Discriminatory Healthcare Experiences and Medical Mistrust in Patients With Serious Illness

Abstract

Context.

Though discrimination in healthcare settings is increasingly recognized, the discriminatory experiences of patients with serious illness has not been well studied.

Objectives.

Describe racial differences in patient-reported experiences with discrimination in the healthcare setting and examine its association with mistrust.

Methods.

We used surveys containing patient-reported frequency of discrimination using the Discrimination in Medical Setting (DMS) and Microaggressions in Health Care Settings (MHCS) scales, mistrust using the Group Based Medical Mistrust (GBMM) scale, and patient characteristics including patient-reported race, income, wealth, insurance status, and educational attainment. Univariable and multivariable linear regression models as well as risk ratios were used to examine associations between patient characteristics including self-reported race, and DMS, MHCS, and GBMM scores.

Results.

In 174 participants with serious illness, racially minoritized patients were more likely to report experiencing discrimination and microaggressions. In adjusted analyses, DMS scores were associated with elements of class and not with race. Black, Native American/Alaskan Native (NA/AN), and multiracial participants had higher MHCS scores compared to White participants with similar levels of income and education. Higher income was associated with lower GBMM scores in participants with similar DMS or MHCS scores, but Black and NA/AN participants still reported higher levels of mistrust.

Conclusion.

In this cross-sectional study of patients with serious illness, discriminatory experiences were associated with worse mistrust in the medical system, particularly for Black and NA/AN participants. These findings suggest that race-conscious approaches are needed to address discrimination and mistrust in marginalized patients with serious illness and their families.

Introduction

Black and other racially minoritized patients experience inequities at end-of-life (EOL) resulting in disproportionately high healthcare costs, high symptom burden, and poor quality of life.^1,2,3 This is often attributed to patient- and family-level factors that are frequently targeted in interventions to address these inequities.^2,4–8 These approaches often utilize race-agnostic approaches that do not account for intersectional and racialized experiences of racially minoritized patients.⁹ Recently, we reported the results of a qualitative study of Black patients with serious illness.¹⁰ In that study, participants reported experiencing discrimination from healthcare workers (HCWs) that negatively impacted communication and medical decision-making. Participants reported that they felt these experiences were due to both race and socioeconomic disadvantage including being unhoused and underinsured.

Discriminatory behaviors by HCWs directed toward patients during clinical interactions are increasingly recognized as common experiences deserving of attention.^11–13 While participants in our study and others report experiencing discrimination due to race, discrimination based on other social identities has also been described.^13,14 For example, patient-reported discrimination has been attributed to socioeconomic disadvantage including low levels of educational attainment and income while others have reported unprofessionalism, disrespect, and other harmful behaviors from HCWs related to patients’ gender, sexual orientation, age, and other identities.^11,13–15

Discrimination in other public sectors, including the judicial and education systems, is well described and its link to poor health well documented.^16,17 In addition, discriminatory experiences in healthcare settings are associated with poor quality communication, patient-observed racial power imbalances, and patient- and family-perceived conflict.^18–22 Discriminatory experiences may exacerbate preexisting resentment in marginalized communities, worsen mistrust in the medical system, and contribute to poor expectations in care.^{11–14,23,24} These data suggest that discriminatory experiences related to race, class, and other identities may impact medical decision-making and decisions to seek care for patients or their family members,.^24–27 Communication and decision-making are crucial factors during serious illness, but the discriminatory experiences of patients with serious illness are not well described. It is important to better understand and characterize the discriminatory experiences of patients during serious illness so that they are better validated and addressed.

We report the results of a cross-sectional study to characterize patient-reported experiences of discrimination at an academic county hospital in a cohort of patients with serious illness, and to examine its associations with mistrust in the medical system. We hypothesized that patient-reported discrimination would be associated with socioeconomic status, but that there would still be important differences by patient self-reported race.

