Examples of heuristics in clinical reasoning

A patient is diagnosed with pharyngitis and goes home. The final diagnosis, which was lost along the way, was osteomyelitis. This case, published in a clinical medicine journal, shows how a series of mental shortcuts led to the missed diagnosis, and the author concludes that the solution is not to eliminate the shortcuts, which usually serve clinicians well, but to know the most common errors (Redelmeier, 2005).

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

Heuristics are these shortcuts. This text explains what they are, gives clinical examples of the best known ones, shows when each one helps and when it gets in the way, and proposes exercises to use them better.

Main points

• Heuristics are simple decision strategies that use only part of the available information, and they often work well.
• Representativeness, availability and anchoring are the three classic heuristics described in 1974, and all of them appear in daily clinical practice.
• Occam's razor and Hickam's dictum are opposing heuristics, and knowing when to swap one for the other is part of experience.
• Some simple decision rules work as well as more complex statistical methods.
• Structured reflection on the case helps correct the error of a heuristic when it happens.

Summary

• What heuristics are
• The three classic heuristics in clinical practice
• Occam's razor and Hickam's dictum
• Other common heuristics on shift
• When a heuristic helps and when it gets in the way
• Exercises to train heuristics
• Frequently asked questions
• Checklist for using heuristics safely

What heuristics are

Analyzing every patient from scratch, weighing all possible diseases, would take too long. With experience, the clinician starts to recognize patterns and use quick rules to decide. These rules are heuristics.

A definition

Heuristics are simple decision strategies that ignore part of the available information and rely on a few relevant data points (Marewski and Gigerenzer, 2012). The authors argue that, contrary to the intuition that more information is always better, heuristics can help doctors and patients make good decisions (Marewski and Gigerenzer, 2012).

A shortcut is not a synonym for error

The word "heuristic" often appears alongside "bias", as if the two were the same thing. They are not. Bias is the systematic error that a heuristic can produce in certain situations. The classic 1974 text describes heuristics as very economical and generally effective, but capable of leading to systematic and predictable errors (Tversky and Kahneman, 1974).

A review of heuristics in medicine reaches a similar conclusion: most of the time they are positive, but they can bring negative consequences when used inappropriately (Whelehan et al., 2020). The same authors note that personality, experience and context influence the use of heuristics (Whelehan et al., 2020).

Why they exist

Marewski and Gigerenzer point out among the qualities of heuristics their accuracy, transparency, broad access and the little time they require (Marewski and Gigerenzer, 2012). On a shift with many patients and little time, deciding with a few well-chosen data points is a necessity. The challenge is knowing which data to choose and recognizing when the shortcut leads to the wrong place. If you want to understand biases more deeply, the text on differential diagnosis with clues explains how they get in the way of the list of hypotheses.

How heuristics are formed

Heuristics are usually learned in practice, more than in class. In the study with nephrologists, the interviewees reported that clinical decision-making was learned implicitly, through exposure to cases, and not through explicit teaching (Ferrara et al., 2026). A curious finding of the same study is that residents, with their questions, broke the diagnostic routines of more experienced doctors in a way that reduced premature closure (Ferrara et al., 2026).

This has two consequences for those who are learning. The first is that every case you see feeds your heuristics, for better and for worse, and that is why the variety of cases matters. The second is that asking "why can't it be something else?" to the preceptor is not disrespect: it is a way to help everyone review the anchor.

The three classic heuristics in clinical practice

In 1974, Tversky and Kahneman described three heuristics used to judge in situations of uncertainty: representativeness, availability, and adjustment from an anchor (Tversky and Kahneman, 1974). All three appear in clinical reasoning every day.

Representativeness

Representativeness is used to judge the probability of something belonging to a category by how much it resembles that category (Tversky and Kahneman, 1974). In clinical practice, it is pattern recognition: the picture "looks like" the classic description of a disease.

Example where it helps: a patient with a typical picture of a common disease, with all the findings that describe it, probably has that disease. Recognizing the pattern saves time and usually gets it right.

Example where it gets in the way: when the presentation departs from the classic one, representativeness pushes away from the right diagnosis. A serious disease that presents atypically, for example in older adults, may be ruled out for "not looking like" the book. Another risk is forgetting frequency: a picture that resembles a rare disease may still more likely be a common disease that presented unusually.

Availability

Availability is used to estimate the frequency or plausibility of something by the ease with which examples come to mind (Tversky and Kahneman, 1974). In the case of the pharyngitis that hid an osteomyelitis, the author describes availability as judging probability by the ease with which examples come to memory (Redelmeier, 2005).

