Affective bias: when emotion enters diagnosis

Third visit by the same patient to the emergency department in a week. He arrives irritated, complains about the wait, interrupts the questions and says that nobody solves anything. The resident feels his own face get hot, does the exam faster than he would with another patient and writes "complaint already investigated" before finishing listening. Only later, at the shift handover, does someone notice a new piece of data that had been there from the start.

NEXO is a daily medical diagnosis game: one clinical case a day, six progressive clues, and ICD-10 proximity feedback across five levels.

Classic heuristics explain this scene poorly. What weighed was emotion: the patient's and, above all, the doctor's. This guide explains what affective bias is, what experiments show about patients considered difficult and about the mood of the person making the diagnosis, and how to notice emotion before it decides in place of reasoning.

Key points

• Affective bias is the influence of the professional's emotions, about themselves and about the patient, on clinical decisions.
• A broad review pointed to sources such as patient characteristics, the work environment, the time of day and the professional's own state.
• In experiments with vignettes, patients with difficult behavior reduced the diagnostic accuracy of residents, with the same time spent on the case.
• The most accepted explanation is attention: part of the mind is occupied with the behavior, and the clinical findings are processed less.
• Positive emotions also change reasoning, and a recent study found an effect on emotions without an effect on diagnosis.

What affective bias is

Emotion as a source of error

A literature review from several fields concluded that the emotional state of the health professional can influence their decision, lead to affective biases and result in errors and adverse events (Croskerry et al., 2010). The authors listed various sources for this state: patient characteristics, conditions of the work environment, variations in time of day, cycle and season, as well as problems of the professional themselves (Croskerry et al., 2010).

The same review points out that clinical reasoning and judgment can be especially sensitive to emotion when they depend on intuitive judgments (Croskerry et al., 2010). It is the fast mode, the one that recognizes patterns, that is most easily carried away.

Different from classic heuristics

Availability, representativeness, anchoring and framing are shortcuts about information, described in the guide to examples of heuristics in clinical reasoning. Affective bias works another way: it changes how much attention the doctor pays, how much they ask and how willing they are to reconsider. It often pushes reasoning toward a shortcut, such as anchoring on an explanation that ends the conversation earlier.

Close neighbors

Affective bias touches on other topics in this series. Stereotypes about groups of patients, implicit bias, also change attention, but through another mechanism. The diagnostic pause is one of the institutional ways to interrupt hurried reasoning. And judging an error after the outcome is known suffers from hindsight bias. Each has its own guide.

What the experiments show

Difficult patients and accuracy

In a randomized experiment, 63 family medicine residents assessed six vignettes in which the same patient appeared with difficult or neutral behavior (Schmidt et al., 2017). Mean diagnostic accuracy, on a scale from 0 to 1, was 0.54 with difficult patients and 0.64 with neutral patients (Schmidt et al., 2017).

The time to reach the diagnosis was similar in both versions, and difficult patients received lower sympathy ratings (Schmidt et al., 2017). Deliberate reflection on the case improved the initial diagnosis, from 0.60 to 0.68, whatever the patient's behavior (Schmidt et al., 2017).

What stays in memory

A second experiment, with 74 internal medicine residents, used eight vignettes identical except for the patient's behavior (Mamede et al., 2017). Mean accuracy was 0.41 in the difficult versions and 0.51 in the neutral ones, with similar diagnostic time (Mamede et al., 2017).

After each case, residents remembered fewer clinical findings and more behaviors of difficult patients: 29.82% versus 32.52% for findings and 25.51% versus 17.89% for behaviors (Mamede et al., 2017). The authors concluded that part of the mental resources goes to dealing with the behavior, and clinical findings are processed insufficiently (Mamede et al., 2017).

Emotions reported in the emergency department

In a mixed-methods study, 94 experienced emergency professionals, 50 physicians and 44 nurses, described recent encounters that stirred anger or satisfaction or involved a patient with a mental disorder (Isbell et al., 2020). The professionals reported that emotions influenced their decisions more often in anger encounters, and they associated the emotions of those encounters with more perceived risk to patient safety (Isbell et al., 2020).

