Medical mistrust and willingness to use mental health services among Black adults

Abstract

Objective:

Black adults experience greater severity of depression compared to their White counterparts yet are also less likely to receive treatment from a mental health professional. This study aimed to examine the relationship between mistrust and willingness to seek mental health care.

Methods:

The authors conducted an online cross-sectional survey of 1043 Black adults in the United States. The primary variables of interest were mistrust (negative attributions about physicians and healthcare workers, measured with the 12-item Group Based Medical Mistrust Scale; GBMMS) and a single item derived from the General Help-Seeking Questionnaire (GHSQ) assessing willingness to seek mental health care. They hypothesized mistrust would be negatively correlated with willingness to seek help from a mental health professional (and trust would be positively correlated). To estimate the association between mistrust and willingness to seek care, they fitted gamma regression models with a log link, adjusting for age, ethnic identity/origin, education, insurance status, personal income, citizenship status, and length of time in the US.

Results:

A statistically significant non-linear relationship was observed, with a positive correlation between willingness to seek help and mistrust at lower levels of mistrust or GBMMS scores ≤3 (RR=1.55, p<0.001; 95%CI=1.44, 1.67) and a negative correlation at higher levels of mistrust or GBMMS scores >3 (RR=0.74, p<0.001; 95%CI=0.66, 0.83). This inverted “U” relationship was observed for both mistrust and trust.

Conclusion:

Lower levels of mistrust may have a counterintuitive positive correlation with mental health service use that does not undermine willingness to engage in help-seeking behavior.

Introduction

Black adults in the United States experience higher chronicity and severity of depression compared to their White counterparts.^1–3 Black adults are less likely to receive treatment from a clinician or mental health professional⁴ and are less likely to attend outpatient mental health visits or use psychiatric medication.⁵ Several factors contribute to these patterns of mental health service utilization, including financial limitations (e.g., access to adequate health insurance⁶), clinician cultural competency in accurately diagnosing mental health disorders among non-White patients,⁷ and the absence of mental health clinics or hospital systems embedded within underserved communities.^8,9 Even when controlling for socioeconomic status and insurance, Black adults are still less likely to seek care.¹⁰ Additionally, it is noted that the COVID-19 pandemic may have widened disparities in treatment utilization.¹¹

Other potential explanatory structural factors have received less attention in this literature. Previous studies show that Black adults are hesitant to seek professional help for mental health problems.^12–14 This hesitation is due to in part to concerns associated with racial discrimination within healthcare settings, which contribute to medical mistrust.^15,16 Medical mistrust, which is rooted in both historical and ongoing discrimination by clinicians and healthcare organizations towards underserved persons, is a phenomenon that stems from systems of oppression and unfair power structures.¹⁷ Although it is often conceptualized as a lack of interpersonal trust driven by individual-level psychological variables, the entity of medical mistrust has been more rigorously defined as the general sense of suspicion and belief that an institution or individuals will not meet agreed upon expectations to provide optimal care. A more robust conceptual framework presented by Richmond and colleagues defines mistrust as the belief that the medical system associated with a society’s dominant culture will act in a manner that will “not optimize well-being or worse, actively harm a patient.”^18,19 This concept is distinct from the lack of or absence of trust, as mistrust is characterized by the belief that the healthcare system or its personnel is acting against an individual’s well-being.^{20, 21} Whereas, medical trust conceptual frameworks focus on positive attributions of honesty, communication, fairness, and competency towards medical personnel and systems. In this study, we use a widely accepted trust and mistrust assessment, the Group Based Medical Mistrust Scale (GBMMS) and we approach trust and mistrust as distinct measures, consistent with existing literature that address trust and mistrust as distinct, rather than opposites.¹⁸ For example, a healthcare consumer may have a high degree of trust in their personal family doctor but harbor significant mistrust for medical institutions as a whole. Mistrust can also be driven by spillovers from larger systemic factors outside of the healthcare system (e.g., racial disparities in the incidence of police violence).^22,23

Medical mistrust is associated with decreased healthcare utilization.^{24 25} This phenomenon has been documented for a variety of health contexts such as delayed cancer screening and vaccination,^26–29 decreased adherence to HIV treatment or prophylaxis, ^30–34 and care seeking for Ebola virus disease.^35,36 Medical mistrust stemming from the Tuskegee Study of Untreated Syphilis in the Negro Male was shown to account for up to one-third of the life expectancy gap between Black and White adults.³⁷ Thus, the phenomenon of medical mistrust is significant from a public health perspective.

