Health disparities have been found among patients with gynecologic cancers, with the greatest differences arising among groups based on racial, ethnic, and socioeconomic factors. Although there may be multiple social barriers that can influence health disparities, another potential influence may stem from healthcare system factors that unconsciously perpetuate bias toward patients who are racially and socioeconomically disadvantaged. More recent research suggested that providers hold these implicit biases (automatic and unconscious attitudes) for stigmatized populations with cancer, with emerging evidence for patients with gynecologic cancer. These implicit biases may guide providers’ communication and medical judgments, which, in turn, may influence the patient’s satisfaction with and trust in the provider. This narrative review consolidated the current research on implicit bias in healthcare, with a specific emphasis on oncology professionals, and identified future areas of research for examining and changing implicit biases in the field of gynecologic oncology.
Introduction
Gynecologic cancers and their precancerous diagnoses remain prevalent worldwide. According to a report from the Centers for Disease Control and Prevention, it was estimated that approximately 94,000 cases of gynecologic cancers were diagnosed yearly from 2012 to 2016. Specifically, it was estimated that by the end of 2021, there will be 66,570 new cases of endometrial cancers, 14,480 new cases of invasive cervical cancers, and 21,410 new cases of ovarian cancers in the United States. These numbers may be especially troublesome for racial and ethnic minority group patients who are disproportionately diagnosed with gynecologic cancers at higher rates. For example, Hispanic, African American, American Indian, and Alaskan Native patients are more likely to be diagnosed with cervical cancer than their White counterparts. As Temkin et al discussed in their review, health disparities in cervical cancers are related to several social factors, including insurance, socioeconomic status, and lower literacy levels. In addition, compared with non-Hispanic White patients, disproportionate differences in cancer-related mortality, cancer screening use, and late-stage cancer diagnoses were found in patients with gynecologic cancer of low-socioeconomic status and racial and ethnic minority groups. These differences in health outcomes among social groups were recognized as health disparities.
Health disparities can be influenced by environmental, biological, genetic, and societal factors. Factors that are recognized as social determinants of health, including access to quality healthcare and education and income barriers, have been closely linked to health disparities, including those found among individuals living with cancer. Among these social determinants of health disparities for populations with cancer, specifically among patients living with gynecologic cancers, are the negative stereotypes and attitudes that providers can hold for certain patient groups. Research has suggested that when people hold such biases against a social group, contact with that group can activate the biases, which then guide judgment and behavior. Therefore, members of the group become subject to the type of discrimination that causes disparities in health outcomes.
Decades of study have shown that most providers (eg, physicians and nurses) consciously acknowledge that many who work in healthcare hold negative stereotypes and prejudices about racial and ethnic minority group patients. Nevertheless, because most providers also hold strong egalitarian goals for treating all patients fairly and without malice, they deny and reject the biases for themselves. Although conscious rejection of bias is a potential starting point for reducing racial and ethnic disparities, a growing body of work indicates that the biases that occur unconsciously may influence provider behavior automatically and without their awareness. This form of bias is known as implicit bias.
Research on implicit bias in healthcare has found that these biases exist among several professionals and specialties. In particular, recent research has used empirical measures of implicit bias to demonstrate that practitioners in the field of gynecologic oncology hold implicit biases toward their patients. This is particularly problematic as implicit biases can manifest through the use of negative, nonverbal (eg, posture, eye contact, and facial expressions) and paraverbal (eg, aspects that characterize one’s speech, including tone, pitch, and volume) behaviors, which can impose on components of the patient-provider relationship and impair the patient’s care and cancer-related health outcomes. Increasing the knowledge of this research among cancer care professionals, specifically in the field of gynecologic oncology, will be instrumental in reducing the inequitable, interpersonal practices and health disparities evidenced in cancer care.
This narrative review aimed to summarize the literature examining the role of implicit bias across different healthcare interactions, with a particular emphasis on cancer providers and patients. This information will be extrapolated to inform the knowledge about implicit bias in the field of gynecologic oncology. We concluded the review with a discussion about potential areas of future research on implicit bias in the field of gynecologic oncology and the current intervention and training program efforts that are being tested to decrease implicit biases among providers and trainees.