Methods

This study used the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting Guideline. We recruited patients with serious illness hospitalized at an academic county hospital in Seattle, Washington between January 2021 and August 2023 into a prospective cohort study to examine associations between racism experiences, communication, and healthcare intensity. Potential participants were screened through the electronic health record (EHR) and approached on hospital day two or three. Potential participants were at least 18 years old, without cognitive dysfunction, and with a serious illness defined as a diagnosis associated with a median life expectancy of two years or less.²⁸ After consenting, participants completed a survey containing demographic information including self-reported race, the Discrimination in Medical Setting (DMS) scale,¹⁸ the Microaggressions in Health Care Settings (MHCS) scales,^29,30 and the Group Based Medical Mistrust (GBMM) scales.³¹ Surveys were completed by participants themselves or verbally if the participant was unable to write or visually impaired. A $20 gift card was provided for enrollment into the cohort and completion or partial completion of the survey. This study was approved by the University of Washington Institutional Review Board (STUDY00011422).

We quantitatively analyzed racial differences in patient-reported discriminatory experiences in the healthcare setting using complete case analysis. Self-reported race was categorized according to the National Institutes of Health (NIH) guidelines as Black, White, Hispanic/Latinx, Asian, Native American/Alaskan Native (NA/AN), Native Hawaiian/Pacific Islander (NH/PI), Multiracial, and Other/Decline to Answer. To assess patient-reported frequency of discriminatory experiences in the healthcare setting, we used the DMS and MHCS scales. Discrimination refers to the differential treatment of members of different social groups and is usually the behavioral manifestation of prejudice toward those who are marginalized, both by race and social group.³² Microaggressions are the everyday, subtle interactions or behaviors that communicate bias toward marginalized groups, both unintentional and intentional, and are a form of discriminatory behavior.^33,34 The DMS is a nine-item scale with excellent internal consistency, test-retest reliability, and convergent and discriminant validity among Black and Latinx participants.^18,35,36 The survey asks respondents to report the frequency of specific types of discriminatory events on a Likert scale from 1 (never) to 5 (always) and is scored by the sum of responses with higher scores indicating more discriminatory experiences.³⁷ The MHCS is a 6-item survey designed to examine patient-reported experiences with perceived microaggressions from clinicians.²⁹ It is a unidimensional scale with good internal consistency, reliability and discriminant validity in Black and Latinx participants.^29,30 It asks respondents to report the frequency of microaggression events on a three-point Likert scale from 1 (this never happened) to 3 (this happened, and I was bothered by it) and is scored by the overall mean of responses with higher scores indicating more microaggression events. Health literacy was measured by the Rapid Estimate of Adult Literacy in Medicine – Short Form (REALM-SF) Health literacy instrument.³⁸ Participants are shown seven medical words to read aloud and are scored by the number of words correctly read. The Group Based Medical Mistrust (GBMM) scale has 12 items measuring ethnicity-based trust in clinicians.³¹ Participants rate their agreement to provided statements using a Likert scale from 1 (strongly disagree) to 5 (strongly agree). It is scored by the sum of responses. Higher scores indicate more mistrust. The GBMM highly correlates with health outcomes, yields similar results to more inclusive measures of wealth, and is easily obtained.^40,41 Wealth was measured by patient-reported ownership of five assets (checking account, savings account, retirement funds, vehicle, and home ownership), and scored from 0 to 5.³⁹ Using variance inflation factors, we evaluated for collinearity between the DMS, MHCS, GBMM scores and income, educational attainment, wealth, and REALM scores and found no collinearity between any of the variables. The Supplemental Material provides a complete description of all variables, including the data-collection method and operationalization of the variables.