Emerging research on the impact of medical mistrust on mental health outcomes has demonstrated that higher levels of mistrust are associated with delayed treatment seeking and decreased willingness to utilize psychiatric medications.^19,38 Early recognition and treatment onset, and shortened duration of untreated illness, are associated with improved mental health outcomes across a wide range of psychiatric disorders.^39–41 Therefore, it is imperative to further our understanding of the impact of medical mistrust on help-seeking behaviors. To address this gap in the literature, we conducted this observational study of Black adults to better understand the relationships between medical trust, mistrust, and willingness to seek help for mental health problems from a mental health professional. We also assessed the extent to which demographic factors would moderate the association between medical mistrust/trust and willingness to seek help from a mental health professional. We had two initial primary hypotheses: Mistrust is negatively correlated with willingness to seek help from a mental health professional; and Trust is positively correlated with willingness to seek help from a mental health professional.

Methods

Study Design

We conducted an online cross-sectional survey of Black adults residing in the United States between September 2020 and July 2021. We specifically surveyed people who identified as African American, African, and Afro-Caribbean. We partnered with local community-based organizations (the United African Organization, World Relief Chicago, Coalition for Immigrant Mental Health) to disseminate the survey to their networks. We also recruited research participants through social media (Twitter and Facebook). The final sample size was therefore a convenience sample (N=1043). Current research uses variable terms for race/ethnicity for Black people.⁴² In our study, participation was limited to individuals who identified as Black (n = 1043) and: 1) also identified as African American (n = 805), African (n = 187), Afro-Caribbean (n = 44), or other (n = 7); 2) were 18–65 years of age; 3) resided in the United States at the time of the study; and 4) were English speaking. In our survey, ethnic identity/origin referred to people who shared a background with one of four options: African, African American, Afro-Caribbean and other. Written informed consent was obtained for each participant prior to engagement in study activities. Each participant received $25 compensation for their time. Northwestern University (ID STU00213136) provided IRB approval for study activities.

Assessments

The survey instrument elicited socio-demographic factors including age, ethnic identity/origin, education, insurance status, personal income, citizenship status, and length of time in the U.S. The full survey was part of a larger study. The average length to complete each scale was about 5–7 minutes. The two primary variables of interest in this analysis were the 12-item Group Based Medical Mistrust Scale (GBMMS)⁴³ and a single item derived from the General Help-Seeking Questionnaire (GHSQ).⁴⁴

The GBMMS was designed to assess the tendency to mistrust people or systems outside of one’s ethnic group. Each item is scored on a 5-point Likert scale ranging from “strongly disagree” to “strongly agree”. A principal component analysis of the 12 items revealed two factors that were analyzed: 1) mistrust, and 2) trust. Factor 1 (mistrust) items were focused on negative attributions about doctor and healthcare workers including an item that endorses actual negative experience perpetuated by doctors. One sample item was “People of my ethnic group cannot trust doctors and healthcare workers.” Factor 2 (trust) items were not structured as a direct opposite to the Factor 1 (mistrust) items. These items focused on receiving the same type of care across ethnicities. One sample item was “People of my ethnic group are treated the same as people of other groups by doctors.” Internal consistency in the present sample was 0.62 for the total GBMMS, 0.83 for the mistrust subscale, and 0.65 for the trust subscale. This was consistent with previous psychometric analyses of the GBMMS among Black adults, with internal consistency for the total GBMMS and its subscales ranging from 0.65 – 0.89.^45,46 GBMMS also showed strong construct validity.⁴⁵

The GHSQ was designed to assess intentions to seek help from 10 different sources (e.g., partner, friend, helpline, or mental health professional) for two different mental health concerns: personal/emotional problems and suicidal ideation. Each item is scored on a 7-point Likert-type scale ranging from “extremely unlikely” to “extremely likely”. For this study, we used one of the 10 items that assessed intentions to seek help (for personal/emotional problems and for suicidal ideation) specifically from a mental health professional such as a counselor, social worker, or psychologist. Previous psychometric analysis for the full GHSQ scale have shown strong construct validity for the specific item used in our analysis.^44,47

Analysis

We initially examined the relationship between mistrust and willingness to seek help from a mental health professional by calculating mean willingness scores across different values of the total GBMMS. We observed a non-linear relationship, with a positive correlation between willingness and mistrust for GBMMS scores ≤3 and a negative correlation between willingness and mistrust for GBMMS scores >3. This inverted “U” relationship was observed for the total GBMMS and both subscales. We therefore decided to stratify the sample into subgroups with GBMMS scores ≤3 vs. >3.