Methods
A comprehensive search was performed by the first author. Coauthors supplemented the initial search by providing additional search terms and articles not identified by the first author. Articles were found through PubMed and Google Scholar search engines and were identified using the following search words: “implicit bias,” “gynecologic,” “cancer,” “healthcare,” “health disparities,” and “health outcomes.” On the initial search starting in March 2020, articles were identified by having a title and/or abstract that contained at least one of the search words and that were published up to November 2021. From there, articles were included in our review if they were written in English, if they were published in a peer-reviewed journal, and if they presented findings on at least one of the following criteria: (1) implicit biases among healthcare providers, (2) the influence of implicit biases on providers’ medical judgments and communication toward patients, and/or (3) the mechanisms by which this impaired patient-physician communication affects patients’ health outcomes and disease prognoses. Moreover, the reference lists of the included articles were reviewed for any relevant articles.
Results
Health disparities in gynecologic oncology
Similar to other areas of healthcare, health disparities have been found among gynecologic cancers. The greatest health disparities in the field of gynecologic oncology have been observed between racial and ethnic minority group and non-Hispanic White patients. For instance, research has found that effective cancer screenings and standard treatments were used less often by racial and ethnic minority group patients than White patients. Furthermore, compared with non-Hispanic White women, non-Hispanic Black women who were diagnosed with ovarian, endometrial, and cervical cancers tended to receive less evidence-based treatment recommendations and more diagnoses at advanced stages. Of note, 1 recent study found a similar level of disparities such that the incidence of higher-grade lesions of human papillomavirus (HPV) was greater among non-Hispanic Black patients than non-Hispanic White patients. Recent findings from Ford et al may explain this disparity where non-Hispanic Black women had lower rates of being informed of their Papanicolaou test results and had higher rates of not receiving a Papanicolaou test recommendation from their provider. Sequentially, this biased care could delay non-Hispanic Black women from receiving early detection of HPV and prompt treatment. These racial and ethnic differences in cancer outcomes are more problematic as they can inflict greater financial burden to the individual and healthcare system extensively. Furthermore, this research found the greatest differences in cancer-related health outcomes between non-Hispanic White and non-Hispanic Black women diagnosed with gynecologic cancers. Among endometrial cancer diagnoses, non-Hispanic White women were found to have a higher incidence than non-Hispanic Black women. However, mortality in this population was higher among non-Hispanic Black women than non-Hispanic White women. Similarly, non-Hispanic Black women were twice as likely to die from endometrial and cervical cancers, more likely to be diagnosed at higher stages, and have poorer disease prognoses and outcomes than non-Hispanic White women. , , Moreover, these disparities have been found among American Indian or Alaskan Native women. Compared with non-Hispanic White patients, incidents and mortality from cervical cancer and mortality from ovarian cancer were greater for American Indian or Alaskan Native patients.
Although the exact cause of these disparities remains undetermined, several multisite cohort studies have suggested that social determinants of health, such as socioeconomic and insurance statuses, unequal access to care, and quality of care by site, may contribute to the widening of these health disparities in gynecologic cancers. , , In addition to race, other person-specific factors, such as insurance status, age, and region, have been linked to the unequal likelihood of receiving nonstandard treatments and guideline-adherent care as outlined by the National Comprehensive Cancer Network (NCCN) and, therefore, a greater risk of gynecologic cancer-related mortality. For instance, in an analysis of a multisite cohort of patients with ovarian cancer from the National Cancer Database, 47% of the patients who were diagnosed with advanced ovarian cancer between 2003 and 2006 were given nonstandard treatments, with non-Hispanic Black and Hispanic women accounting for most patients. In a more recent analysis of patients with cervical cancer from the California Cancer Registry, only 47% of patients between 1995 and 2009 received NCCN guideline–adherent care. Socioeconomic status was a factor in receiving this care, whereas being non-Hispanic Black and receiving nonadherent care were associated with greater mortality from cervical cancer. Moreover, limited access to care because of socioeconomic barriers seemed to be a contributor to health disparities within gynecologic cancers. Silvera et al found that Papanicolaou testing for cervical cancer screening was higher among patients who had insurance in the past 2 years than patients who were uninsured.
In certain populations, health disparities continue to persist when certain social determinants are distributed equally. However, when similar treatments are administered, non-Hispanic Black women are still more likely to be diagnosed with stage IV diseases and with higher tumor grades than non-Hispanic White women. In addition, non-Hispanic Black women have lower survival rates than non-Hispanic White women when treatment, sociodemographic, comorbidity, and histopathologic variables are held equal, whereas Hispanic women attained greater survival rates than non-Hispanic White women when these same variables are held equal. Given that health disparities between racial and ethnic minority group and non-Hispanic White patients continue to persist even when gynecologic care practices are held equal, researchers have deduced that these disparities may result from factors unrelated to the patients’ differences, such as provider implicit bias. A summary of these findings is captured in the Table .