A result that calls for caution

In a more recent experiment, 134 experienced emergency physicians saw four computer-simulated patients, calm or irritated (Isbell et al., 2026). Faced with the irritated ones, the physicians reported more anger, anxiety and fatigue and judged those patients as worse informants and less cooperative, but the patients' behavior left clinical decisions, tests ordered and diagnosis intact (Isbell et al., 2026).

The contrast with the 2017 studies is useful. The physician's emotion changes consistently; the effect on diagnosis depends on the study format, the type of patient and perhaps the experience of the person providing care. The authors of the 2026 study note that, even without an effect on diagnosis, the reaction can worsen the relationship with the patient (Isbell et al., 2026).

Why emotion gets in the way of reasoning

The negative affect hypothesis

A synthesis of 43 experimental studies on context in diagnostic reasoning proposed two hypotheses (Schmidt et al., 2024). Under the negative affect hypothesis, error arises when the physician's attention shifts from relevant clinical findings to the irrelevant source of the negative affect, such as a patient's aggressiveness (Schmidt et al., 2024).

The second hypothesis, that of the primacy of the initial diagnosis, attributes errors to the disproportionate weight of the first hypothesis over the gathering and interpretation of the rest of the case (Schmidt et al., 2024). The two can add up: irritation speeds up the first hypothesis, and the first hypothesis then filters the rest.

Time is far from explaining everything

It would be intuitive to think that the irritated physician errs because he spends less time. The two 2017 experiments showed similar time in the difficult and neutral versions (Schmidt et al., 2017; Mamede et al., 2017). The difference lies in the quality of attention within the same time, and that makes the bias harder to notice.

Fatigue amplifies

The 2010 review lists time of day, long shifts and the professional's own state among the sources of emotional influence (Croskerry et al., 2010). An intern at the end of a 24-hour shift has less room to notice his own irritation.

When emotion is positive

A good mood also changes reasoning

In a classic experiment, third-year students received a positive mood induction before choosing, among six hypothetical patients, the one with the highest probability of a diagnosis (Isen et al., 1991). The groups were correct at the same rate, but the positive-mood students reached the choice sooner, went beyond the task asked and integrated the information better, with fewer signs of confusion (Isen et al., 1991).

Liking as a risk

Positive emotion toward the patient as a person is another matter. Liking someone very much can lead to avoiding an uncomfortable examination, interpreting a symptom as benign or postponing a difficult question. Emergency professionals associated positive encounters with higher-quality care (Isbell et al., 2020), but that is a perception, and affective bias also acts when the perception is good.

How to notice affective bias

Name the emotion

The 2010 review recommends raising awareness of these factors, including emotional skills training in medical education and making room in teaching to talk about feelings toward patients (Croskerry et al., 2010). In practice, the first step is simple: notice and name it to yourself. "I am irritated with this patient." Naming the emotion is not the opposite of empathy: it is what allows you to keep welcoming the patient without letting the feeling decide the diagnosis.

Separate the behavior from the finding

Before closing the case, list only the clinical findings, without the adjectives about the patient. If the list changes the impression, the behavior was taking up too much space. The exercise connects with metacognition in medicine.

Ask one more question

With patients who irritate, the tendency is to end the conversation. One extra open question ("is there anything else that has changed since last time?") is a simple way to make up for the attention that went to the behavior.

Reflect deliberately

In the 2017 experiments, deliberate reflection improved the initial diagnosis in all versions (Schmidt et al., 2017). Reviewing the hypotheses of the differential diagnosis, asking what supports and what contradicts each one, is the everyday version of that reflection.

Ask for a second look

When the emotion is strong, telling the case to a colleague who has never seen the patient helps: they receive the data without the emotional charge. The guide to feedback in clinical cases covers how to ask for and give that feedback, and the guide to calibrating confidence in diagnosis shows why trusting the first impression too much is risky.

Take care of your own state

Hunger, sleep and overload change your mood. During the internship, the guide on how to study during the internship with little time also talks about breaks and rest.

