotherstudy demonstrated that nursing students were less pre-judiced, more willing to help and desired more socialinteraction with patients with brain injury, when com-pared with qualified nurses [58]. Exposure to communi-cation skills training was not associated with lowerrace-IAT scores for physicians [23]. However, individ-uals with mental health training demonstrated morepositive implicit and explicit evaluations of people withmental illness than those without training [42]. Yet inthe same study, graduate students had more positiveimplicit attitudes towards the mentally ill than mentalhealth professionals.We included all types of implicit bias in our review,
not only race bias, partly in an effort to capture non-USstudies, hypothesising that the focus on race in the USleaves fewer resources for investigation into other biases.It is possibly the case that a wider range of biases wereinvestigated in non-US countries, but there is notenough evidence to deduce this from our review alone.For instance, two British studies examine bias againstbrain-injured patients who are perceived as having con-tributed to their injury [58, 59], and two Australianstudies looked at bias against intravenous drug users[48, 49], but the sample size of studies is too small towarrant drawing any conclusions from this.Is it possible that there are implicit associations that
are justified because they are based on prevalence datafor diseases? One study in our review aimed to test thestatistical discrimination hypothesis by asking physi-cians to estimate the prevalence data among males andfemales for coronary heart disease in addition topresenting them with vignettes of a female or male cor-onary heart disease patient. It found that 48% of physi-cians were inconsistent in their population-level andindividual level assessments and that the physicians’gender-based population prevalence assessments werenot associated with the certainty of their diagnosis ofcoronary heart disease. There was no evidence to sup-port the theory of statistical discrimination as an ex-planation for why physicians were less certain of theirdiagnoses of CHD in women [36]. Another exploratorystudy looked at the diseases that were stereotypicallyassociated with African-Americans and found thatmany diseases were associated with African-Americans
that did not match prevalence data, such as drug abuse[20]. The danger in these cases is that a physician mayapply a group-level stereotype to an individual and fail tofollow-up with a search for individuating information.Impartial treatment of patients by healthcare profes-
sionals is an uncontroversial norm of healthcare. Impli-cit biases have been identified as one possible factor inhealthcare disparities and our review reveals that theyare likely to have a negative impact on patients fromstigmatized groups. Our review also indicates that theremay sometimes be a gap between the norm of imparti-ality and the extent to which it is embraced by health-care professionals for some of the tested characteristics.For instance, explicit anti-fat bias was found to beprevalent among healthcare professionals [60]. Sinceweight can be relevant to diagnosis and treatment, it isunderstandable that it is salient. It is nonethelessdisturbing that healthcare professionals exhibit thesame explicit anti-fat attitudes prevalent in the generalpopulation.The most convincing studies from our review are
those that combine the IAT and a method measuringthe quality of treatment in the actual world. Thesestudies provide some evidence for a relationship be-tween bias as measured by the IAT and behaviour byclinicians that may contribute to healthcare disparities.More studies using real-world interaction measureswould be helpful because studies using vignettes remainopen to the criticism that they do not reveal the truebehaviour of healthcare professionals. In this respect,the three studies using measures of physician-patientinteraction are exemplary [22–24], in particular whenusing independent evaluators of the interactions [23].Overall, our review reveals the need for discussion ofmethodology and for more interaction between differ-ent literatures that focus on different biases.
ConclusionOur findings highlight the need for the healthcare pro-fession to address the role of implicit biases in disparitiesin healthcare. In addition to addressing implicit biases,measures need to be taken to raise awareness of the po-tential conflict between holding negative explicit atti-tudes towards some patient characteristics, such asobesity, and committing to a norm to treat all patientsequally.Our review reveals that this is an area in need of more
uniform methods of research to enable better compari-son and communication between researchers interestedin different forms of bias. Important avenues for furtherresearch include examination of the interactions be-tween patient characteristics, and between healthcareprofessional and patient characteristics, and of possible
FitzGerald and Hurst BMC Medical Ethics (2017) 18:19 Page 16 of 18
ways in which to tackle the presence of implicit biases inhealthcare.
Endnotes1There are conceptual problems with this distinction
as used in psychology that have been pointed out by phi-losophers, but we will ignore these for the purposes ofthis review.
2Interestingly, physicians were also asked for how theyexpected their colleagues to rate the vignette, and inthese ratings there was a negative bias towards bothpatients from Ghana and from Serbia.