Shapiro-Wilk tests were used to evaluate for normal distribution of the DMS, MCHS, and GBMM scores.⁴² To test for differences in DMS, MCHS, and GBMM scores across binary variables, we used two-sample t-tests for variables that were normally distributed and Wilcoxon rank sum tests for variables that were not normally distributed. We used χ² and Fisher exact tests to compare patient-reported race with categorical variables. To evaluate risk to participants from each racial group reporting forms of discrimination or mistrust on any item in the scales, we calculated risk ratios with 95% confidence intervals comparing likelihood a patient in one racial group reported: 1) a mean DMS value of 3 (“Sometimes”) or higher; 2) an MHCS score of 2 (“This happened but I wasn’t bothered by it) or higher; and 3) a mean GBMM value of 4 (“Disagree” for standardly score items, “Agree” for reverse scored items) or higher to others outside of that racial group. To compare means between categorical variables and race, and DMS, MCHS, and GBMM scales, we performed analyses of variance (ANOVA). These analyses were used to identify variables that were statistically significant or trending significant as potential confounders. Variables that were identified in prior qualitative interviews, literature review, and the DMS survey questions in which respondents identify why they felt discriminated against in healthcare settings were also included. We performed multivariable linear regression models, using combinations of these variables, to identify models that included race and variables with explanatory power. Additional combinations of variables that were not included in these models were also explored to ensure that no variables were excluded that could provide explanatory power. Finally, we performed multivariable linear regressions using DMS and MCHS scores as predictors and GBMM as an outcome, hypothesizing that more discriminatory experiences would result in higher levels of mistrust in the medical system. Interaction terms were generated between race and income, wealth, educational attainment, and health literacy, but no interactions were significant. We did not combine racial groups, even though the numbers of participants in some racial groups were small, as participants from different racial groups have different racialized experiences. One participant declined to report their race and was excluded from analyses. For our final regression models, we selected the most parsimonious model that included race and best correlation structure using the Akaike’s information criteria. We defined statistical significance as P < 0.05 in two-tailed tests. Analyses were conducted with Stata version 17 and Python version 3.11.1.

Results

Demographics

Table 1 summarizes the demographic characteristics of 175 survey respondents. Participants had a mean age (SD) of 59.5 (12.6) years and were mostly male (66.7%). A total of 41.1% of participants self-identified as Black, 23.6% as White, and 10.3% as multiracial individuals. Participants had a Modified Charlson Comorbidity Index equal to or greater than six (33.9%), New York Heart Association class III or IV congestive heart failure (33.3%), metastatic cancer or inoperable lung cancer (15.5%), or end stage renal disease with diabetes or an albumin level less than 2.5 g/dL (12.5%). A majority of participants were insured through either Medicaid (58.6%) or Medicare (20.7%). Participants on average had low levels of educational attainment, health literacy, income, and wealth. There were no racial differences in income, wealth, REALM scores, or educational attainment. A greater proportion of participants who did not identify as White had Medicaid insurance.

Table 1.

Demographics of Participants Who Completed Surveys

Variable	N	Statistic
Agea	175	59.2 (12.6)
Sexb	174
Male		116 (66.7)
Female		58 (33.3)
Self-reported raceb	175
Black		72 (41.1)
White		57 (32.6)
Multiracial		18 (10.3)
American Indian		11 (6.3)
Hispanic/Latinx		7 (4.0)
Native Hawaiian		6 (3.4)
Asian		3 (1.7)
Other		1 (0.6)
Diagnosesb
Charlson comorbidity index score ≥6		59 (33.9)
NYHA III or IV CHF		58 (33.3)
Metastatic cancer or inoperable lung cancer		27 (15.5)
ESRD with DM or albumin <2.5 g/dL		21 (12.5)
COPD FEV1<35% or O2 dependent		15 (8.6)
Child’s class C cirrhosis or MELD>17		5 (2.9)
PAH with 6MWT distance <250 feet		1 (0.6)
Aged 75 yrs or olderc		1 (0.6
Years of schoolinga	169	12.9 (2.1)
Religionb	174
Christian		102 (58.6)
None		38 (21.8)
Other		24 (13.8)
Muslim		5 (2.9)
Buddhist		4 (2.3)
Jewish		1 (0.6)
Household income/yearb	170
Less than $25,000		120 (70.6)
$25,000–$34,999		13 (7.7)
$35,000–$49,999		12 (7.1)
$50,000–$74,999		9 (5.3)
$75,000–$99,999		7 (4.1)
$100,000–$149,999		5 (2.9)
$150,000–$199,999		4 (2.4)
Wealthb	171
0 assets		53 (31.0)
1 asset		35 (20.5)
2 assets		34 (19.9)
3 assets		17 (9.9)
4 assets		14 (8.2)
5 assets		18 (10.3)
Insuranceb	174
Medicaid		102 (58.6)
Medicare		36 (20.7)
Private		25 (14.4)
Uninsured		6 (3.5)
Other		3 (1.7)
Military		2 (1.2)
REALM-SFa	123	5.9 (1.9)
Black	48	5.5 (2.2)
White	39	6.2 (1.6)
Multiracial	15	6.4 (1.6)
Native American/Alaskan Native	8	5.1 (2.4)
Hispanic/Latinx	5	7.0 (0.0)
Native Hawaiian/Pacific Islander	6	6.3 (1.2)
Asian	2	7.0 (0.0)

