Can this be done by at 8 pm 4/22

American Economic Review 2019, 109(12): 4071–4111
https://doi.org/10.1257/aer.20181446

4071

Does Diversity Matter for Health?
Experimental Evidence from Oakland†

By Marcella Alsan, Owen Garrick, and Grant Graziani*

We study the effect of physician workforce diversity on the demand
for preventive care among African American men. In an experi-
ment in Oakland, California, we randomize black men to black or
non-black male medical doctors. We use a two-stage design, measur-
ing decisions before ( pre-consultation) and after ( post-consultation)
meeting their assigned doctor. Subjects select a similar number of
preventives in the pre-consultation stage, but are much more likely to
select every preventive service, particularly invasive services, once
meeting with a racially concordant doctor. Our findings suggest
black doctors could reduce the black-white male gap in cardiovascu-
lar mortality by 19 percent. (JEL I12, I14, C93)

African American men have the lowest life expectancy of any major demographic
group in the United States (Arias, Heron, and Xu 2017) and live on average 4.5 fewer
years than non-Hispanic white men (Murphy et al. 2017). Reasons for this disparity
are multifactorial and include lack of health insurance, lower socioeconomic status,
and structural racism (IOM 2003). Approximately 60 percent of the difference in
life expectancy between black and white men is attributable to chronic diseases,
which are amenable to primary or secondary prevention (Harper,  Rushani,  and

* Alsan: Harvard Kennedy School, 79 JFK Street, Cambridge, MA 02138 (email: [email protected].
edu); Garrick: Bridge Clinical Research, 333 Hegenberger Road, Suite 208, Oakland, CA 94621 (email: owen.
[email protected]); Graziani: University of California, Berkeley, Evans Hall, Berkeley, CA 94720 (email:
[email protected]). Esther Duflo was the coeditor for this article. We are grateful to an anonymous coed-
itor and four anonymous referees. We thank Pascaline Dupas and the J-PAL Board and Reviewers who provided
important feedback that improved the design and implementation of the experiment. We thank Ran Abramitzky,
Ned Augenblick, Jeremy Bulow, Kate Casey, Arun Chandrasekhar, Raj Chetty, Stefano DellaVigna, Mark Duggan,
Karen Eggleston, Erica Field, Matthew Gentzkow, Gopi Shah Goda, Susan Godlonton, Jessica Goldberg, Michael
Greenstone, Guido Imbens, Seema Jayachandran, Damon Jones, Supreet Kaur, Melanie Morten, Maria Polyakova,
Matthew Rabin, Al Roth, Kosali Simon, Ebonya Washington, Crystal Yang, and seminar participants at UC
Berkeley, Stanford, Cornell, MIT, UCLA, UCSB, Harvard Kennedy School, University of Chicago, and IFPRI for
their helpful comments. Javarcia Ivory, Matin Mirramezani, Edna Idna, Anlu Xing, and especially Morgan Foy pro-
vided excellent research assistance. We thank the study doctors and field staff team for their participation. We thank
the administration at Stanford, SIEPR, and J-PAL particularly Lesley Chang, Rhonda McClinton-Brown, Dr. Mark
Cullen, Dr. Douglas K. Owens, Ann Dohn, Ashima Goel, Atty. Ann James, Atty. Tina Dobleman, Nancy Lonhart,
Jason Bauman, Sophie Shank, James Turitto, Florian Grosset, and Luke Sonnet. A working paper version of this
paper was submitted for pre-publication re-analysis to the Abdul Latif Jameel Poverty Action Lab (J-PAL), where
a code replication exercise was conducted on the analysis. For more details about J-PAL’s replication work, visit
https://osf.io/be432/. We thank Uber for donating ride-sharing services, Alameda County for donating vaccinations,
and the Lenoirs for subletting their clinic. The study was made possible by a grant from the J-PAL—Health Care
Delivery Initiative with supplemental support from NBER P30AG012810. The trial is registered at clinicaltrials.gov
(NCT03481270) and in the AEA RCT Registry (0002497). The authors declare they have no conflicts of interest.