3Bias against patients who are seen as contributing totheir injury initially seems to be an odd category com-pared to the more familiar ones of race and gender.Clinicians may treat brain injured patients differently ifthey are somehow seen as ‘responsible’ for their injury,for instance, if they were engaging in risk-taking behavioursuch as drug taking. Our review was intended to capturestudies such as these that identify biases that are spe-cific to clinical contexts and thus of particular interestto clinicians.
Appendix 1Search StrategyPubmed
� The following combination of subject headings andfree text terms was used:
(“Prejudice” [MAJR] AND “Attitude of healthpersonnel” [MAJR]) OR (“Attitude of healthpersonnel/ethnology” [MH] AND “Prejudice”[MH])OR (“Stereotyping”[MH] AND “Attitude of healthpersonnel”) OR (“Prejudice”[MH] AND “Healthcaredisparities” [MH]) OR (“Prejudice”[MH] AND“Cultural Competency” [MH]) OR (“Social Class”[MH] AND “Attitude of health personnel” [MH])OR (“Prejudice”[MH] AND “Physicians” [MH]) OR(“Prejudice”[MAJR] AND “Delivery of HealthCare”[MAJR] AND “stereotyping”[MAJR]) OR(“Physician-Patient Relations” [MH] AND “healthstatus disparities”[MH]) OR (“Prejudice”[MH] AND“Obesity”[MH]) OR (“African Americans/psychology” [MH] AND “Healthcare disparities”[MH]) OR (“Prejudice”[MH] AND “Mentally IllPersons”[MH]) OR (“Prejudice”[MH] AND“Women’s Health”[MH]) OR “aversive racism” OR“anti-fat bias” OR “racial-ethnic bias” OR “racial-ethnic biases” OR “ethnic/racial bias” OR “ethnic/racial biases” OR (“disabled persons”[MAJR] AND“prejudice”[MAJR])� Dates: 1st March 2003 to 31st March 2013� Final number of retrieved articles: 2510
PsychINFO and PsychARTICLE
� The following combination of subject headings andfree text terms was used was used:
Health personnel AND (prejudice OR bias)� Dates: 1st March 2003 to 31st March 2013� Other filters: Scholarly journals� Final number of retrieved articles: 377� Final result when duplicates removed: 360.
CINAHL
� The following combination of subject headings andfree text terms was used was used:
Prejudice [MM Exact Major Subject Heading] ORstereotyping [MM Exact Major Subject Heading]OR Discrimination [MM Exact Major SubjectHeading] OR implicit bias OR unconscious bias� Dates: 1st March 2003 to 31st March 2013� Other filters:
– Exclude Medline records– Peer reviewed
� Final number of retrieved articles: 897
AcknowledgementsNot applicable. Only the two authors were implicated in the review.
FundingThis work was carried out with the support of grants from the Swiss NationalScience Foundation under grants numbers: PP00P3_123340 and 32003B_149407.
Availability of data and materialsThe search strategy is available in the Appendix to the paper.
Authors’ contributionsBoth authors discussed to select the databases and decide on the researchquestion, based on CF’s knowledge of the field of implicit bias and SH’sknowledge of systematic reviews and bioethics literature. CF compiled thekey words for the search strategy with constant advice and input from SH.CF drafted the inclusion criteria and received constant input on this from SH:CF carried out the search and downloaded the relevant articles to bescrutinised. CF and SH both independently read all the initial titles to selectwhich were relevant, then the abstracts, and then the final included articlesand discussed at each stage to resolve any disagreements. CF drafted theinitial tables including the information from the studies and this was revised bySH. SH particularly revised the statistical methods used by the studies and bothreviewed their methodology. CF drafted the manuscript and it was revised withcomments by SH a number of times until both authors were satisfied with themanuscript. Both authors read and approved the final manuscript.
Competing interestThe authors declare that they have no competing interests.
Ethics approval and consent to participateNot applicable.
Received: 19 October 2016 Accepted: 14 February 2017
References1. Holroyd J, Sweetman J. The Heterogeneity of Implicit Bias. In Brownstein,
Michael, Saul, Jennifer, editors. Implicit Bias and Philosophy, Volume 1:Metaphysics and Epistemology 1. Oxford: Oxford University Press; 2016.p. 80–103.