Train reasoning without the pressure of the consultation on NEXO

On NEXO, the case of the day arrives in up to six clues, with no patient in the room, no queue and no irritation. It is an environment to train the part of reasoning that looks only at the findings: which clues support the hypothesis and which contradict it. With each guess, the proximity in ICD-10 across five levels shows how far the hypothesis was.

The cases are fictional with a real clinical presentation, and the guide what NEXO is explains the whole game. How the modes are unlocked is on the NEXO economy page.

If you coordinate a medical school, a student league or a residency and want to discuss biases with your class, see NEXO Closer. Researchers in medical education can propose studies through NEXO Researcher. Doctors who review cases take part through NEXO Curator, and those who create health content on social media can take part through NEXO Creators.

Frequently asked questions

What is affective bias?

It is the influence of the professional's emotions, about themselves and about the patient, on clinical decisions (Croskerry et al., 2010).

Does a difficult patient lead to more diagnostic error?

In experiments with vignettes, yes: accuracy was lower in the versions with difficult behavior (Mamede et al., 2017). A recent study with computer simulation found an effect on emotions, with no effect on diagnosis (Isbell et al., 2026).

Why does emotion get in the way?

One hypothesis is that attention goes to the source of the emotion, and the clinical findings are processed less (Schmidt et al., 2024).

Is positive emotion also a bias?

It can be. A positive mood changes the way of reasoning, and liking a patient very much can lead to avoiding uncomfortable questions or exams.

How can affective bias be reduced?

By naming the emotion, separating behavior from finding, reflecting deliberately and asking for a second look.

Is affective bias the same as implicit bias?

Both change attention, but implicit bias starts from stereotypes about groups, and affective bias starts from the emotion of the moment. Each has its own guide in this series.

Affective bias checklist

1. Name the emotion the patient aroused in you.
2. List only the findings, without adjectives about the patient.
3. Ask one more open question before closing.
4. Review the hypotheses with what supports and what contradicts each one.
5. Tell the case to a colleague when the emotion is strong.
6. Notice your state: sleep, hunger and overload.
7. Be wary of sympathy as much as of irritation.

Want to train reasoning with the findings alone? Play the case of the day on NEXO.

References

1. Croskerry P, Abbass A, Wu AW. Emotional influences in patient safety. J Patient Saf. 2010. https://pubmed.ncbi.nlm.nih.gov/21500605/
2. Schmidt HG, Van Gog T, Schuit SC, Van den Berge K, Van Daele PL, Bueving H, et al. Do patients' disruptive behaviours influence the accuracy of a doctor's diagnosis? A randomised experiment. BMJ Qual Saf. 2017. https://pubmed.ncbi.nlm.nih.gov/26951795/
3. Mamede S, Van Gog T, Schuit SC, Van den Berge K, Van Daele PL, Bueving H, et al. Why patients' disruptive behaviours impair diagnostic reasoning: a randomised experiment. BMJ Qual Saf. 2017. https://pubmed.ncbi.nlm.nih.gov/26951796/
4. Isbell LM, Tager J, Beals K, Liu G. Emotionally evocative patients in the emergency department: a mixed methods investigation of providers' reported emotions and implications for patient safety. BMJ Qual Saf. 2020. https://pubmed.ncbi.nlm.nih.gov/31988259/
5. Isbell LM, Huff NR, Liu G, Bessen H, Kan V, Levesque I, et al. When emergency physicians meet patients displaying irritable behaviours: a randomised vignette-based experiment investigating physicians' emotions and clinical reasoning. BMJ Qual Saf. 2026. https://pubmed.ncbi.nlm.nih.gov/42437698/
6. Schmidt HG, Norman GR, Mamede S, Magzoub M. The influence of context on diagnostic reasoning: A narrative synthesis of experimental findings. J Eval Clin Pract. 2024. https://pubmed.ncbi.nlm.nih.gov/38818694/
7. Isen AM, Rosenzweig AS, Young MJ. The influence of positive affect on clinical problem solving. Med Decis Making. 1991. https://pubmed.ncbi.nlm.nih.gov/1881279/