To estimate the association between mistrust and willingness to seek care, we fitted gamma regression models with a log link. In these models, GBMMS was specified as the explanatory variable of interest, with the mistrust and trust subscales included in separate models. The dependent variables were specified (in separate regression models) as willingness to seek care from a mental health professional for (1) a personal or emotional problem and (2) a suicide crisis. We fitted separate models for those with a GBMMS ≤3 and GBMMS >3. The rate ratios derived from these models represented the percent change in the average willingness to seek care per 1 unit change in the mistrust score. We further adjusted for demographic factors including age, ethnic identity/origin, education, insurance status, personal income, citizenship status, and length of time in the US.

Results

Tables

Table 1 shows the descriptive demographics and sample characteristics of the participants.

Table 1.

Descriptive demographics and sample characteristics (N=1,043)

	n	%
Age
18-44	749	72
45-65	294	28
Sex
Male	677	65
Female	365	35
Marital status
Married	665	64
Unmarried, living with a romantic partner	162	16
Never married	139	13
Separated	24	2
Divorced	37	3
Widowed	16	2
Education
Less than high school	29	3
High school diploma / GED	147	14
Trade school/vocational school	167	16
Some college, no degree	280	27
2-year college	143	14
4-year college degree	217	21
Master’s degree	43	4
Doctoral degree	17	2
Annual income
$9,999 or less	86	8
$10,000 to $29,999	250	24
$30,000 to $49,999	349	34
$50,000 to 89,999	204	19
$90,000 or more	154	15
Insurance
None	209	20
Public	620	59
Private	209	20
CHAMPUS, CHAMPVA, VA, or Other military healthcare	<10	1
Ethnic identity/origin
African	187	18
African-American	805	77
Afro-Caribbean	44	4
Other*	<10	1
Length of time in US
Less than 2 years	17	2
2 to 5 years	142	14
6 to 10 years	222	21
More than 10 years	662	64
Citizenship status
Citizen	864	83
Non-citizen	179	17

^*“Other” includes people who identified as Black but who did not select any of the three ethnic identities/origin

Primary Analysis

After adjusting for age, ethnic identity/origin, education, insurance status, personal income, citizenship status, and length of time in the US, mistrust had a nonlinear association with willingness to seek help from a mental health professional for personal and emotional problems (Table 2). At low levels of mistrust (GBMMS mistrust subscale ≤3), each unit increase in mistrust was associated with an increase in the probability of seeking care from a mental health professional (RR=1.55; 95%CI=1.44, 1.67; p<0.001). At high levels of mistrust (GBMMS mistrust subscale >3), each unit increase in mistrust was associated with a decrease in the probability of seeking care from a mental health professional (RR=0.74; 95%CI=0.66, 0.83; p<0.001). Similar patterns were observed in the association between mistrust and willingness to seek help from a mental health professional for suicidal problems.

Table 2.

Group Based Medical Mistrust Scale (mistrust and trust subscales) and willingness to seek help from a mental health professional

	For personal or emotional problems				For suicidal ideation
	Rate ratio	Lower	Upper	P-value	Rate ratio	Lower	Upper	P-value
Mistrust subscale	Adjusted *
<3, n=586	1.55	1.44	1.67	<0.0001	1.44	1.33	1.56	<0.0001
>3, n=457	0.74	0.66	0.83	<0.0001	0.74	0.66	0.84	<0.0001
Trust subscale	Adjusted *
<3, n=627	1.29	1.19	1.40	<0.0001	1.25	1.15	1.36	<0.0001
>3, n=416	0.64	0.57	0.72	<0.0001	0.66	0.59	0.74	<0.0001

^*Adjusts for age group, ethnicity, education, insurance status, personal income, citizenship status, and length of time in US

The association between trust and willingness to seek help showed a similar pattern to mistrust. At high levels of trust (GBMMS trust subscale >3), each unit increase in trust was associated with a decrease in the probability of seeking care from a mental health professional (RR=0.64; 95%CI=0.57, 0.72; p<0.001). A low levels of trust (GBMMS trust subscale ≤3), each unit increase in the trust score was associated with an increase in the probability of seeking care from a mental health professional (RR=1.29; 95%CI=1.19, 1.40; p<0.001). Similar patterns were observed in the association between trust and willingness to seek help from a mental health professional for suicidal problems.