FitzGerald and Hurst BMC Medical Ethics (2017) 18:19 Page 17 of 18
2. Blair IV, Steiner JF, Havranek EP. Unconscious (implicit) bias and healthdisparities: where do we go from here? Perm J. 2011;15:71.
3. Ambady N, Shih M, Kim A, Pittinsky TL. Stereotype susceptibility in children:effects of identity activation on quantitative performance. Psychol Sci. 2001;12:385–90.
4. Louise A. Bias: Friend or Foe? Reflections on Saulish Skepticism. InBrownstein, Michael, Saul, Jennifer, editors. Implicit Bias and Philosophy,Volume 1: Metaphysics and Epistemology 1. Oxford: Oxford University Press;2016. p. 157–190.
5. Nosek BA, et al. Pervasiveness and correlates of implicit attitudes andstereotypes. Eur Rev Soc Psychol. 2007;18:36–88.
6. Nosek BA, Riskind RG. Policy implications of implicit social cognition. SocIssues Policy Rev. 2012;6:113–47.
7. Jost JT, et al. The existence of implicit bias is beyond reasonable doubt: arefutation of ideological and methodological objections and executivesummary of ten studies that no manager should ignore. Res Organ Behav.2009;29:39–69.
8. Bertrand M, Mullainathan S. Are Emily and Greg more employable thanlakisha and Jamal? a field experiment on labor market discrimination. AmEcon Rev. 2004;94:991–1013.
9. Rooth D-O. Automatic associations and discrimination in hiring: real worldevidence. Labour Econ. 2010;17:523–34.
10. Dovidio JF, Gaertner SL. Aversive racism. Adv Exp Soc Psychol. 2004;36:1–52.11. Martin AK, Tavaglione N, Hurst S. Resolving the conflict: Clarifying’
Vulnerability’ in health care ethics. Kennedy Inst Ethics J. 2014;24:51–72.12. De-Shalit A, Wolff J. Disadvantage. Oxford: Oxford University Press; 2007.13. Burgess D, van Ryn M, Dovidio J, Saha S. Reducing racial bias among health
care providers: lessons from social-cognitive psychology. J Gen Intern Med.2007;22:882–7.
14. Stone J, Moskowitz GB. Non-conscious bias in medical decision making:what can be done to reduce it? Med Educ. 2011;45:768–76.
15. Shavers VL, et al. The state of research on racial/ethnic discrimination in thereceipt of health care. Am J Public Health. 2012;102:953–66.
16. Zestcott, C. A., Blair, I. V. & Stone, J. Examining the presence, consequences,and reduction of implicit bias in health care: A narrative review. GroupProcess Intergroup Relat. 2016. doi:10.1177/1368430216642029.
17. Hall WJ, et al. Implicit racial/ethnic bias among health care professionalsand its influence on health care outcomes: a systematic review. Am J PublicHealth. 2015;105:e60–76.
18. Drewniak D, Krones T, Sauer C, Wild V. The influence of patients’immigration background and residence permit status on treatmentdecisions in health care. Results of a factorial survey among generalpractitioners in Switzerland. Soc Sci Med. 2016;161:64–73.
19. De Houwer J, Moors A. How to define and examine the implicitness ofimplicit measures. In Wittenbrink B, Schwartz N. (eds.). Implicit measures ofattitudes: Procedures and controversies. New York: Guilford Press; 2007. pp.179–194.
20. Moskowitz GB, Stone J, Childs A. Implicit stereotyping and medicaldecisions: unconscious stereotype activation in practitioners’ thoughtsabout African americans. Am J Public Health. 2012;102:996–1001.
21. Stepanikova I. Racial-ethnic biases, time pressure, and medical decisions.J Health Soc Behav. 2012;53:329–43.
22. Blair IV, et al. Assessment of biases against Latinos and African americansamong primary care providers and community members. Am J PublicHealth. 2013;103:92–8.
23. Cooper LA, et al. The associations of clinicians’ implicit attitudes about racewith medical visit communication and patient ratings of interpersonal care.Am J Public Health. 2012;102:979–87.
24. Penner LA, et al. Aversive racism and medical interactions with blackpatients: a field study. J Exp Soc Psychol. 2010;46:436–40.
25. Protière C, Viens P, Rousseau F, Moatti JP. Prescribers’ attitudes towardelderly breast cancer patients. Discrimination or empathy? Crit Rev OncolHematol. 2010;75:138–50.