NYHA = New York Heart Association; CHF = congestive heart failure; ESRD = end stage renal disease; DM = diabetes mellitus; COPD = chronic obstructive pulmonary disease; FEV₁= forced expiratory volume in 1 second; O₂ = oxygen; MELD = Model for End-stage Liver Disease; PAH = pulmonary arterial hypertension; 6MWT = six-minute walk test; REALM = rapid estimate of adult literacy in medicine − Short Form.

^aMean (standard deviation).

^bNumber (percentage).

^cAged 75 years or older with diagnosis of at least one of the life-limiting chronic illnesses noted above, although possibly of lesser severity.

Discrimination in Medical Settings

Table 2 shows participant-reported frequency of certain discriminatory events. Approximately half of respondents reported sometimes, most of the time, or always being treated with less courtesy, less respect, and receiving poorer service than other participants. Approximately half of participants also reported that sometimes, most of the time, or always doctors or nurses act as if they do not think they are smart or act if they are better than them. Almost 60% of participants reported that a doctor or nurse did not listen to them sometimes, most of the time, or always. Educational attainment or income was chosen by 40.7% of participants as a reason why they were discriminated against (Table 3). Some participants reported that some aspect of their appearance (18.6%) or race (17.7%) explained why they experienced discrimination from HCWs. While completing the survey, participants identified additional factors or different interpretations of questions in the surveys, or identified gaps that the surveys did not address. For example, many participants pointed toward time constraints, often observing that HCWs are “busy and overwhelmed.” In addition, participants noted that they felt discriminated against because of their struggles with addiction or mental illness, an option not offered on the DMS survey. Multiracial (1.3, 95% CI 1.1, 1.5) and NA/AN (1.3, 95% CI 1.1, 1.6) participants were at increased risk of reporting experiencing discrimination at least “Sometimes” on any item on the DMS survey compared to other participants outside of their racial group (Table 4). Black participants were approximately two times more at risk (1.8, 95% CI 1.2, 2.8) compared to all participants outside of their racial group of having a DMS score of 3 or higher, reflecting an average response of “Sometimes” or higher on the DMS survey (Table 5). Additionally, participants who were not White (2.1, 95% CI 1.2, 3.9) were approximately two times more at risk of having a DMS scores of 3 or higher compared to White participants (Table 5).

Table 2.

Percentage of Participants Who Reported the Frequency of Experiencing Certain Discrimination Events as “Sometimes”, “Most of the Time”, or “Always” on Individual DMS Survey Items

Survey Item	Never	Rarely	Sometimes	Most of the time	Always
1. You are treated with less courtesy than other people. (N = 171)	40 (23.5)	44 (25.9)	59 (34.5)	20 (11.7)	7 (4.1)
2. You are treated with less respect than other people. (N = 172)	44 (25.7)	43 (25.2)	50 (29.1)	23 (13.4)	11 (6.4)
3. You receive poorer service than other people. (N = 171)	48 (28.2)	40 (23.5)	54 (31.6)	22 (12.9)	6 (3.5)
4. A doctor or nurse acts as if they think you are not smart. (N = 170)	53 (31.4)	32 (18.9)	52 (30.6)	20 (11.8)	12 (7.1)
5. A doctor or nurse acts as if they are afraid of you. (N = 171)	89 (52.4)	33 (19.4)	32 (18.7)	12 (7.0)	4 (2.3)
6. A doctor or nurse acts as if they are better than you. (N = 171)	58 (34.1)	35 (20.6)	42 (24.6)	19 (11.1)	16 (9.4)
7. You feel like a doctor or nurse is not listening to you (N = 172)	36 (21.1)	33 (19.3)	53 (30.8)	23 (13.4)	26 (15.1)

Table 3.