† Go to https://doi.org/10.1257/aer.20181446 to visit the article page for additional materials and author
disclosure statements.

https://doi.org/10.1257/aer.20181446

mailto:[email protected]

mailto:[email protected]

mailto:[email protected]

mailto:[email protected]

mailto:[email protected]

https://osf.io/be432/.

http://clinicaltrials.gov

https://doi.org/10.1257/aer.20181446

4072 THE AMERICAN ECONOMIC REVIEW DECEMBER 2019

Kaufman 2012; Silber et al. 2014). Some examples are poorly controlled hyperten-
sion (associated with stroke and myocardial infarction), diabetes (associated with
end organ disease including kidney failure), and delayed diagnosis of cancers. These
data suggest at least part of the mortality disparity is related to underutilized preven-
tive health care services.

One frequently discussed policy prescription put forth by the Institute of Medicine
(IOM) as well as the National Medical Association (NMA), the Association of
American Medical Colleges (AAMC), and the American Medical Association
(AMA) to address racial health disparities is to diversify the health care profession
by increasing the number of underrepresented minorities.1 Blacks comprise approx-
imately 13 percent of the US population but only 4 percent of physicians and less
than 7 percent of recent medical school graduates (AAMC 2014, 2016). Evidence
on whether patient and physician racial concordance improves satisfaction and
health outcomes is mixed, perhaps due to methodological or contextual differences.
Recent studies have found that gender- or race-match between doctors and patients
in a hospital setting reduces mortality (Greenwood, Carnahan, and Huang 2018;
Hill, Jones, and Woodworth 2018) yet in the outpatient setting, the results are less
clear. Meghani et al. (2009) performs a meta-analysis of 30 observational studies in
public health and medicine concerning four racial and ethnic groups. They conclude
that the evidence in favor of patient-doctor concordance in medical care is incon-
clusive and recommend additional research. We advance this literature by provid-
ing experimental evidence on whether and to what extent diversity in the physician
workforce improves medical decisions and outcomes among minority populations.

Our study builds upon several findings in economics. First, randomized trials in
development economics have demonstrated puzzlingly low demand for high return
preventive health care services among low-income populations (for a review, see
Dupas 2011 and Banerjee and Duflo 2011). Similar patterns are found in the United
States. Compared to non-Hispanic white men, African American men are six per-
centage points less likely to visit the doctor and eight percentage points less likely
to report receipt of the flu shot; insurance and education do not fully explain these
gaps (Blewett et al. 2018a).

Many factors likely contribute to this puzzle including lack of information, inad-
equate or low-quality health care supply, and misperceptions about the etiology of
disease. Given the prominent history of neglect and exploitation of disadvantaged
populations by health authorities, mistrust of the medical establishment is sometimes
invoked as a contributing factor. Evidence consistent with historical abuse damp-
ening demand and increasing mistrust has been found specifically among African
American men in the immediate aftermath of the US Public Health Service syphi-
lis experiment in Tuskegee, Alabama (Alsan and Wanamaker 2018) and persisting
decades after colonial medical campaigns in Central Africa (Lowes and Montero
2018). Recent studies in public health demonstrate that African American men

1 See Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care (IOM 2003); Addressing
Racial Disparities in Health Care: A Targeted Action Plan for Academic Medical Centers (AAMC 2009); “Major
Minority Physician Associations Come Together” (NMA 2018); and “Reducing Disparities in Health Care” (AMA
2018).

4073ALSAN ET AL.: DOES DIVERSITY MATTER FOR HEALTH?VOL. 109 NO. 12

continue to score higher on medical mistrust measures than other groups (Kinlock
et al. 2017, Nanna et al. 2018, Hammond et al. 2010).