26. Mackay N, Barrowclough C. Accident and emergency staff’s perceptions ofdeliberate self-harm: attributions, emotions and willingness to help. Br J ClinPsychol. 2005;44:255–67.
27. Arber S, et al. Influence of patient characteristics on doctors’ questioningand lifestyle advice for coronary heart disease: a UK/US video experiment. BrJ Gen Pract. 2004;54:673–8.
28. McKinlay J, et al. How do doctors in different countries manage the samepatient? results of a factorial experiment. Health Serv Res. 2006;41:2182–200.
29. Lutfey KE, et al. Diagnostic certainty as a source of medical practicevariation in coronary heart disease: results from a cross-national experimentof clinical decision making. Med Decis Making. 2009;29(5):606–18.
30. McKinlay JB, et al. Sources of variation in physician adherence with clinicalguidelines: results from a factorial experiment. J Gen Intern Med. 2007;22:289–96.
31. Bönte M, et al. Women and men with coronary heart disease in threecountries: are they treated differently? Womens Health Issues. 2008;18:191–8.
32. Arber S, et al. Patient characteristics and inequalities in doctors’ diagnosticand management strategies relating to CHD: a video-simulationexperiment. Soc Sci Med. 2006;62:103–15.
33. Lutfey KE, Eva KW, Gerstenberger E, Link CL, McKinlay JB. Physician cognitiveprocessing as a source of diagnostic and treatment disparities in coronaryheart disease results of a factorial priming experiment. J Health Soc Behav.2010;51:16–29.
34. Maserejian NN, Link CL, Lutfey KL, Marceau LD, McKinlay JB. Disparities inphysicians’ interpretations of heart disease symptoms by patient gender:results of a video vignette factorial experiment. J Womens Health. 2009;18:1661–7.
35. Lutfey KE, Link CL, Grant RW, Marceau LD, McKinlay JB. Is certainty moreimportant than diagnosis for understanding race and gender disparities?:An experiment using coronary heart disease and depression case vignettes.Health Policy. 2009;89:279–87.
36. Maserejian NN, Lutfey KE, McKinlay JB. Do physicians attend to base rates?prevalence data and statistical discrimination in the diagnosis of coronaryheart disease. Health Serv Res. 2009;44:1933–49.
37. Kales HC, et al. Race, gender, and psychiatrists’ diagnosis and treatment ofmajor depression among elderly patients. Psychiatr Serv. 2005;56:721–8.
38. Kales HC, et al. Effect of race and Sex on primary care Physicians’ diagnosisand treatment of late-life depression. J Am Geriatr Soc. 2005;53:777–84.
39. Dehlendorf C, et al. Recommendations for intrauterine contraception:a randomized trial of the effects of patients’ race/ethnicity andsocioeconomic status. Am J Obstet Gynecol. 2010;203:319–e1.
40. Tamayo-Sarver JH, et al. The effect of race/ethnicity and desirable socialcharacteristics on physicians’ decisions to prescribe opioid analgesics. AcadEmerg Med. 2003;10:1239–48.
41. Barnato AE, et al. A randomized trial of the effect of patient race onphysician ICU and life-sustaining treatment decisions for an acutely unstableelder with end-stage cancer. Crit Care Med. 2011;39:1663.
42. Peris TS, Teachman BA, Nosek BA. Implicit and explicit stigma of mentalillness: Links to clinical care. J Nerv Ment Dis. 2008;196:752–60.
43. Burgess DJ, et al. Patient race and physicians’ decisions to prescribe opioidsfor chronic low back pain. Soc Sci Med. 2008;67:1852–60.
44. Blair IV, et al. Clinicians’ implicit ethnic/racial bias and perceptions of careamong black and Latino patients. Ann Fam Med. 2013;11:43–52.
45. Sabin JA, Greenwald AG. The influence of implicit bias on treatmentrecommendations for four common pediatric conditions: pain, urinary tractinfection, attention deficit hyperactivity disorder, and asthma. Am J PublicHealth. 2012;102:988–95.
46. Green AR, et al. Implicit bias among physicians and its prediction ofthrombolysis decisions for black and white patients. J Gen Intern Med. 2007;22:1231–8.
47. Sabin JA, Rivara FP, Greenwald AG. Physician implicit attitudes and stereotypesabout race and quality of medical care. Med Care. 2008;46:678–85.