Reasons Participants Felt They Were Experiencing Discrimination in the Healthcare Setting

	Education or Income	Another Aspecta	Race	Age	Sexual Orientation	Weight	Religion	Ancestry	Gender	Religion
Black (N = 49)	15 (30.6)	7 (14.3)	14 (28.6)	7 (14.3)	2 (4.1)	2 (4.1)	1 (2.0)	1 (2.0)	0 (0.0)	1 (2.0)
White (N = 29)	13 (44.8)	5 (17.2)	0 (0.0)	6 (20.6)	2 (6.9)	1 (3.4)	1 (3.4)	0 (0.0)	1 (3.4)	1 (3.4)
Multiracial (N = 15)	10 (66.7)	4 (26.7)	0 (0.0)	1 (6.7)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
NA/AN (N = 19)	3 (15.8)	1 (5.3)	4 (21.1)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	1 (5.3)	0 (0.0)	0 (0.0)
Hispanic/Latinx (N = 5)	2 (40.0)	3 (60.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
NH/PI (N = 3)	2 (66.7)	1 (33.3)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Asian (N = 3)	1 (33.3)	0 (0.0)	2 (66.7)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
All (N = 113)	46 (40.7)	21 (18.6)	20 (17.7)	14 (12.4)	4 (3.5)	3 (2.7)	2 (1.8)	2 (1.8)	1 (0.9)	2 (1.8)

^aAnother aspect of my appearance. NA/AN: Native American/Alaskan Native; NH/PI: Native Hawaiian/Pacific Islander. Participants were able to choose more than one reason as to why they felt they were experiencing discrimination and not all participants indicated a reason for their experiences.

Table 4.

Risk Ratios of Reporting Discriminatory Experience(s) or Mistrust on Any Item on the DMS, MHCS, or GBMM Scales^a

	DMSb RR (95% CI)	MHCSc RR (95% CI)	GBMMd RR (95% CI)
Black (N = 67)	1.1 (0.9,1.3)	1.8 (1.4, 2.3)	1.2 (1.1, 1.4)
Multiracial (N = 18)	1.3 (1.1,1.5)	1.2 (0.9,1.8)	1.0 (0.7, 1.2)
Native American/Alaskan Native (N = 10)	1.3 (1.1,1.6)	1.2 (0.7,1.9)	1.3 (1.2, 1.4)
Hispanic/Latinx (N = 7)	1.2 (0.9,1.6)	1.3 (0.8, 2.1)	0.7 (0.4, 1.3)
Native Hawaiian/Pacific Islander (N = 6)	0.7 (0.3,1.5)	0.3 (0.1,1.8)	0.6 (0.3, 1.4)
Asian (N = 3)	0.5 (0.1, 2.3)	1.2 (0.5, 2.7)	0.8 (0.4, 1.8)
Non-white (N = 54)	1.2 (1.0,1.5)	2.4 (1.6, 3.8)	1.1 (0.9, 1.3)

DMS = discrimination in medical settings; MHCS = microaggressions in health care settings; GBMM = group based medical mistrust.

^aFor each racial group, risk ratios were calculated to compare the probability of reporting discrimination or mistrust on the scales compared to all other racial groups.

^bThe risk a participant in a particular racial group answered at least 3 or higher (“Sometimes”, “Often”, “Always”) to any question in the DMS survey.

^cThe risk a participant in a particular racial group answered 2 or higher (“This happened but it didn’t bother me” or “This happened and I was bothered by it”) to any question in the MHCS survey.

^dThe risk a participant in a particular racial group answered 4 or higher (”Disagree” and “Strongly Disagree”) to any question in the GBMM survey. This takes into consideration items in that are reverse scored.

Table 5.

