Bias - State of the art

Research on bias in health professions education (HPE) has expanded significantly over the past two decades. What began as a focus on individual attitudes has evolved into a broader examination of systems, structures, and institutional design.

What is the current understanding of bias?

Most recent research has been on implicit bias. Implicit bias in health professions education is now widely recognized as pervasive, consequential, and multi-level. Current scholarship integrates insights from cognitive psychology, social science, and organizational theory to show that bias influences how learners are evaluated, whose knowledge is valued, and how professional judgments are made — often without conscious intent.

A major shift in the field has been moving beyond the idea that bias is primarily an individual problem. While cognitive processes remain important, researchers increasingly emphasize that bias is shaped, amplified, or constrained by educational systems. Structures that rely on unstructured judgment, vague criteria, informal norms, and high-stakes interpretation create fertile ground for patterned inequities.

As a result, contemporary work in the field focuses not only on raising awareness but on redesigning systems. Transparency, structured processes, reflexivity, and accountability are now understood as components of educational quality — not optional additions.

Expert Insight:
“The field has moved from asking “Who is biased?” to asking “What conditions make bias more or less likely to shape outcomes?” 

– Nicole A. Perez

How did we get to current understanding?

Modern HPE developed through reforms that prioritized standardization, efficiency, and scientific rigor. These reforms strengthened the profession in important ways. At the same time, they shaped norms about who belongs in the health professions, what excellence looks like, and how competence should be expressed.

Assessment systems were largely designed to sort and rank learners, not to interrogate inequities. Over time, these systems became normalized.

In the late 20th century, concerns about inequity were often addressed through individual-level interventions,(e.g. cultural competence training or bias awareness workshops). These efforts were important in naming bias and opening conversation, however, evidence accumulated showing that awareness alone did not eliminate persistent group-level differences in evaluation, advancement, and recognition.

This prompted deeper inquiry. Scholars began examining how assessment practices, institutional incentives, informal sponsorship patterns, and definitions of professionalism might reproduce inequities, even in the absence of explicit prejudice.

Expert Insight:
“Today we know that bias is not only about individual cognition, but also about how systems are designed.”    – Nicole Perez

What the evidence shows

Empirical research across HPE reveals several consistent patterns.

In assessment, small differences in narrative language or ratings can accumulate into significant downstream consequences. Subtle variations in tone, specificity, or interpretation influence grades, awards, residency placement, and long-term career trajectories. These differences often emerge in contexts of time pressure and subjective judgment.

In feedback, learners from historically underrepresented groups are more likely to receive personality-focused or corrective comments, while historically overrepresented groups receive more actionable and developmental guidance. Over time, these patterns shape confidence, opportunity, and professional identity.

In clinical reasoning and decision-making, cognitive biases influence what information is prioritized, how uncertainty is interpreted, and which explanations are considered plausible. Notably, these effects occur even among highly trained clinicians, reinforcing that expertise alone does not eliminate bias.

Taken together, the evidence has shifted the conversation from identifying biased individuals to examining patterned outcomes produced by everyday practices. 

What are tomorrows challenges

The current frontier in bias research focuses on moving from recognition to transformation.

This includes:

  • Designing assessment systems that are more resistant to bias by default;
  • Embedding equity into definitions of educational quality and patient safety;
  • Supporting longitudinal faculty development rather than one-time training; and
  • Making outcome data visible and reviewing disparities systematically.

A growing area of inquiry also concerns data-driven tools and artificial intelligence in education and clinical care. While often framed as objective, these technologies can reproduce or amplify existing inequities if their assumptions and training the data informing are not critically examined.

Across these developments, a shared insight is emerging: addressing implicit bias is less about correcting individuals and more about building institutions capable of identifying and learning from their own biased outcome patterns.