As shown in Figures 1 and 2, both trust and mistrust had an inverted U-shaped association with willingness to seek care: at lower levels of mistrust (or trust), mistrust (or trust) was associated with an increased willingness to seek care, whereas at higher levels of mistrust (or trust), mistrust (or trust) was associated with a decreased willingness to seek care. The patterns were consistent irrespective of the subscale or outcome analyzed.

Figure 1.

Panel A. Probability of willingness to seek (mental health) care vs mistrust subscale score

Panel B. Probability of willingness to seek (mental health) care vs trust subscale score

Figure 2.

Panel A. Probability of willingness to seek (suicidal) care vs mistrust subscale score

Panel B. Probability of willingness to seek (suicidal) care vs trust subscale score

Discussion

In this cross-sectional analysis of survey data collected from 1,043 Black people living in the United States, we found evidence of a nonlinear association between medical mistrust and willingness to use mental health services. At lower levels of mistrust, there was a positive correlation (i.e., an increased willingness to seek help from a mental health professional among Black adults) between mistrust and willingness to seek help. At higher levels of mistrust, the correlation was negative. This finding suggests that there is a certain degree of wariness that may have a counterintuitive positive role (in terms of promoting mental health care seeking) but that beyond a certain point, mistrust has adverse effects on health behavior. We found a similar inverted U-pattern for the association between trust and help seeking. As expected, and consistent with previous studies, as trust increased, there was an increase in willingness to engage in health services. However, above a certain level of trust (high trust), we found an increase in trust was associated with decreased willingness to engage in mental health services. One study showed that generalized trust is rooted in optimism and a sense of control.⁵⁴ It is possible that when people have very high levels of trust, their overall sense of optimism and control may lead to a perception that they are less in need of support or mental health services. Among Black adults, one way an elevated sense of optimism and a sense of control may present is with the notion that even if one has a personal or emotional mental health problem or suicidal ideation, these issues may self-resolve and do not warrant formal intervention.^55,56

There were unintuitive findings in our study that may be driven by differences in other psychosocial factors, such as how minority groups may approach health systems and their clinicians through shared decision making. In a systematic review involving 18 studies⁴⁸, Jolles and colleagues found that some minority groups have low health literacy and may encounter barriers in communication and difficulty engaging in shared decision making because of fears around interrupting the conversation. In addition, minority groups more often preferred a clinician driven communication process. In the review, they also noted that familiarity over time with health systems may help foster trust. The authors acknowledged that they did not assess for specific differences within the minority groups based on race and ethnicity, and hence, noted their aggregate reporting may have missed possible subgroup differences. Our study specifically assessed the correlation between trust and mistrust and mental health service use. Considering the role of shared decision making and Jolles’ framework, trust and mistrust may be influenced by willingness to engage in shared decision. It is possible that patients with very high levels of trust may not feel the need to seek mental health services because they feel their doctors will make decisions on their behalf and tell them if services are needed so they may not themselves seek out services. Whereas patients with very high levels of mistrust may not feel the need to seek mental health services because they inherently have mistrust of the available mental health services.

Previous studies show that medical mistrust is associated with delay in preventive health screenings and low use of general health services for Black adults.^48–50,26 However, the specific association between medical mistrust and its relationship with the willingness to use mental health services is not well studied.^51,52 Previous studies on medical mistrust often focus on general health services, or focus on the historical origins of mistrust. In this study, we focused on willingness to use mental health services, and the current implications of mistrust in its association with willingness to engage with mental health services. Consistent with previous studies on use of general health services, higher mistrust (above a score of 3) was associated with lower willingness to use mental health service.

The measurement of mistrust varies across studies, 25% of studies on mistrust used either a single-item or a few items to measure mistrust.⁵³ While the GBMMS is one of three of the most commonly used scales to measure mistrust in a more robust manner, less robust measures are likely to be limited in value. The measurement of mistrust in our study is more comprehensive than previous studies that use single items in addressing both mistrust (8 items) and trust (4 items) as distinct variables. One reason why mistrust and trust do not function as opposite terms in this study is that the constructs within the GBMMS scale assess trust and mistrust in distinct ways. For example, an item assessing mistrust states – “doctors and healthcare workers treat people of my ethnic group like ‘guinea pigs’,” whereas an item assessing trust states - “People of my ethnic group receive the same medical care from doctors and healthcare workers as people from other groups.” The specific items used within the GBMMS scale for mistrust and trust are not opposite constructs.