Example where it helps: during an epidemic in the region, thinking first of the disease that is circulating is reasonable, because its probability really has increased.

Example where it gets in the way: after seeing a striking case, the disease becomes "easy to remember" and starts to seem more likely than it is. In an experiment, residents read a Wikipedia entry about a disease and, hours later, made more diagnostic errors on cases that only superficially resembled it (Schmidt et al., 2014). Mean accuracy fell to 0.56 in the cases where the bias was expected, versus 0.70 in the others (Schmidt et al., 2014).

Anchoring

The third heuristic is adjustment from an anchor, an initial value or impression (Tversky and Kahneman, 1974). In clinical practice, the anchor is usually the first hypothesis, and anchoring is the tendency to hold on to it (Redelmeier, 2005).

Example where it helps: starting from a working hypothesis organizes the investigation. Without an anchor, data collection becomes scattered.

Example where it gets in the way: the diagnosis from triage or the previous shift is repeated without review, and new data are read to confirm the anchor, not to test it. Anchoring is among the shortcuts the author identified in the case of the pharyngitis that hid an osteomyelitis (Redelmeier, 2005).

Summary of the three

Heuristic	Question it answers	When it helps	When it gets in the way
Representativeness	Does this look like that disease?	Typical picture of a common disease	Atypical presentation, rare disease that "looks like" it but is unlikely
Availability	How easily do I remember cases like this?	Disease that is in fact more frequent at the moment	After a striking case or a news story
Anchoring	Where do I start from?	Organize the investigation with a working hypothesis	First impression never reviewed

Source of the table: heuristics described by Tversky and Kahneman, 1974; examples of use and error organized by the Editorial team from Redelmeier, 2005 and Schmidt et al., 2014.

Occam's razor and Hickam's dictum

Two classic heuristics of medicine say opposite things, and both are right part of the time.

The two rules

Occam's razor favors a single diagnosis that explains all the findings. Hickam's dictum reminds us that the patient can have several diseases at the same time. A qualitative study with nephrologists describes diagnostic reasoning as a navigation between these two competing heuristics (Ferrara et al., 2026).

How experienced clinicians alternate

In the interviews, the nephrologists said that, with experience, they began to apply parsimony to routine presentations and to think of several causes for patients with many diseases (Ferrara et al., 2026). They also reported that time and resource pressure pushes toward the simplest diagnosis, whatever the complexity of the case (Ferrara et al., 2026).

The study interviewed few physicians, from one specialty, and does not measure outcomes. But the authors' observation is useful for those who are learning: the ability to switch heuristics is usually learned implicitly, and they suggest teaching it explicitly (Ferrara et al., 2026).

A practical guide

Situation	Heuristic that usually works best	Why
Young person, without previous diseases, acute condition	Occam	It is more likely that a single cause explains everything
Older adult with several chronic diseases and many medications	Hickam	Several causes can coexist, including medication effects
Finding that does not fit the single hypothesis	Hickam	There may be a second hidden problem
Many findings from several systems at the same time	Occam, tested carefully	A systemic disease can unite the findings

Source of the table: teaching guidance from the Editorial based on the idea of alternating between parsimony and multiple causes according to the patient's complexity, described by Ferrara et al., 2026.

Other common heuristics on shift

Besides the classic ones, clinical daily life is full of short rules, many passed from preceptor to resident. Some have names; others are teaching phrases, such as those discussed in the text about what a clinical pearl is.

"Common is common"

The phrase about horses and zebras sums up the idea: before thinking about the rare diagnosis, think about the frequent one. It is a way to remember pretest probability, explained in the text about pretest probability. The limit is that, in referral services, zebras appear more often, and the frequency that matters is that of the place where you practice.

"First rule out what kills"

Faced with a symptom such as chest pain, the rule says to consider serious causes before benign ones, even if they are less likely. It works as a counterweight to the previous rule: reasoning considers probability and severity at the same time. The text about clinical decision-making shows how this balance enters the choice of tests.

Simple decision rules

Some heuristics are formal: a few questions, in order, that lead to a decision. Marewski and Gigerenzer discuss when heuristics like these outperform methods that use a lot of information, such as regressions in diagnosis (Marewski and Gigerenzer, 2012).

One example comes from a 100-bed rural hospital (Green and Mehr, 1997). Researchers tested a tool with probability tables to decide who went to the coronary care unit, but use of the unit had already fallen before the test, right after the tool was presented at a clinical meeting (Green and Mehr, 1997). Just by knowing the tool, the residents reached sensitivity and specificity equal to those of its best performance, even without doing the calculations, and the authors suggest that teaching simple decision strategies can be effective (Green and Mehr, 1997).