| Author | Health disparity type | Cancer type | Source of disparity |
|---|---|---|---|
| Temkin et al, 2018 Chatterjee et al, 2016 Ford et al, 2021 Bristow et al, 2015 Chase et al, 2012 Bristow et al, 2013 Pfaendler et al, 2018 Silvera et al, 2020 | Screening and treatment | Ovarian, endometrial, cervical |
|
| Chatterjee et al, 2016 Miller et al, 2020 Long et al, 2013 Allard et al, 2009 | Late-stage and higher-grade diagnosis | Ovarian, endometrial, cervical |
|
| Bruegl et al, 2020 | Incidence | Cervical |
|
| American Cancer Society, 2021 Collins et al, 2014 Long et al, 2013 Bruegl et al, 2020 Pfaendler et al, 2018 | Mortality | Endometrial, cervical, ovarian |
|
| Long et al, 2013 | Disease prognosis | Endometrial, cervical |
|
| Bregar et al, 2017 | Survival rates | Endometrial |
|
Implicit bias in healthcare
Implicit bias among healthcare providers has been evidenced across several stigmatized groups. In their review of implicit bias in healthcare, Zestcott et al cited research demonstrating that healthcare providers hold more implicit biases toward racial and ethnic minority groups, including non-Hispanic Black, Latino or Hispanic, and American Indian or Alaskan Native patients, than non-Hispanic White patients. In addition to race and ethnicity, healthcare providers have been found to hold implicit biases toward patients who are obese, identify as gay or lesbian, mentally ill, and of low-socioeconomic status. ,
Alternatively, research has found that physicians hold implicit prejudices toward certain groups even when their explicit biases were lower or nonexistent. For instance, doctors possessed negative implicit attitudes toward American Indian, Latino or Hispanic, , and non-Hispanic Black , , patients, even in cases where they did not outwardly endorse negative attitudes toward these groups. Despite the clinical competencies they are expected to uphold, compared with the general public, providers demonstrated equivalent levels of bias and negative attitudes toward historically marginalized groups. , When providers have low self-awareness of their negative attitudes, these biased prejudices and stereotypes may inadvertently influence their communication behaviors and interactions with patients from stigmatized groups. Overall, implicit bias is thought to influence patient outcomes, including notable health disparities between racial and ethnic minority groups and cancer groups, through the provider’s poor communication ( Figure 1 ). Provider implicit bias operates by influencing their choice of poor communication strategies, lack of empathy, and biased medical decisions. These behaviors can inadvertently lead to less equitable and individualized medical care for each patient and further the health disparity divide between different social groups. Therefore, understanding the relationship between implicit bias and provider communication behaviors has become a significant emphasis of this research and a key target for intervening and reducing the impact of the providers’ implicit biases.
Implicit bias affects provider communication
Research has shown that providers communicate implicit bias through subtle and indirect means. In a study examining racially discordant interactions, greater implicit racial bias among non-Black oncologists was related to shorter interactions with non-Hispanic Black patients, lower patient ratings of the physician’s patient-centered communication, and lower confidence in the physician’s treatment recommendations. In addition, the patients recalled less of the content in their conversations with the physicians. A similar study found that physicians who had higher racial implicit bias were more likely to use words that reflect social dominance when meeting with non-Hispanic Black patients, such as a heightened use of first-person plural pronouns and anxiety-related words (eg, worried, afraid, or nervous). Furthermore, there is evidence indicating that within these racially discordant interactions, patients report higher levels of physician-to-patient talk ratios, lower levels of interpersonal care, and lower levels of patient-centered communication from their physician. , These findings were particularly concerning as biased provider-patient interactions can affect patients’ ability and willingness to adhere to treatments and to manage their illness. Furthermore, provider implicit bias may influence the type of health information communicated to patients. Lowe et al found that genetic counselors who demonstrated greater pro-White bias provided less patient-centered communication (eg, individually tailored medical information) to simulated patients who were non-Hispanic Black or Latino or Hispanic.
Specifically, when interacting with a member of a group for whom bias is held, providers may consciously express positive verbal information, but, at the same time, may express this information to the patient through negative, nonverbal and paraverbal behaviors. The incongruence between the verbal and nonverbal or paraverbal information causes the stigmatized patient to feel less positive toward the physician and less satisfied with the encounter, compared with encounters where the provider hold low levels of implicit bias toward the patient’s group. Sequentially, this can influence the patients’ satisfaction and trust in the provider and the provider’s recommendations, potentially contributing to poor patient outcomes.