Second, contributions in cultural economics have highlighted how norms of
behavior are influenced by social identity (Akerlof and Kranton 2000; Benjamin,
Choi, and Strickland 2010). Most notably, Tabellini (2008) shows how cooper-
ation can be sustained in a one-shot prisoner’s dilemma among agents who per-
ceive a non-economic benefit from cooperating with those closer in social distance.
Third, natural experiments in labor and education have underscored how diversity,
or lack thereof, may be particularly relevant in asymmetrical power relationships.
For instance, Glover, Pallais, and Pariente (2017) finds that minority workers exert
less on-the-job effort in grocery stores with biased majority managers. Additional
evidence on how diversity affects hiring and job performance can be found in Stoll,
Raphael, and Holzer (2004); Giuliano, Levine, and Leonard (2009); Hjort (2014);
and Bertrand et al. (2019). A spate of studies has found that same race or same
gender teachers are positively correlated with grades and career path, potentially
through a role model effect (Ehrenberg, Goldhaber, and Brewer 1995; Dee 2004,
2005; Bettinger and Long 2005; Carrell, Page, and West 2010; Fairlie, Hoffmann,
and Oreopoulos 2014; and Lusher, Campbell, and Carrell 2018).

There are several ways in which racial diversity could play a role in medicine,
specifically as it relates to the patient-doctor relationship. Taste-based discrimina-
tion (Becker 1957) on the part of the patient or doctor could imply that individuals
are averse to interacting with those who do not share their racial background. On the
other hand, internalized racism, or negative beliefs about one’s racial group, could
lead to the opposite phenomenon. Third, a common racial background might facili-
tate communication, a critical component of clinical care as both patient and physi-
cian have potentially life-saving information to exchange. Fourth, and not mutually
exclusive, concordance may foster trust leading to cooperation (i.e., compliance
with doctors’ advice or willingness to engage). As noted by Arrow (1963, p. 951),
“… it is a commonplace that the physician-patient relation affects the quality of the
medical care product.”

In this study, we examine whether doctor race affects the demand for preventive
care among African American men. We induce exogenous variation by randomly
assigning subjects to black and non-black doctors.2 Our experiment was conducted
in Oakland, California, where we recruited over 1,300 black men from about 20
local barbershops and 2 flea markets. At these recruitment sites, subjects filled out
baseline questionnaires and received a coupon for a free health screening. To facili-
tate our experiment, we set up a clinic to provide preventive services to the subjects.
The clinic was staffed with 14 black and non-black male doctors from the Bay Area
as well as a diverse team of receptionists. Doctors and staff were told the study was
designed to improve the take-up of preventive care among black men in Oakland,
but not specifically informed about the role of doctor race. Subjects learned of their
(randomly) assigned doctor via tablet in the privacy of their own patient room.

The experiment proceeded in two stages and cross-randomized doctor race
with incentives for the flu vaccine at the individual level. In the pre-consultation

2 Throughout the paper, we use “black” to refer to African Americans and “ non-black” to refer to Caucasian
and Asian Americans.

4074 THE AMERICAN ECONOMIC REVIEW DECEMBER 2019

stage,  patients were introduced to their doctor via the tablet by way of text and
photo, both standardized as described in Section I. Subjects were then provided the
opportunity to select which, if any, of the four advertised cardiovascular screen-
ing services they would like to receive. These services included body mass index
(BMI) measurement, blood pressure measurement, diabetes screening, and choles-
terol screening. The last two tests required a blood sample, and subjects were made
aware of this feature. After making their selections for cardiovascular screening,
subjects were informed they could also elect to receive a flu shot, administered by
their assigned doctor. For subjects randomized to receive a flu incentive to encour-
age vaccine selection, the incentive amount was also listed. We conjectured that if
subjects disliked doctors who did not share their racial background, those randomly
assigned to non-black doctors would, on average, demand fewer preventives simply
based on the tablet photo.

In the second stage, subjects met their assigned doctor in person. We refer to this
stage throughout the paper as post-consultation (since decisions occur after inter-
acting with their doctor). Subjects could revise their choice of preventives during
this stage, after which the doctor administered the selected services. We therefore
measure how black versus non-black doctors affect demand between the pre- and
post-consultation stages, which we refer to as the delta, since it represents the change
in selected services across the two periods. These are two choice events occurring
after randomization and both represent experimental outcomes. Following the
patient-doctor interaction, subjects filled out feedback forms and exited the clinic.