Risk Ratios of Reporting Mean Scores Reflective of Frequent Discriminatory Experiences and Mistrust^a

	DMSb	MHCSc	GBMMd
Black (N = 71)	1.8 (1.2, 2.8)	2.5 (1.6, 3.9)	1.5 (1.1, 2.2)
Multiracial (N = 18)	1.5 (0.9, 2.7)	1.3 (0.7, 2.4)	1.5 (0.9, 2.2)
Native American/Alaskan Native (N = 11)	1.2 (0.5, 2.7)	1.5 (0.8, 3.0)	1.7 (1.2, 2.6)
Hispanic/Latinx (N = 7)	0.9 (0.3, 3.0)	0.9 (0.3, 3.30)	1.0 (0.4, 2.3)
Native Hawaiian/Pacific Islander (N = 6)	-	0.5 (0.1, 3.2)	-
Asian (N = 3)	-	1.1 (0.2, 5.3)	0.8 (0.2, 3.8)
Non-white (N = 56)	2.1 (1.2, 3.9)	6.2 (2.4, 15.9)	2.2 (1.3, 3.5)

DMS = discrimination in medical settings; MHCS = microaggressions in health care settings; GBMM = group based medical mistrust.

^aFor each racial group, risk ratios were calculated to compare the probability of reporting discrimination or mistrust on the scales compared to all other racial groups.

^bThe risk of a mean score of 3 or higher (“Sometimes”, “Often”, “Always”) in the DMS survey.

^cThe risk of a mean score of 2 or higher (“This happened but it didn’t bother me” or “This happened and I was bothered by it”) in the MHCS survey.

^dThe risk of a mean score of 3 or higher (“Neutral,” ”Disagree,” and “Strongly Disagree”) in the GBMM survey. This takes into consideration items in that are reverse scored. Risk ratios without values indicates that no participant within that racial group had scores that reached the pre-specified mean for that particular survey instrument.

Higher DMS scores were associated with higher GBMM scores (1.03, 95% CI 0.86, 1.2). Participants with high school (−4.1, 95% CI −7.2, −1.0), some college (−3.9, 95% CI −6.9, −0.8), and college level (−7.3, 95% CI −11.9, −2.8) education had lower DMS scores than those who did not complete high school. Higher levels of wealth were associated with lower DMS scores (−1.0, 95% CI −1.6, −0.40), specifically those who reported owning all five assets compared to those with none (−5.0, 95% CI −8.7, −1.4). In unadjusted analyses, Black (3.7, 95% CI 1.3, 6.0), NA/AN (4.9, 95% CI 0.5, 9.4), and multiracial participants (4.5, 95% CI 0.9, 8.0) had DMS scores that were higher than White participants. In multivariable analyses, race was not significantly associated with DMS scores (Fig. 1, panel a). However, higher educational attainment, specifically some college (−4.7, 95% CI −8.9, −0.5) and college level education (−9.1, 95% CI −13.9, −3.2), higher REALM scores (0.8, 95% CI 0.1, 1.4) and higher levels of wealth (−1.3, 95% CI −2.0, −0.5) were associated with lower DMS scores.

Fig. 1.

Adjusted DMS, MHCS, and GBMM scores by participant self-identified race.

Microaggressions in Health Care Settings

Table 6 shows the patient-reported frequency of certain microaggression events. Approximately a quarter of participants reported being bothered by HCWs being insensitive about their cultural group (26.9%), denying cultural biases (24.6%), having stereotypes about their cultural group (27.2%), or minimizing the importance of their cultural issues (20.9%). Higher levels of income (0.1, 95% CI −0.1, −0.01) and wealth (−0.07, 95% CI −0.12, 0.007) were associated with lower MCHS scores. Higher educational attainment, specifically high school level (−0.4, 95% CI −0.7, −0.1), some college (−0.3, 95% CI −0.6, −0.01) and college level (−0.75, 95% CI −1.2, −0.3) education, were associated with lower MHCS scores. Black (0.6, 95% CI 0.4, 0.9), Hispanic/Latinx (0.5, 95% CI 0.2, 0.9), AI/AN (0.5, 95% CI 0.1, 0.9), and multiracial (0.5, 95% CI 0.2, 0.9) participants had higher MCHS scores than White participants. Black (1.8, 95% CI 1.4, 2.3) and NA/AN (2.4, 95% CI 1.6, 3.8) were more at risk of reporting experiencing a microaggression compared to all participants outside of their racial groups. Black participants had over two times the risk of reporting mean MHCS of two or more compared to all other participants outside of their racial groups (2.5, 95% CI 1.6, 3.9). Participants who were not White had over six times the risk of having an MHCS score of 2 or higher indicating frequently experiencing microaggressions from HCWs that may or may not have bothered them (6.2, 95% CI 2.4, 15.9).

