Implicit Bias in Healthcare: 2018
Implicit Bias in Healthcare: 2018Kirwan’s previous State of the Science reports emphasized differential treatment of patients due to implicit bias among healthcare providers (see also Burgess et al., 2014; Chae et al., 2014; Moskowitz et al., 2012; Sabin et al., 2009). Consistent with this theme, studies published in 2018 continued to investigate the pervasiveness of implicit racial bias within healthcare systems and its adverse impacts on the quality of care as well as patient outcomes. A total of 18 articles from the year 2018 were included in the article dataset. In the following section, we examine insights here as they pertain to the manifestation of implicit bias in healthcare. Specifically, we synthesize the findings regarding the potential negative impacts of implicit racial bias on patient care, including a discussion of disparate quality of care and differential standards of medical decision-making (Goodwin, 2018). The impact of implicit bias isn't confined to general healthcare alone but extends its reach to various specialized healthcare services, each presenting unique challenges.
Our 2018 Healthcare State of the Science takes a deep dive into the question of who in the healthcare system is influenced by implicit bias. We found articles focused on healthcare providers across specialties (Maina et al., 2018; Tajeu et al., 2018); medical trainees (Backhus et al., 2019; Hernandez, 2018); and mitigation strategies. The three articles that explicitly focus on mitigation are combined with recommendations from other years into a single cumulative mitigation strategies section for the healthcare sector. Articles published in 2018 affirm the pervasiveness of implicit racial bias among providers from multiple specialties (pediatrics, obstetrics, mental health) and trainees undergoing medical education (medical students and residents). Even patients themselves (Greene et al., 2018) are susceptible to being influenced by implicit racial bias as they make decisions about their healthcare.
Impact of Healthcare Provider’s Implicit Racial Bias on Patient Care
Our data collection strategy uncovered three medical specialties that used a variety of methods (quantitative, qualitative, mixed methods and meta-analysis) to arrive at their conclusions. In this section we focus on obstetrics, pediatrics, and mental health implicit racial bias research published in 2018. We cover additional specialties in 2019 and 2020 as applicable based on the results of our data collection strategy. All three of the specialties we analyzed for 2018 acknowledged the intersection of the healthcare system with other important and relevant systems that are also grappling with significant racial disparities: the child welfare system (obstetrics and pediatrics) and the criminal justice system (mental health).
Obstetrics
Obstetrics and gynecology as a medical specialty has a particularly fraught history with explicit racial bias and discrimination, which has had a direct effect on the quality of patient care. Several articles published in 2018 allude to this broader history as well as provide evidence of ongoing implicit racial bias. Brenda Pereda and Margaret Montoya (2018) provide a broad review of the role of implicit bias and note specifically the history of non-consensual medical experimentation on enslaved women and forced sterilization of Latinas as part of the embedding of bias. In a similar vein Erin Thomas (2018)’s study of bias among lactation consultants suggests that such bias originates from larger inequalities in society and must be addressed at their root.
Through semi-structured interviews with 36 lactation consultants, Thomas (2018) documented instances of race-based discrimination against patients seeking lactation care and connected these findings to the implicit bias literature. Through an analysis of interview transcripts, the researcher identified three subthemes of implicit bias, including lower quality and less quantity of care. Thomas used actual statements given by the interviewees to show how Black and Hispanic mothers were denied proper assistance with breastfeeding because healthcare providers believed it would be a waste of resources. Similar biases toward various Asian groups were also observed, particularly those who spoke English as a second language. Thomas (2018) explicitly connected these biases to quality of care, finding that White patients received more services and less referrals to social services in lactation care compared to Black patients. This was especially prevalent in cases where a patient used drugs or showed signs of mental illness. Although the study did not directly measure implicit bias, it presented us with compelling evidence of its existence. Of special interest to our analysis, Thomas noted negative pediatric impacts of positive implicit racial bias, where White infants were at higher risk when parents displayed evidence of drug abuse or mental illness left unacknowledged by lactation consultants.
