Bias in healthcare is any tendency, conscious or unconscious, that leads a provider to treat patients differently based on characteristics like race, gender, weight, age, or income rather than medical need alone. It shapes which tests get ordered, how seriously symptoms are taken, and what treatment options are offered. The consequences are measurable: misdiagnoses, delayed care, and worse outcomes for millions of people every year.
Implicit vs. Explicit Bias
Healthcare bias falls into two broad categories. Explicit bias involves preferences and beliefs a person is aware of and can articulate. A provider who knowingly refuses treatment based on a patient’s background is acting on explicit bias. These cases, while harmful, are relatively straightforward to identify.
Implicit bias is harder to pin down. These are unconscious mental processes that create automatic associations and reactions without the person’s awareness. A clinician may genuinely believe they treat every patient equally while still making faster, less thorough assessments for certain groups. Because implicit bias operates below the level of conscious thought, the person carrying it often cannot recognize or report it. That disconnect between intention and behavior is what makes it so persistent in clinical settings.
Racial Bias in Pain Treatment
One of the most thoroughly documented forms of healthcare bias involves pain management. A systematic review and meta-analysis covering 2011 to 2021 found that Black patients were roughly 17% less likely to receive opioid pain medication than White patients with similar conditions. Hispanic patients fared even worse, at about 20% less likely to receive the same prescriptions. These gaps persisted even when researchers controlled for the type of pain and clinical setting.
The disparity extends well beyond prescriptions. Between January 2018 and June 2024, the pregnancy-related mortality rate for Black individuals was 68.0 deaths per 100,000 live births, compared to 26.3 for White individuals. That’s a ratio of roughly 2.6 to 1. For context, peer nations typically have a national maternal mortality rate of 10 per 100,000 or lower. The rate for Black individuals in the U.S. alone exceeded 50 in 2023.
Gender Bias in Heart Disease
Women are 50% more likely than men to be misdiagnosed when having a heart attack, despite carrying comparable cardiovascular risk. Their symptoms are frequently attributed to anxiety or gastrointestinal problems rather than cardiac events. This pattern of misdiagnosis leads to longer wait times for treatment and fewer referrals for follow-up care.
The downstream effects compound quickly. Women with coronary heart disease are 2.5 times less likely to be referred to a cardiologist than men with the same condition. They receive fewer diagnostic tests, including angiography and electrocardiograms, and are less frequently hospitalized. None of this reflects a biological difference in how heart disease should be managed. It reflects assumptions about who “looks like” a heart attack patient.
Weight Stigma and Missed Diagnoses
Providers tend to spend less time consulting with patients in larger bodies and invest less effort in building rapport, being empathetic, or involving those patients in shared decision-making. Health conversations often narrow to weight loss, crowding out other concerns. One study found that for every increase in BMI, providers were significantly less likely to ask patients about sexual health or family planning, meaning routine preventive care quietly drops off.
The consequences go beyond a shorter appointment. Women in larger bodies have been refused removal of implanted contraceptives, denied infertility treatment until they lost weight, and barred from natural or home births. Reduced referrals for cancer screenings like pap smears and breast exams can delay early detection. Perhaps most critically, patients who experience weight stigma in healthcare settings are more likely to avoid seeking care in the future, creating a cycle where the bias itself becomes a barrier to treatment.
Age and Socioeconomic Bias
Ageism in clinical settings swings in both directions. Undertreatment happens when providers dismiss treatable conditions, like joint pain, as a “normal” part of aging, allowing them to worsen without intervention. Overtreatment looks like aggressive chemotherapy for an older patient who has clearly expressed a preference for comfort-focused care. In some cases, providers skip conversations about preventive care altogether, as though screening and early detection no longer matter past a certain age.
Income and insurance status introduce their own distortions. Research has found that doctors offer more aggressive treatment options to patients with higher-status occupations, even when those same doctors explicitly state that social status plays no role in their decisions. Patients with higher education and income are significantly more likely to have their heart rhythm monitored before a cardiac arrest, while older adults with low income and no insurance are less likely to receive a primary cardiac diagnosis at all. Communication style shifts too: providers tend to adopt a more collaborative, shared decision-making approach with white-collar patients while defaulting to a more directive, doctor-centered model with blue-collar patients.
Women with lower socioeconomic status face a specific version of this pattern. Studies have found they receive biased guidance that steers them toward longer-term and sometimes irreversible contraceptive options, narrowing their reproductive choices in ways that wealthier patients don’t experience.
How Algorithms Reproduce Bias
Bias doesn’t require a human in the room. A landmark study published in Science examined an algorithm used to manage care for millions of patients across the U.S. health system. The tool was designed to identify patients who would benefit from extra medical support, but it used healthcare spending as a proxy for healthcare need. Because Black patients historically have less money spent on their care due to unequal access, the algorithm concluded they were healthier than equally sick White patients.
The scale of the error was striking. Fixing the algorithm’s logic would have increased the percentage of Black patients flagged for additional help from 17.7% to 46.5%, meaning more than half of the Black patients who should have been identified for extra care were being missed. The problem wasn’t malicious design. It was a seemingly reasonable shortcut, using cost data instead of illness data, that quietly encoded decades of inequality into an automated system.
Why Bias Training Has Limits
Many hospitals and health systems now mandate implicit bias training for staff. Since January 2023, the Joint Commission has required accredited hospitals to collect sociodemographic data, stratify quality and safety metrics by patient characteristics, and gather feedback from patients and staff to identify perceptions of bias. About 95% of hospitals now collect self-reported race and ethnicity data, and roughly 78% train staff on how to do so sensitively.
The infrastructure is growing, but evidence for the training itself remains mixed. Literature reviews have found that many bias-reduction interventions show no measurable effect on clinical behavior, and some may even worsen implicit biases, particularly when participants are simply told to avoid stereotyping. Much of the supporting research has lacked control groups, longitudinal follow-up, or real-world clinical settings. A meta-analysis of the Implicit Association Test, the most widely used tool for measuring unconscious bias, found it accounted for only about 2% of the variance in actual behavior.
This doesn’t mean awareness efforts are pointless, but it does suggest that training alone, without changes to clinical workflows, decision-support systems, and institutional accountability structures, is unlikely to close the gaps that bias creates.