It is important to note that the study provided only preventive (i.e., care rec-
ommended during a state of relatively good health to avoid future illness, such as
screenings and immunizations) as opposed to curative (i.e., care needed during a
state of illness to restore health) interventions. Individuals often have imperfect
knowledge regarding the health benefits of prevention, perhaps because they have
been misinformed, never informed, or informed by someone they don’t trust, which
can dampen demand.3 Hence, the role of study doctors was limited to information
provision on the benefits of receiving care even when not feeling sick and then pro-
viding those chosen.

Approximately one-half of the subjects we recruited from the community vis-
ited our clinic. Those who presented were negatively selected relative to those
who completed the barbershop survey but did not come to the clinic. Subjects who
redeemed the clinic coupon were 13 percentage points more likely to be unem-
ployed (compared to 18 percent among non-participants) and 19 percentage points
more likely to have a high school education or less (compared to 44 percent among
non-participants). In terms of health and health care utilization, they had signifi-
cantly lower self-reported health, were less likely to have a primary care physician,
and more likely to have visited the emergency room.

Once at the clinic, subjects randomly assigned to a black doctor elected to receive
the same number of preventive services as those assigned to a non-black doctor in
the pre-consultation stage. In sharp contrast, we find that subjects assigned to black
doctors, upon interacting with their doctor, are 18 percentage points more likely to

3 According to the CDC, up to 40 percent of annual deaths in the United States are deemed preventable (CDC
2014).

4075ALSAN ET AL.: DOES DIVERSITY MATTER FOR HEALTH?VOL. 109 NO. 12

take up preventives relative to those assigned to non-black doctors. These findings
are robust to corrections for correlated error structures within doctor; the inclusion
of fixed effects for clinic date, field staff, and recruitment location; as well as various
permutations of the study doctors, including dropping the “best” black and “worst”
non-black doctor.

Why would black male subjects randomly assigned to black male doctors elect to
receive more services upon interacting with them? We provide several pieces of evi-
dence that better communication between black subjects and black doctors explains
our results, and discuss alternative mechanisms below. First, in our controlled study
environment, the role of the doctor was circumscribed to informing subjects about
the benefits of preventive services, and then providing those chosen. Second, for
non-invasive tests (those that do not require blood or an injection), both non-black
and black doctors shifted out demand in the post-consultation stage relative to the
pre-consultation stage, though the effect was larger for the latter. Yet, for invasive
tests, those that carry more risk and thus likely require more trust in the person
providing the service, only subjects assigned to black doctors responded: increasing
their take-up of diabetes and cholesterol screenings by 20 and 26 percentage points
(49 percent and 71 percent), respectively. Third, subjects are more likely to talk to
their assigned doctor about health issues if the assigned doctor is African American,
a result which is particularly strong among those who obtain an invasive exam.

The experimental findings highlighting improved communication for black
male patients paired with black male doctors are consistent with those collected in
a non-experimental manner. We surveyed 1,490 black and white adult males who
matched our sample in terms of educational attainment. The respondents were asked
to select a doctor of a particular race based on accessibility, quality, and communi-
cation. With respect to quality (i.e., which doctor is the most qualified) black and
white respondents both selected doctors of the same race about 50 percent of the
time. However, for questions regarding communication, in particular which doctor
would understand your concerns, the proportion of respondents choosing doctors of
their own racial background jumped to nearly 65 percent for blacks and 70 percent
for whites.

An alternative interpretation of our results is that the estimated treatment effect
is picking up an attribute correlated with the race of the doctor in our sample and
which affects the outcome of interest. A prominent candidate for a hard-to-measure
characteristic that may correlate with doctor race is quality.4 The non-experimental
findings cited above demonstrate black male respondents believe that non-black
doctors are as qualified as black doctors. Yet, actual doctor quality within the con-
text of our study could vary.

We address the possibility of differential quality across doctor race in the study
setting in several ways. First, doctors were balanced on observables in age, expe-
rience, and medical school rank, characteristics we collected from their resumes.
Moreover, all of the non-black doctors, but only 67 percent of black doctors, prac-
ticed internal medicine. In addition, we created a survey for the study doctors

4 This could arise if, for example, black doctors are more qualified than non-black doctors in the population and
we failed to draw our sample from an area of overlapping support, or if the distributions were similar, but we drew
from different tails.