If you only have time to read a few papers

The following readings offer entry points into key shifts in the field:

Sukhera, J., Watling, C. J., & Gonzalez, C. M. (2020). Implicit Bias in Health Professions: From Recognition to Transformation. Academic Medicine, 95(5), 717–723. https://doi.org/10.1097/ACM.0000000000003173

This article argues that most implicit bias curricula in the health professions focus narrowly on awareness and attitude change, which limits their impact on equity. The authors propose using transformative learning theory to design longitudinal, reflexive, and system oriented approaches that move from simple recognition of bias toward genuine transformation of learners and institutions. 

Mangalindan, K. E., Wyatt, T. R., Brown, K. R., Shapiro, M., & Maggio, L. A. (n.d.). Investigating the Road to Equity: A Scoping Review of Solutions to Mitigate Implicit Bias in Assessment within Medical Education. Perspectives on Medical Education, 14(1), 92–106. https://doi.org/10.5334/pme.1716

This scoping review maps interventions aimed at mitigating implicit bias in assessment, categorizing them by the types of bias addressed, the targets of intervention (eg faculty, systems), and how effectiveness is defined and measured. The authors conclude that although promising strategies exist, the evidence base is fragmented and calls for multi institution, theory informed, and rigorously evaluated interventions that tackle structural as well as individual level bias.

Teherani, A., Perez, S., Muller-Juge, V., Lupton, K., & Hauer, K. E. (2020). A Narrative Study of Equity in Clinical Assessment Through the Antideficit Lens. Academic Medicine, 95(12S), S121–S130. https://doi.org/10.1097/ACM.0000000000003690

This narrative study interviews UIM fourth year medical students and senior residents about their own stories of achievement in clinical training and how assessment practices can equitably capture those successes, intentionally shifting focus from deficit narratives to an antideficit lens. UIM learners described equitable assessment as sound, achievement focused systems that include frequent direct observation with real time feedback, clear expectations, longitudinal relationships, narrative assessments, minimized peer comparison, explicit attention to learner identity, and supervisor training to reduce bias.

Kakara Anderson, H. L., Govaerts, M., Abdulla, L., Balmer, D. F., Busari, J. O., & West, D. C. (2025). Clarifying and expanding equity in assessment by considering three orientations: Fairness, inclusion and justice. Medical Education, 59(5), 494–502. https://doi.org/10.1111/medu.15534

This conceptual paper clarifies equity in assessment by distinguishing three orientations, fairness oriented assessment, assessment for inclusion, and justice oriented assessment, each reflecting different priorities and strategies. The authors argue that moving beyond a narrow focus on fairness toward inclusion and justice requires rethinking assessment purposes, stakeholder roles, and the broader sociopolitical context in which assessment occurs.

Bornstein, B. H., & Emler, A. C. (2001). Rationality in medical decision making: A review of the literature on doctors’ decision‐making biases. Journal of Evaluation in Clinical Practice, 7(2), 97–107. https://doi.org/10.1046/j.1365-2753.2001.00284.x

This review synthesizes evidence on how a variety of cognitive biases, such as anchoring, availability, and framing effects, lead physicians to gather and interpret diagnostic and treatment information in suboptimal ways. It also discusses strategies to reduce these biases, including decision support tools, feedback, and structured reflection, while noting that complete debiasing is unlikely and that system level safeguards remain important. It is one of the foundational papers that shaped how the field first understood cognitive bias.

Norman, G. R., Monteiro, S. D., Sherbino, J., Ilgen, J. S., Schmidt, H. G., & Mamede, S. (2017). The Causes of Errors in Clinical Reasoning: Cognitive Biases, Knowledge Deficits, and Dual Process Thinking. Academic Medicine, 92(1), 23–30. https://doi.org/10.1097/ACM.0000000000001421

This article examines causes of errors in clinical reasoning and argues that while cognitive biases in fast, Type 1 thinking are often blamed, knowledge deficits and limitations of slower, Type 2 reasoning also play substantial roles. The authors conclude that improving diagnostic accuracy requires strengthening domain specific knowledge and pattern recognition in parallel with thoughtful use of analytic reasoning rather than relying solely on generic debiasing techniques.