Mistrust functions on a gradient in terms of how it affects use of mental health services. Based on our findings, a goal to reduce or eliminate mistrust may not be the ideal target to improve willingness to engage in mental health services. Rather, an ideal goal may be to navigate the nuances of mistrust as both a barrier and a tool in how it affects mental health service engagement. As shown in this study, one could have high mistrust below a certain gradient and continue to have an increasing willingness to engage in mental health services.

In light of our findings, eliminating mistrust may not be the most appropriate focus -- rather, addressing it may be what is more beneficial from the perspective of improving health seeking behavior and, potentially, health outcomes. Our results have demonstrated a non-linear relationship between mistrust and help-seeking. This may be reflective of the role of mistrust in the context of a healthcare system that is susceptible to systemic and structural racism that affects Black patients.^16,51 Mistrust arises from individual, systemic, and structural racism. Therefore, addressing the root causes of mistrust, rather than attempting to eliminate mistrust could lead to greater benefits for Black patients. Additionally, an approach that prioritizes institutional reform avoids imposing an undesired agenda that places the burden of overcoming mistrust on the population that experiences systemic marginalization. Such strategies to decrease medical mistrust and improve mental health outcomes include cultural humility training and supervision, strengthening workforce development initiatives to increase the representation of diverse groups of clinicians.

Study Limitations

Interpretation of our findings should be tempered by the following limitations. First, sampling was done on a convenience basis. Men are relatively overrepresented based on US population demographics. Given the focus of the study, we limited enrollment to African/Afro-Caribbean/African-American participants, which may limit the generalizability of findings. Future research should explore within racial group variations in ethnic identity and country of origin. Second, the terms used to refer to healthcare workers in the GBMMS and the GHSQ differed slightly. For example, the term “doctor and healthcare workers” is included in the GBMMS, while the term “mental health professional (e.g. psychologist, social worker, counselor)” is specified in the GHSQ. It is possible that mistrust in doctors, specifically, may be higher than mistrust in social workers and/or counselors. While it has previously been shown that medical mistrust is lower among Black people who report their usual source of medical care to be a physician’s office (vs. other sources, such as health centers, clinics, health maintenance organizations, emergency departments, and hospital outpatient department,⁵⁷ we are unaware of any data specifically comparing levels of trust stratified by type of health care provider. If true, the direction of bias in our estimated association (between mistrust and willingness to use services) would be unpredictable given the complex nonlinear association observed. Third, our data collection occurred during the COVID-19 pandemic, a period characterized by worsened population mental health⁵⁸ and an increase in awareness around racial inequities and medical mistrust in the United States.^59,60 These factors may have influenced participant responses. If the milieu increased respondents’ propensities to provide responses consistent with mistrust and also increased respondents’ propensities to provide responses consistent with not seeking mental health care, that could have biased our estimates (of the association between mistrust and mental health care seeking behavior) away from the null . However, we are unaware of a reasonable pattern of responses that could have biased responses in such a way as to produce the inverted-U shaped association we observed.

Conclusion

Given the history of discrimination by medical institutions towards underserved Black people and the continued discrimination that takes place in the healthcare setting, it is important to understand how medical mistrust informs help-seeking behaviors among Black adults. Our analysis has demonstrated a non-linear association between both trust and mistrust and willingness to seek professional help for personal and emotional problems as well as for suicidal ideation. As expected, higher levels of mistrust (scores above 3) were associated with less willingness to use mental health services. Our finding suggests that lower levels of mistrust may play counterintuitive positive role that does not prohibit willingness to engage in help-seeking behavior. Similarly, we expected trust to be associated with higher willingness to use mental health services, however, we found excessive trust may be associated with other characteristics that make help-seeking less likely, such as optimism. Understanding the role of trust and mistrust and developing interventions that engage with and address mistrust will likely promote increased willingness to seek treatment for mental health concerns, further research is needed in this area.

Highlights.

• The conceptualization of mistrust is not simply the opposite of trust, there is a need to consider the specific content of mistrust and trust as distinct concepts.

• While mistrust is known to reduce willingness to use mental health services, a more robust assessment of the effect of mistrust on use of mental health services takes into consideration the differing ranges of mistrust, such that at lower levels of mistrust, mistrust may have a counterintuitive positive correlation in its relationship with mental health service utilization.

• The authors show evidence that addressing the nuances of mistrust may be a more effective end goal in efforts to increase mental health service engagement, rather than simply attempting to eliminate mistrust.