This study was accompanied in 2018 by the publication of a national study of 286 perinatal physicians by Natasha Shapiro and colleagues (2018), which used a vignette strategy to explore the role of implicit bias regarding both race and socioeconomic status. In this study, physician participants were instructed to view a vignette featuring a female patient in imminent labor. The patient's race was indicated by a photograph of either a White or Black individual, while their socioeconomic status was implied by mentioning the care is provided by either a "Private Obstetrician" for high socioeconomic status or by the "Resident OB clinic" for low socioeconomic status. After being randomly assigned to view one of the four versions of the vignette, participants were then asked about how likely they would recommend the patient for intensive care versus comfort care.
Following this choice, participants’ implicit biases were also measured through two Implicit Association Tests (IATs) focused on race and socioeconomic status respectively. The results revealed that the participating physicians demonstrated an implicit preference for White patients and those with higher socioeconomic status. However, it was found that these implicit biases did not influence the physicians' treatment decisions based on the patient's race. The researchers suggested that this could be a result of physician’s higher self-awareness of racial biases and their impact on decision-making, making them more cognizant of not letting these biases influence their recommendations. The researchers also started with a sample of 971 physicians who initially started the study; 685 physicians did not complete the entire study so there is more work to be done to understand the relationship between implicit racial bias and implicit socioeconomic bias in obstetrics. Interestingly, some prior research suggests that someone whose behavior does not necessarily reflect the biases might still be influenced by social or cultural norms in their responses to IATs Fazio & Olson (2003), which is consistent with some of the literature in medical education that points to mitigation strategies that focus on the meso- and macro-level of analysis. We take up this point in the mitigation strategies section of this report.
Pediatrics
As with obstetrics scholarship in 2018, pediatrics research published in the same year also focused on the impact of implicit biases on quality of care in both qualitative and quantitative studies. Michele Goodwin (2018)’s analysis of the Jahi McMath case, where she documented the tragic experiences of a 13-year-old girl who suffered severe complications after receiving a routine tonsillectomy, which is the second most commonly performed pediatric surgical procedure in the United States (Jaryszak et al., 2011). Goodwin cited several statements from Jahi’s family questioning whether she received the standard quality of care. While acknowledging the tragedy of McMath’s death Goodwin suggested that there are numerous social implications regarding the influence of healthcare providers’ implicit biases upon the quality of care received by patients who are members of minoritized racial groups.
In a similar vein Kent Hymel et al. (2018) conducted a cross-sectional analysis of 500 pediatric patients under three years of age with acute head injuries to examine the potential influence of implicit bias among physicians who specialize in Abusive Head Trauma (AHT). Here the researchers used the evaluation and reporting frequencies of AHT as an indirect measure of implicit racial bias. Their findings showed that racial and ethnic minority patients were evaluated and reported more frequently for suspected AHT than White (non-Hispanic) patients, which is considered an indicator of implicit racial bias in reporting standards. Similar to the Thomas (2018) study, Hymel et al. (2018) suggest “positive” implicit racial bias, like giving White parents the “benefit of the doubt,” when assessing for potential child abuse, can have negative outcomes for White non-Hispanic babies under three years old. The consequences are critical, even if they exist predominantly in lower-risk patients, because delays in AHT diagnosis have been shown to lead to re-injury or even death.
Both obstetrics and pediatrics authors suggest that having too many false positives among minoritized caregivers in AHT cases and too few positives in lactation cases are detrimental both for the young patients themselves (infants and children up to three years old) but also for the parents, who either face unwarranted referrals to child welfare services or law enforcement OR do not get adequate treatment for the drug dependence or mental health concern, leaving children at risk in either outcome.