Table 6.

Participant-Reported Frequency of Experiencing Specific Microaggressions on Individual MHCS Survey Items N (%)

My Healthcare Provider:	This Never Happened	This happened, But it Didn’t Bother me	This Happened, and I Was Bothered by it
1: Avoided discussing or addressing cultural issues. (N = 171)	120 (70.2)	22 (12.9)	29 (17.0)
2: Sometimes was insensitive about my cultural group when trying to understand or treat my issues. (N = 171)	109 (63.7)	16 (9.4)	46 (26.9)
3: Seemed to deny having any cultural biases or stereotypes. (N = 171)	103 (60.2)	26 (15.2)	42 (24.6)
4: At times seemed to over-identify with my experiences related to my race or culture. (N = 171)	119 (69.6)	25 (14.6)	27 (15.8)
5: At times seem to have stereotypes about my cultural group, even if he or she did not express them directly. (N = 169)	99 (58.6)	24 (14.2)	46 (27.2)
6: Sometime minimized the importance of cultural issues. (N = 172)	111 (64.5)	25 (14.5)	36 (20.9)

Higher MHCS scores were associated with higher GBMM scores (9.2 95% CI 7.2, 11.2). In multivariate analyses, higher incomes (−0.1, 95% CI −0.1, −0.01) were associated with lower MCHS scores. Additionally, high school (−0.4, 95% CI −0.6, −0.1) and college level (−0.5, 95% CI −1.0, −0.1) education were associated with lower MCHS scores. However, Black (0.6, 95% CI 0.3, 0.8), Hispanic/Latinx (0.5, 95% CI 0.01, 1.0), NA/AN (0.4, 95% CI 0.04, 0.8), and multiracial (0.5, 95% CI 0.2, 0.8) participants had higher MCHS scores compared to White participants even after adjusting for educational attainment and income (Fig. 1, panel b).

Group Based Medical Mistrust

Black (1.2, 95% CI 1.1, 1.4) and NA/AN (1.3, 95% CI 1.2, 1.4) were more at risk of reporting mistrust on the GBMM by indicating that they disagreed with a statement in the survey. Black (1.5, 95% CI 1.1, 2.2) NA/AN (1.7, 95% CI 1.2, 2.6) participants, and participants who were not White (2.2, 95% CI 1.3, 3.5) were more at risk of having mean GBMM scores of 3 or higher, scores reflective of neutral or worse mistrust in the medical system. In unadjusted analyses, Black (8.3, 95% CI 4.9, 11.7), NA/AN (11.2, 95% CI 5.0, 17.4), and multiracial (7.9, 95% CI 2.6, 13.2) participants had higher mistrust scores compared to White participants. Higher levels of income (−1.8, 95% CI −2.8, −0.8,) and wealth (−1.8, 95% CI −2.7) were associated with lower levels of mistrust, in particular owning five assets of wealth compared to none (−10.3, 95% CI −15.8, −4.9). In adjusted analyses, higher levels of income were associated with lower levels of mistrust, but participants who self-identified as Black (4.3, 95% CI 1.7–7.0) and NA/AN (5.2, 95% CI 0.4, 9.9) still had higher mistrust scores compared to White participants (Fig. 1, panel c). While responding to items about equal treatment between groups, some White participants volunteered that they disagreed or strongly disagreed with this statement, citing articles in the lay press about disparate healthcare. However, other White participants disagreed or strongly disagreed with this statement, but cited personal, racist beliefs about racially minoritized participants that explicitly blamed them for unequal treatment.

The most parsimonious model for both DMS and MHCS scores on GBMM scores included only race and income (Fig. 2). Participants who self-identified as Black (4.3, 95% CI 1.7–7.0) and NA/AN (5.2, 95% CI 0.4, 9.9) had higher mistrust scores compared to White participants who had similar DMS scores and similar levels of income. NA/AN (6.5, 95% CI 1.1,11.9) participants had higher mistrust scores compared to other participants with similar levels of income and MHCS scores. Higher levels of income (−1.0, 95% CI −1.8, −0.13) were associated with lower levels of mistrust in participants from the same racial group and with similar DMS or MHCS scores.