4076 THE AMERICAN ECONOMIC REVIEW DECEMBER 2019

designed to assess their typical patient characteristics, their persuasiveness, and
their current medical knowledge using questions typically found on medical creden-
tialing exams. Interestingly, the non-black study doctors were more likely to state
their patients comply with medical advice and that they are able to persuade both
white and black adult male patients to take up testing they had initially refused.

If black doctors were higher quality than non-black doctors we might have
expected them to be rated higher on feedback forms, yet black and non-black
doctors were rated equally (highly). This compression likely reflects the design.
Differences in quality that would stem from diagnostic or treatment skills were not
elicited in our study, which narrowly focused on encouraging the take-up of preven-
tives. Furthermore, if black doctors were higher quality, they should perform better
with all patients and on all tests. Although our recruitment efforts were focused on
African American men, 12 clients identified as from another racial or ethnic back-
ground.5 Among this out-of-sample group, individuals were 20 percentage points
less likely to choose invasive services in the post-consultation stage when random-
ized to black doctors (a finding that is more extreme than 97 percent of bootstrap
coefficients on draws of 12 in-sample subjects). Moreover, for the in-sample sub-
jects, the differences in post-consultation preventive test take-up were much more
muted for non-invasive screenings (e.g., blood pressure) than for exams that required
blood (e.g., cholesterol). Thus, in order for an attribute correlated with the race of
black doctors to be driving our results, it must manifest only when treating African
American male patients and especially for invasive exams.

This leads to another competing explanation: perhaps black male doctors exerted
more effort with patients who shared their racial background. Since communication
requires some amount of effort, this is not an interpretation to which we object
(though we note if communication is more natural within concordant pairs, black
doctors might be expending less effort to achieve the same or better results: i.e.,
communication may be more efficient). Time spent with patients has been used as a
proxy for provider effort (Das et al. 2016). Equating time spent with effort is prob-
lematic in our setting because it reflects many different factors. A longer time spent
could simply reflect the treatment effect (i.e., subjects elect to receive more services
from black doctors), low quality (i.e., difficulty performing the services), or com-
munication (i.e., a better patient-doctor connection facilitating credible information
exchange). We find that black doctors indeed spent more time with subjects, but
this finding is driven by the treatment effect: the difference in visit lengths is small
and statistically insignificant after adjusting for the selected services. If we examine
another potential proxy for effort, the allocation of services to the “highest need”
subjects, we fail to find evidence that doctors of either race were expending effort
to target interventions. Lack of targeting also reflects our instruction to the study
doctors to try and encourage all patients to take up preventives.

Although years of experience in the medical field do not differ by race of doctor,
it is possible that black male doctors have more familiarity with serving black male
patients. This sorting would be consistent with national statistics on doctor-patient
pairings as well as with the tendency for minority physicians to work in medically

5 To avoid conflict, we provided services for the handful of people from other backgrounds who were consented
in to the study but deleted them from the main analytical sample.

4077ALSAN ET AL.: DOES DIVERSITY MATTER FOR HEALTH?VOL. 109 NO. 12

underserved areas with more low-income and minority patients (Komaromy et al.
1996; Walker, Moreno, and Grumbach 2012). Our study doctor survey reveals that
black doctors were more likely to have seen at least five black adult male patients a
week, though this experience does not predict doctor fixed effects. In addition, in the
context of our own experiment, non-black physicians did not “close the gap” with
black doctors vis-à-vis post-consultation preventive care take-up over time.

Lastly, we do not find evidence for the controversial hypothesis that subjects or
doctors were discriminating against each other. First, there was no race-preference
elicited in the pre-consultation (tablet) stage. Second, the comments and ratings
on feedback forms were consistently positive for both sets of doctors. As for
provider-level discrimination, all doctors who were involved in the study knew the
goal was to improve the preventive care of black men (though were blind to the
notion that their race was being randomized, thus we could not administer implicit
association tests). Taste-based discrimination by doctors would again be inconsis-
tent with non-black doctors being rated as highly as black doctors. We also failed
to find evidence that doctors of different races were using distinct thresholds to test
patients for disease, consistent with Chandra and Staiger (2010).

