Sex and Gender Bias in Clinical Practice: Evidence from a Simulation-Based Quasi-Experimental Study in Spain
A quasi-experimental study across four Spanish medical faculties reveals that implicit sex and gender bias significantly lowers the quality of care provided to female patients during the early, interpretative stages of clinical decision-making, such as symptom assessment and diagnosis, while later protocolized stages remain unaffected.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the quiet moments before a diagnosis is made, a doctor stands at a crossroads of information. They listen to a patient's story, weigh physical signs, and decide what questions to ask next. This process relies on two things: the hard facts of biology and the human act of interpretation. Biology is the body's machinery, where hormones and genetics create real differences between men and women in how diseases appear and progress. But the human mind also plays a role. Doctors, like all people, carry unconscious assumptions—subtle mental shortcuts formed by years of observing the world. These assumptions can shape how a clinician listens, what they notice, and how they speak to a patient. When these hidden biases align with deep-seated stereotypes about gender, they can subtly alter the quality of care a person receives. The question is not whether doctors are biased, but where in the complex journey of medical care these biases take root, and whether they fade away as a doctor gains more experience.
To find the answer, researchers in Spain set up a unique experiment that looked like a medical school exam but was designed to catch something invisible. They gathered nearly two thousand medical students and junior doctors from four universities and placed them in a controlled environment where they had to treat simulated patients. These were not real people in distress, but actors trained to play specific roles, presenting with one of thirty-three different medical conditions ranging from infections to injuries. The students did not know who they would be treating until the moment the simulation began. They were assigned to a case simply by the schedule, meaning they had no choice in whether their patient was a man or a woman. This setup created a rare, fair test: if the doctors treated the patients differently, it could not be blamed on the patient's behavior or the doctor's prior knowledge, but rather on the sex of the person they were seeing.
The researchers watched these interactions closely, breaking down the medical encounter into seven distinct steps. They looked at how the doctors assessed symptoms, formed an initial guess about the illness, monitored the patient, ordered tests, confirmed the final diagnosis, administered treatment, and communicated with the patient. The results revealed a clear pattern of inequality that appeared right at the start of the process and then largely disappeared. When the simulated patient was a woman, the doctors scored significantly lower on the very first tasks. In assessing symptoms, the quality of care dropped by nearly fifteen percent compared to when the patient was a man. In forming an initial diagnosis, the score fell by eighteen percent. The way the doctors communicated with the female patients also suffered, with a fifteen percent drop in quality.
However, this gap did not last. As the simulation moved into more technical and rule-based stages, the difference vanished. When it came time to order diagnostic tests, confirm the main diagnosis, or administer treatment, the doctors performed just as well for women as they did for men. The bias was concentrated entirely in the early, interpretative phases of care, where the doctor must listen to a story and make sense of vague clues. It seems that once the process shifted to following established protocols and checking objective data, the unconscious bias receded. The study also found that this pattern held true regardless of how much training the doctor had. Whether the participant was a student in their third year of medical school or a resident doctor with more experience, the initial gap in care for female patients remained. Experience did not seem to wash away the bias; if anything, in some areas like communication, the disparity persisted or even grew slightly larger for the more advanced trainees.
The study suggests that the root of the problem lies in the human side of medicine, specifically in how doctors listen and speak before they begin to act. The early stages of a medical visit rely heavily on narrative and subjective judgment, areas where hidden stereotypes can easily slip in. The later stages, which depend on checklists and hard numbers, act as a safeguard, forcing the doctor to follow a standard path that leaves less room for bias to interfere. The researchers did not find that the doctors were less skilled overall, nor did they find that the bias was limited to specific diseases like heart conditions or mental health issues. Instead, the disparity appeared across a wide variety of medical problems, suggesting a structural issue in how clinical reasoning is applied to women.
This work does not claim to have solved the problem of bias in healthcare, but it has pinpointed exactly where it happens. By showing that the inequality is most severe in the first moments of a consultation, the study highlights a critical window for change. It suggests that training doctors to recognize these early interpretative gaps and to communicate more effectively with female patients could improve the quality of care long before a diagnosis is ever written down. The findings confirm that even in a controlled setting where doctors are trying their best, the sex of the patient can still quietly influence the quality of the care they receive, particularly when the doctor is asked to listen and understand rather than simply follow a rule.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.