Implicit bias and the gender inequity in liver transplantation

Abstract

Women with cirrhosis awaiting liver transplantation are less likely to receive a transplant and more likely to die than men. While differences in body size and estimation of kidney function are well-studied contributors to this gender inequity, what has received relatively little mention as a potential contributing factor is the possibility of implicit bias. Implicit bias is defined as “any unconscious or unacknowledged preference that affects a person’s outlook or behavior.” The undeniable presence of implicit bias, a factor that is known to negatively influence health outcomes for women, within our health care system means that patients interacting within our transplant system may still experience unequal treatment despite our best efforts to modify the allocation system at the national level. Awareness of this additional source of gender-based disparities is the first step. In this article, we posit that implicit bias in liver transplantation may exacerbate the gender inequity in transplant access and provide examples in the literature to support this assertion. Lastly, we offer strategies that could be applied at the individual or the healthcare delivery system levels to help reduce the influence of implicit bias on the gender inequity in liver transplantation.

Women with cirrhosis awaiting liver transplantation are less likely to receive a transplant and more likely to die than men.^1–7 This has been reported since the implementation of the Model for End-Stage Liver Disease (MELD) score for liver allocation and persists to this day. Several explanations have been proposed. The two leading explanations—and the ones with the greatest weight of evidence to support them—are: (1) the underestimation of mortality risk due to renal dysfunction by serum creatinine in the MELD score, and (2) the relative scarcity of size-appropriate livers due to the majority male (and therefore, physically larger) donor pool.^5,8 But what has received relatively little mention as a potential contributing factor to this gender inequity is the possibility—if not, inevitability—of implicit bias.

Implicit bias is defined as “any unconscious or unacknowledged preference that affects a person’s outlook or behavior” (Oxford English Dictionary). These preferences can be favorable or unfavorable and are thought to be “automatic” in that they are triggered without an individual’s intention or conscious awareness.⁹ There are many forms of implicit bias—including affinity/similarity bias, height bias, halo effect, contrast effect, to name a few—highlighting the complexity of this issue and the near inevitability of implicit bias affecting healthcare delivery. Also, the influence of implicit bias in healthcare is well-established. A systematic review of 42 studies evaluating implicit bias in healthcare professionals concluded that physicians and nurses manifested implicit biases to a degree similar to that of the general population, but that these biases impact clinical judgment and behavior.¹⁰ While the majority of these studies focused on implicit bias and race or ethnicity, a few highlighted the relationship between implicit bias and gender-related preconceptions (and none focused on the intersectionality between race and gender). For example, in a study of physicians’ interpretations of an actor or actress playing the role of a 55-year old patient presenting with classic coronary heart disease symptoms, physicians were twice more likely to diagnose the patient with a mental health condition if the patient in the scenario was a woman versus a man.¹¹ In a separate study, women were asked fewer questions about potential coronary heart disease risk factors and received less counseling about smoking cessation than men.¹² Among patients being evaluated for advanced heart failure therapies, women, particularly Black women, were perceived to have less adequate social support than men.¹³ Among patients with end-stage renal disease, gender, age, race, and obesity have all be found to be associated with perceptions of suitability and therefore referral for kidney transplantation.^14–16 No study, to our knowledge, has purposefully evaluated the role of implicit gender-related bias in liver transplantation.

While implicit bias, per se, does not necessarily result in unfavorable outcomes—it could have a positive or even no influence on outcomes—we posit that implicit bias in liver transplantation negatively impacts outcomes for women awaiting liver transplantation, exacerbating the gender inequity in transplant access. There is evidence to support this assertion. Women awaiting liver transplantation experienced a 10% increased risk of delisting for being “too sick” for transplantation compared to men despite similar metrics of disease severity.¹⁷ In other words, transplant centers were more likely to actively remove a woman from the waitlist than a man (as opposed to patients dying on the waitlist), potentially reflecting differences in provider behavior based on patient gender rather than biologic differences in the patients. Among waitlist candidates (both women and men) of “smaller stature” in the same match run position, transplant clinicians were more likely to decline a liver offer for women of average stature awaiting liver transplantation compared to men of average stature; however, small women experienced an additional 18% higher odds of organ decline compared to small men, a difference that cannot be explained by size mismatch.⁶ Lastly, in a study evaluating clinicians’ subjective assessments of “overall health” in their own patients awaiting liver transplantation, gender was the only patient characteristic that was associated with the subjective clinician assessment; clinicians, on average, rated their female patients as having worse overall health (on a scale of 0 to 5) than men.¹⁸ Taken together, we offer, as a matter of conjecture, that perceptions of frailty and “strength” to tolerate surgery, ingrained beliefs about the adequacy of social support based on “traditional” family roles, or stereotypes about mental health in women may be applied differentially in women versus men, albeit subconsciously. In the absence of objective data about the patient, this might lead to inaccurate conclusions that women are “too frail” for transplant, have “inadequate social support,” or have an “uncontrolled” mental health condition that poses a barrier to transplant, compared to male counterparts (all other factors being equal).