Racial concordance between subjects and doctors appears to be a particular com-
ponent of social distance that is influential in affecting demand. Alternative concor-
dance measures, such as whether subjects and assigned doctors share approximately
the same age or educational attainment, do not predict health care demand in any
meaningful way. Nor does race interact with these other concordance measures.
Such findings should be interpreted with caution since these characteristics were
not randomized.

Similar to prior scholarship on incentives for preventives among low-income
communities (Banerjee et al. 2010, Cohen and Dupas 2010, Thornton 2008), we
find that financial incentives for the flu shot increased demand for the vaccine: by
19 percentage points for a $5 incentive and 30 percentage points for a $10 incen-
tive in the pre-consultation stage. Yet, not all those who selected an incentivized
flu shot actually received it. About 18 percent of subjects randomized to black
doctors and 26 percent randomized to non-black doctors declined the shot in the
post-consultation stage (many cited contraindications). And regardless of incentive
level, black doctors increased demand in the post-consultation stage, convincing
about 26 percent of subjects who initially turned down an incentive and refused a
flu shot to obtain it, suggesting subsidies and (meeting with) black doctors are not
perfect substitutes.

In the setting of imperfect information regarding the benefit of health care, demand
curves cease to be a sufficient statistic for welfare calculations (Pauly and Blavin
2008; Baicker, Mullainathan, and Schwartzstein 2015; Chetty, Looney, and Kroft
2009). Furthermore, we incentivized take-up for only one preventive yet demand
for every preventive was affected by a black doctor treatment. Thus, to make prog-
ress on valuation, we combine published estimates on the health value of interven-
tions offered in our clinic with results from our study. The health value estimates
come from cost-effectiveness simulations in which the screen-positive population
obtains and complies with guideline-recommended therapy. Using this approach,
we calculate that black doctors would reduce mortality from cardiovascular disease
by 16 deaths per 100,000 per year, accounting for 19 percent of the black-white

4078 THE AMERICAN ECONOMIC REVIEW DECEMBER 2019

gap in cardiovascular mortality (Kahn et al. 2010; Dehmer et al. 2017; Murphy et
al. 2017; and Harper, Rushani, and Kaufman 2012). If these effects extrapolate to
other leading causes of death amenable to primary or secondary prevention, such as
HIV/AIDS or cancer, the gains would be even larger.

These calculations presume that there is a supply of African American male
doctors who could screen and treat black male patients. This might not be a safe
assumption. Black males are especially underrepresented in the physician work-
force, comprising about 12 percent of the US male population but only 3 percent
of male doctors (AAMC 2014, Census Bureau 2013). According to a recent report
by the AAMC (2015), the number of black male medical students has been roughly
constant since 1978 (when 542 matriculated into medical school compared to 515
in 2014). Returning to the non-experimental results, black male respondents were
26 percentage points less likely than white respondents to state that a doctor who
matched their race and sex was available to them.

The remainder of the paper proceeds as follows. Section I describes the exper-
imental design and the hypotheses tested. Section II describes the data, empirical
approach, and the characteristics of study subjects. Section III presents the main
findings and Section IV explores potential mechanisms and validity concerns.
Section V discusses health benefits and Section VI concludes.

I. Experimental Design and Hypotheses Tested

A. Design

The experiment was conducted in Oakland, California, in the fall and winter
of 2017–2018 (see Figure 1 for study design and flow). We recruited men from
19 black barbershops as well as 2 flea markets in and around the East Bay (about
88 percent of all recruited were at barbershops). Field officers (FO) approached men
in the barbershops to enroll in the study. After obtaining written informed consent,
the subject was given a short baseline survey.6 The baseline survey included ques-
tions on socio-demographics, health care, and mistrust. For completing the survey,
the men received a coupon (worth up to $25) for their haircut or, at the flea market,
a cash incentive. After completing the baseline survey, the subjects were given a
coupon to receive a free health screening for blood pressure, BMI, cholesterol, and
diabetes at the clinic we operated on 11 Saturdays (see online Appendix Figure 1).
Subjects were encouraged to come to the clinic promptly, and subjects who did not
have transport could receive a ride to the clinic courtesy of Uber. Field officers used
their own smart phones to obtain the rides. Attendance at …