Mental Health
Four (4) studies published in 2018 analyzed the role of implicit racial biases in mental healthcare, using a variety of empirical methods. Similar to their peers in other specialties, healthcare providers in the mental health field also may hold similar levels of implicit bias against certain patients based on their race, ethnicity, and sexual orientation (Hall et al., 2015; Merino et al., 2018; Snowden, 2003). That said, Yesinia Merino and colleagues (2018) conclude that mental health care is particularly vulnerable to the effects of an individual provider’s implicit racial bias because it heavily relies on one-on-one services where a single provider acts as the gatekeeper to care. Ivy Maina and colleagues (2018) conducted a meta-analysis to consolidate findings of 37 studies published between 1997 and 2015. Their review found that there is growing evidence suggesting that similar to the general U.S. population, a substantial number of healthcare providers across different training levels and disciplines hold implicit biases against Black, Hispanic, American Indian, and dark-skinned individuals more broadly. Maina and colleagues found that studies focused on real-world patient care yielded more consistent findings, which revealed an association between higher implicit bias and less favorable patient-provider interactions.
Merino and colleagues (2018) argued that the potential impact of negative implicit bias reaches various aspects of mental health care, encompassing not only clinical screening but also diagnosis, treatment processes, and crisis response. Furthermore, they also explored the intersection of implicit bias in mental health care and criminal justice institutions. The authors emphasized the importance of this exploration as prison systems have become one of the largest mental health providers in the U.S. (James & Glaze, 2006). In their conclusion, Merino and colleagues reiterated the need to understand how individual practitioners' implicit attitudes influence the delivery of mental health services and treatment effectiveness. Additionally, the authors also urged future investigations to not overlook how structural inequities within the mental health system could perpetuate implicit bias. The structural inequities they identified include the lack of diversity in the mental health workforce and among mental health educators, evidence-based practices that are centered on a majority-group framework, and the lack of diversity on the boards of mental health agencies (Merino et al., 2018).
While Javeed Sukhera et al. (2018b) and George Bermudez (2018) focus more on mitigation strategies, each documents the presence of implicit racial bias among mental health care providers prior to engaging in mitigation. Sukhera et al. focuses on a group of psychiatry faculty and residents, while Bermudez focuses on a group of faculty, students and administrators from various disciplines using psychoanalytic techniques of social dreaming. Analyses of their proposed mitigation strategies in connection with other research can be found in the cumulative mitigation strategies section.
Conclusion
The studies discussed in the 2018 State of the Science use a variety of methods to document the ongoing challenges presented by the role of implicit racial bias among healthcare providers and implicit racial bias’s connection to systemic barriers in the social welfare and criminal justice systems. While recommendations for mitigation are explored in a different section of this report, the articles described here offer both general and specialty-specific ideas for future research. Both Maina et al. (2018) and Thomas (2018) identified a need to move beyond a focus on physicians’ implicit bias to include other types of healthcare providers in the mental healthcare specialty, with Thomas (2018) noting the particular role that lactation consultants play with new parents of infants. To meet this need Tajeu et al. (2018) conducted a statewide study of 107 healthcare staff, which included physicians, nurses, and non-MD/RN personnel such as receptionists. The authors found that White staff members had higher implicit racial bias against Black patients than the physicians or nurses, identifying a potential gap for both further research and mitigation strategy development.
As we noted in the general introduction to the Healthcare State of the Science report, the “who” in healthcare has continued to expand beyond simply physicians. Similar to healthcare providers and non-MD/RN staff, patients can also exhibit implicit racial bias, especially when selecting their physicians, but this also remains a relatively understudied area. Jessica Greene and colleagues (2018) asked study participants to hypothetically choose from four physicians based on their performance quality information. Among the four, two physicians had top performance ratings and two had lower quality performance ratings. The study manipulated the names of the top-performing physicians to represent different races/ethnicities and genders. The findings revealed that respondents more frequently selected a physician with a typically White male name compared to the same physician with either an African American male, African American female, or Middle Eastern name. This bias was more pronounced among White respondents. Greene and colleagues cautioned that it remains unclear if this bias found in an isolated experiment setting can translate into real-world physician selection. If it does, however, it could result in racial and ethnic minority and women physicians having smaller patient panels (Ly et al., 2016). Given the increasing enrollment in health plans with limited provider networks (Polsky & Weiner, 2015) and the painfully slow but steady diversification of the medical profession, the potential influence of implicit bias in patients' choice of new physicians becomes particularly relevant.