Assuming these data represent the influence of implicit bias on the gender inequity in liver transplant waitlist outcomes, can anything be done, or must we simply accept this as an immutable contributing factor? A number of interventions have been studied in nontransplant settings. One recent systematic review of 47 interventions to reduce implicit bias revealed a few effective strategies, such as exposure to counterstereotypical exemplars or intentional strategies to suppress individual’s biases; notably, only one intervention was targeted toward gender biases (all others targeted racial/ethnic biases).¹⁹ Although only a handful of interventions have been applied in the healthcare setting, the Joint Commission, the US institution that accredits healthcare organizations and programs nationally, recommends dedicated clinical training that emphasizes three individual skills—perspective-taking, emotional regulation, and partnership-building.²⁰ The majority of efforts to reduce implicit bias in healthcare have focused on the critical outcomes of reducing implicit bias related to race or ethnicity,¹⁹ but these strategies would likely reduce the influence of gender stereotypes as well. At last, individuals can explore their own implicit biases around a range of topics including gender through the Implicit Association Test (discussed in depth in the book Blindspot: Hidden Biases of Good People²¹) and they are made available at https://implicit.harvard.edu/implicit/index.jsp.

However, if implicit bias is unconscious, automatic, and triggered beyond the level of conscious awareness, then strategies must be applied not just at the individual level but at the center level to be effective at reducing the influence of implicit bias on the gender inequity in liver transplantation. We propose that concrete strategies that liver transplant centers can implement to alleviate the influence of implicit bias on the gender inequity in liver transplantation be implemented. First, decisions regarding whether a patient is “too frail” for liver transplant should be anchored in objective data using standardized frailty assessment tools that have been studied in liver transplant patients²²; frailty assessments should be made at baseline and longitudinally to determine not just whether a patient is too frail but has become too frail during her time on the waitlist. Second, centers should fortify their psychosocial evaluations with transplant-specific standardized assessment tools, but importantly, if there are concerns about a specific aspect of the psychosocial evaluation, then discussion about delisting should be anchored by objective data from that subdomain of the standardized tool. For example, a common transplant-specific tool is the Stanford Integrated Psychosocial Assessment for Transplant,²³ but any concerns about social support (one of the most influential factors in listing decisions²⁴) should focus on the scores associated with social support subdomain specifically, not simply the summary risk assessment. Third, centers should consider assigning a formally designated advocate for each waitlist candidate, whose responsibility is to anchor selection committee discussions and subsequent decisions around objective data, as has been suggested by other experts in the field.²⁵ Better yet, liver transplant selection committee members could be “blinded” to the patient’s gender when forming decisions around the patient’s transplant candidacy (although this would not be possible for the few providers who directly evaluated the patient). At last, it should be required that all transplant clinicians engage in formal diversity, equity, and inclusion programs to better explore one’s own biases and how they can overcome those biases to more effectively collaborate with colleagues in patient selection and transplant decision-making. Such programs established within and mandated by healthcare systems may help to, over time, change the culture to promote gender equity in healthcare delivery.²⁶ We believe that these steps are incremental in that they build upon existing protocols for candidate evaluation and decision-making, and therefore, offer a pragmatic approach that can alleviate the contributions of implicit bias to the gender inequity in liver transplantation.

Additional research and larger structural changes will also be needed. Research on the persistent gender inequity in liver transplant waitlist outcomes have thus far emphasized issues surrounding prioritization of candidates and distribution of organs.⁸ The subsequent debate stimulated by this research has been productive, as efforts are currently underway to modify liver allocation with the explicit intention of reducing the gender inequity.²⁷ However, research regarding how implicit bias impacts the trajectory of women with end-stage liver disease through our medical system and their chances of survival is urgently needed. This starts with bias surrounding diagnosis, treatment, and transplant referral rates. Specifically, within the context of recipient selection, listing/delisting, and organ acceptance, investigation into the impact of implicit bias at the level of individual providers and the selection committee, as well as the level of the surgeon making decisions about individual organ offers, should be a priority. In many specialties, there is an evolving literature regarding the role of provider gender in implicit bias and clinical decision-making, as well as the role of implicit gender-related bias in medical and surgical training itself.^28,29

The undeniable presence of implicit bias within our health care system, a factor that is known to negatively influence health outcomes for women, means that patients interacting within our transplant system may still experience unequal treatment despite our best efforts to modify the allocation system at the national level. However, this can improve over time; one study of 4.4 million tests of implicit attitudes in the U.S. towards a spectrum of factors (sexual orientation, race, skin tone, age, disability, and body weight) demonstrated a shift of implicit attitudes toward neutrality for some factors (sexual orientation, race, and skin tone) over the course of a decade.³⁰ Awareness of implicit bias as a threat to objective transplant decision-making is the first step. With awareness, we hope that champions in each center will step up to enhance existing transplant protocols with the standardized assessment tools and procedures that will reduce the influence of implicit bias on this gender inequity in liver transplantation.

Abbreviation:

MELD

Model for End-Stage Liver Disease

DATA AVAILABILITY STATEMENT

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.