Fig. 2.

Adjusted GBMM scores by participant self-identified race.

Discussion

In this study of patients with serious illness, experiences with discrimination in healthcare settings were not rare events. DMS scores were largely associated with elements of class; however, Black, NA/AN, and multiracial participants still experienced more microaggressions, a form of discrimination related to participants’ cultural beliefs and practices, compared to White participants with similar levels of income and education. While higher levels of income were associated with lower mistrust in participants with similar DMS or MHCS scores, Black and NA/AN participants consistently reported higher levels of mistrust compared to White participants with similar levels of educational attainment and income. The findings in this study add to the small but growing evidence that discriminatory events in healthcare are not rare, are associated with worsened mistrust in the medical system, and may have long-lasting negative impacts on marginalized patients.

To better understand the class-based inequities that contribute to the poor health outcomes of the United States, there is increasing interest in discussing and investigating White poverty and White privilege.^43,44 Our study demonstrates that both White and racially minoritized participants reported experiencing discrimination, and that educational attainment and income in particular had significant explanatory power in our models. In contrast to White participants, racially minoritized participants also reported more frequent discriminatory experiences due to both race and class. These data further reinforce the importance of incorporating intersectionality into analyses, and depict how being multiply marginalized exacerbates negative, discriminatory experiences in the healthcare setting. Knowing that prior discriminatory experiences are associated with both poor quality communication and patient-clinician interactions, it is important to consider the potential impact of discriminatory healthcare experiences on serious illness communication.^18–22 High quality serious illness communication engages patients in values-based discussions about their preferences and goals of care.⁴⁵ However, this and other data urge researchers and clinicians to take seriously the impact of prior and ongoing interpersonal discrimination and other forms of racism in healthcare settings on patients and their families, and how these experiences act as controlling factors that influence patients’ preferences and values.^10,45

It is critical for clinicians to validate and address discrimination and its impact on patients and families as best as they can without causing more harm.⁴⁶ Indeed, conversations around medical decision-making with trusted clinicians, especially ones who are race-concordant, may help provide additional support to patients as they navigate serious illness and near EOL.^47–49 In addition, clinicians and researchers must find ways to address defensiveness, engage in perspective-taking, and develop new, race-conscious models of communication.⁴⁶ Developing and cultivating these skills to promote high-quality patient-clinician communication that addresses and validates prior discrimination and other forms of racism and disrespect in the healthcare setting may help patients and families make medical decisions less encumbered by racist experiences, promote discussions around values, and provide care that is goal-concordant.^10,45,46

Although our study highlights the need to acknowledge the discrimination that patients from marginalized backgrounds experience in the healthcare setting, it has several limitations. First, this was a single center study. We employed recruitment techniques to oversample participants who were not White, but recruitment was stopped due to time limitations. Participant groups were small and while we could have combined racial groups to increase statistical power, we did not since participants from different racial groups have different racialized experiences. We did not perform subset analyses of participants who reported low levels of mistrust or discrimination nor did we gather baseline information about participants’ discriminatory experiences outside of the healthcare setting. Our participant population was low-income and we did not have a robust distribution of participants from varied socioeconomic classes. While we hypothesized that more discriminatory experiences would result in more mistrust, these data only point toward correlations and not causations. Lastly, varying interpretations of survey questions may have contributed to a lack of racial differences in DMS scores.

In conclusion, patients with serious illness reported experiencing discrimination in healthcare settings due in part to race and its intersection with other elements of class. These experiences were associated with worse mistrust in the medical system. These findings suggest that programs or tools that address patient experiences with interpersonal discrimination may help improve quality of serious illness communication for marginalized patients and their families while simultaneously addressing the larger structural and systemic inequities that persist today.

Key Message.

Racially minoritized patients and patients with socioeconomic disadvantage experience discrimination and mistrust in the medical system. While wider systemic and structural changes are necessary, race-conscious and intersectional approaches are needed to address and validate these negative and racist interpersonal experiences, especially during serious illness.