48. von Hippel W, Brener L, von Hippel C. Implicit prejudice toward injectingdrug users predicts intentions to change jobs among drug and alcoholnurses. Psychol Sci. 2008;19:7–11.
49. Brener L, von Hippel W, Kippax S. Prejudice among health care workerstoward injecting drug users with hepatitis C: does greater contact lead toless prejudice? Int J Drug Policy. 2007;18:381–7.
50. Schwartz MB, Chambliss HO, Brownell KD, Blair SN, Billington C. Weight biasamong health professionals specializing in obesity. Obes Res. 2003;11:1033–9.
51. Haider AH, et al. Association of unconscious race and social class biaswith vignette-based clinical assessments by medical students. JAMA. 2011;306:942–51.
52. Oliver MN, Wells KM, Joy-Gaba JA, Hawkins CB, Nosek BA. Do physicians’implicit views of African americans affect clinical decision making? J AmBoard Fam Med. 2014;27:177–88.
53. Blair IV, et al. An investigation of associations between clinicians’ ethnic orracial bias and hypertension treatment, medication adherence and bloodpressure control. J Gen Intern Med. 2014;29:987–95.
FitzGerald and Hurst BMC Medical Ethics (2017) 18:19 Page 18 of 18
54. Hirsh AT, Hollingshead NA, Ashburn-Nardo L, Kroenke K. The interaction ofpatient race, provider bias, and clinical ambiguity on pain managementdecisions. J Pain. 2015;16:558–68.
55. Cole ER. Intersectionality and research in psychology. Am Psychol. 2009;64:170.
56. Burgess DJ, Fu SS, Van Ryn M. Why do providers contribute to disparitiesand what can be done about it? J Gen Intern Med. 2004;19:1154–9.
57. Dominicé Dao M. Vulnerability in the clinic: case study of a transculturalconsultation. J Med Ethics. 2016. doi:10.1136/medethics-2015-103337.
58. Linden MA, Redpath SJ. A comparative study of nursing attitudes towardsyoung male survivors of brain injury: a questionnaire survey. Int J Nurs Stud.2011;48:62–9.
59. Redpath SJ, et al. Healthcare professionals’ attitudes towards traumatic braininjury (TBI): the influence of profession, experience, aetiology and blame onprejudice towards survivors of brain injury. Brain Inj. 2010;24:802–11.
60. Sabin JA, Marini M, Nosek BA. Implicit and explicit anti-fat bias among alarge sample of medical doctors by BMI, race/ethnicity and gender. PLoSOne. 2012;7:e48448.
61. Li L, et al. Using case vignettes to measure HIV-related stigma amonghealth professionals in China. Int J Epidemiol. 2007;36:178–84.
62. Aaberg VA. A path to greater inclusivity through understanding implicitattitudes toward disability. J Nurs Educ. 2012;51:505–10.
63. Abuful A, Gidron Y, Henkin Y. Physicians’ attitudes toward preventivetherapy for coronary artery disease: is there a gender bias? Clin Cardiol.2005;28:389–93.
64. Chow LY, Kam WK, Leung CM. Attitudes of healthcare professionals towardspsychiatric patients in a general hospital in Hong Kong. J Psychiatry. 2007;17(1):3–9.
65. Neauport A, et al. Effects of a psychiatric label on medical residents’attitudes. Int J Soc Psychiatry. 2012;58:485–7.
66. Barnhart JM, Wassertheil-Smoller S. The effect of race/ethnicity, sex, andsocial circumstances on coronary revascularization preferences: a vignettecomparison. Cardiol Rev. 2006;14:215–22.
67. Sabin DJA, Nosek DBA, Greenwald DAG, Rivara DFP. Physicians’ implicit andexplicit attitudes about race by MD race, ethnicity, and gender. J HealthCare Poor Underserved. 2009;20:896.
68. Michael Vallis T, Currie B, Lawlor D, Ransom T. Healthcare professional biasagainst the obese: how do we know if we have a problem? Can J Diabetes.2007;31:365–70.
• We accept pre-submission inquiries
• Our selector tool helps you to find the most relevant journal
• We provide round the clock customer support
• Convenient online submission
• Thorough peer review
• Inclusion in PubMed and all major indexing services
• Maximum visibility for your research
Submit your manuscript atwww.biomedcentral.com/submit
Submit your next manuscript to BioMed Central and we will help you at every step:
Top Related