Implicit bias affects patient outcomes
Biased interactions with providers influence the patients’ attitudes toward the provider and the medical care, which can indirectly impair the patient’s health outcomes. For instance, advantageous and standard of care treatments are less likely to be recommended when physicians believe that their patients may not adhere to them. , These biased recommendations are linked with lower patient ratings of satisfaction with their physician and medical care. , Similar findings have noted that these practices are evident among oncologists.
Some studies have suggested that higher oncologist implicit bias is associated with lower-quality medical judgments, such that when physicians were under greater time restrictions, they were more likely to give late-stage diagnoses among non-Hispanic Black and Hispanic patients and less likely to make referrals to other specialists for non-Hispanic Black patients than non-Hispanic White patients. When faced with limited time to adequately assess the patient’s problem, physicians may rely on their implicit stereotypes to make hasty decisions concerning the patient’s care (eg, diagnostic decisions or adherence tolerance). Of note, 1 line of thought proposes that when physicians default to their implicit stereotypes and prejudices, they are prone to employ medical decisions and clinical judgments based on preconceived and faulty generalizations of a patient group rather than on the patient’s individualized needs. Sequentially, this biased care can result in delayed evaluation and treatment that would otherwise mitigate or alleviate poor health outcomes for racial and ethnic minority group patients.
The impact of time pressures has been documented in the quality of care for gynecologic patients with physical disabilities. Both the providers and patients expressed concerns about the time pressure they experienced during the consultations. This limited time left patients feeling rushed during the examination. Providers reported limits in having the appropriate equipment and clinical resources and in having a lack of formal training toward physically disabled patients. Moreover, providers explicitly reported a high desire to improve the quality of gynecologic care for women living with disabilities, but certain barriers and challenges prevented them from doing so. This example of documented differences in physician and patient concerns demonstrates that there are challenges and barriers experienced within gynecologic care that may lead to the health disparities observed among gynecologic populations, such as those affected by gynecologic cancers.
Implicit bias in gynecologic oncology
Although more research is needed, earlier studies have suggested that oncology providers demonstrate biased attitudes and stigma for certain cancers. Sriram et al found that oncologists were more likely to hold biased implicit attitudes toward patients living with lung cancer than those living with breast cancer. Furthermore, patients, caregivers, other healthcare providers, and members of the general public demonstrated comparable levels of bias toward lung cancer that were comparable with those possessed by the oncologists. These findings were similar to the literature demonstrating physicians’ stigma toward patients living with lung cancer. , Shepherd et al found that physicians reported greater stigma toward case vignettes of patients with cervical cancer than patients with ovarian cancer. Furthermore, these physicians expressed more negative perceptions and attitudes, such as moral disgust, toward patients with cervical cancer when the cause of their cancer was linked to a sexually transmitted infection. This supported the notion that patients are likely to be stigmatized for their illness when the provider believes that the patient willingly engaged in behaviors that directly induce the illness and, therefore, were responsible for causing their illness.
To date, 1 study has been conducted using empirical research methods to examine implicit bias among gynecologic cancer care providers. Liang et al developed an Implicit Association Test (IAT), , a standard measure of implicit bias most used in social psychology research, to measure implicit bias toward patients with cervical cancer. The IAT is a computerized sorting task that measures how long it takes to match words with stimuli. As demonstrated in Figure 2 , a participant’s level of implicit bias is measured by how quickly they match bias congruent, word-stimuli pairs (eg, Papanicolaou test and cervical cancer or angry), compared with bias incongruent, word-stimuli pairs (eg, Papanicolaou test or cervical cancer and empathy). If a provider sorted the bias congruent pairs quicker than the bias incongruent pairs in this task, their score would indicate that they expressed greater implicit bias for the stigmatized group (eg, patients with cervical cancer). The researchers found that gynecologic oncologists exhibited greater levels of implicit bias toward patients with cervical cancer than patients with ovarian cancer. Furthermore, nurses were found to demonstrate higher levels of implicit prejudices and stereotypes toward patients than oncologists. Although the mechanisms are not yet understood, these contrasting levels of implicit bias between medical providers may be a result of the amount and quality of exposure that nurses have with patients compared with physicians. Thus, having more contact with a patient, and potentially having more negative interactions with that patient, can reinforce and activate the provider’s stereotypes and/or prejudices of the patient. Although more research is needed to understand implicit biases in gynecologic cancers, these findings have suggested that gynecologic cancer care providers exhibit implicit biases, specifically for patients with socially stigmatized cancers.