Confirmation

The review on heuristics in medicine points to confirmation heuristic among the three most common, alongside availability and anchoring (Whelehan et al., 2020). It is the tendency to seek and value data that confirm the hypothesis already chosen.

Example where it helps: after forming a hypothesis, seeking the findings it predicts is an efficient way to test it. If you think of a disease, it makes sense to examine what it usually changes.

Example where it hinders: when only supporting data are sought, contrary findings go unnoticed. A good counterweight is to always ask: what finding would I expect to find if my hypothesis were wrong, and did I look for it?

Framing

In the case of pharyngitis, the author also describes the framing effect, in which people decide differently depending on how the information is presented (Redelmeier, 2005). In clinical practice, framing appears in handovers and referrals: "patient with multiple complaints" or "patient with pain that does not improve" describe the same person and pull reasoning in different directions.

A simple exercise is to rewrite the case presentation with neutral words, using only the data, and see if your hypothesis changes.

Group and authority effects

The review on heuristics in medicine also cites herd effect, the tendency to follow what the group thinks (Whelehan et al., 2020). In the case of pharyngitis, the author also describes blind obedience, in which people stop thinking in the presence of an authority, and premature closure, in which alternatives are no longer investigated (Redelmeier, 2005).

When heuristics help and when they get in the way

The same heuristic can be right in one case and wrong in the next. What changes is the context.

Signs that it is worth slowing down

• The case does not fit. An important finding remains unexplained.
• You have just seen something similar. The recent memorable case may be pulling the answer.
• The diagnosis came ready-made. The hypothesis was inherited from triage, from a colleague or from the chart.
• The patient does not improve as expected. The course contradicts the hypothesis.
• The risk of being wrong is high. A serious and treatable disease is among the alternatives.

Experience changes use

In the study on availability bias with residents, the expected error pattern appeared in second-year residents, but not in first-year residents (Mamede et al., 2010). One possible reading is that, as pattern recognition becomes faster, it also becomes more subject to this type of shortcut. The review on heuristics also points to the level of experience as one of the factors that influence their use (Whelehan et al., 2020).

Reflection corrects

In both availability experiments, structured reflection on the case findings helped correct the error (Mamede et al., 2010; Schmidt et al., 2014). In the study with the Wikipedia entry, deliberate reflection brought accuracy back to the level before the bias, with a mean of 0.71 (Schmidt et al., 2014).

Reflecting, in this context, means reviewing the findings one by one: what favors the hypothesis, what contradicts it, what would be present if the hypothesis were different. It is slower, and therefore worth reserving for the moments in the list above.

Exercises to train heuristics

Knowing heuristics by name helps little if you do not recognize the moment when you are using one. The exercises below train that recognition.

Exercise 1: name the shortcut

After solving a case, ask: what heuristic did I use to reach the first hypothesis? Was it a pattern I recognized, a case I remembered, a hypothesis I inherited? Write down the name. Over time, you begin to notice the shortcut while it happens.

Exercise 2: Occam or Hickam?

Before closing a case, ask whether a single disease explains all the findings. If yes, test that hypothesis against each finding. If not, look for the second problem. Say out loud which of the two rules you are applying and why.

Exercise 3: vignettes to recognize heuristics

The vignettes below were written by the Editorial team to train recognition; they do not teach management.

1. The case of the week. Yesterday you saw a patient with a rare and memorable disease. Today another one arrives with some similar symptoms. Which heuristic is at play, and what question would you ask to test it?
2. The inherited diagnosis. A patient arrives transferred with the diagnosis of "anxiety attack" written in the referral. He is tachycardic and short of breath. Which heuristic can get in the way, and what would you review first?
3. The older adult with many problems. An older adult with several chronic diseases arrives confused. Does a single cause explain everything, or is it better to think of several? What data would help decide?

Comment: in the first vignette, availability can exaggerate the probability of the rare disease, and the useful question is what separates the two diseases. In the second, anchoring on the inherited diagnosis is the risk, and the review starts with vital signs and the serious causes of tachycardia and shortness of breath. In the third, Hickam's dictum reminds us that medications, infections and metabolic changes can coexist.

Train on NEXO

On NEXO, the case arrives in up to six clues, one at a time, and you give your guess by the diagnosis name or the ICD-10 code. With each wrong guess, the game shows the proximity in five levels, from different chapter to Exact ICD, and opens the next clue. This creates a good environment for observing your own heuristics: the first clue invites pattern recognition, and the following ones test whether you revise the anchor.

The Diary brings one case per day, the same for everyone, and the Archive gathers the previous daily cases to practice exercise 1 at volume. Continents changes the background epidemiology, which affects the rule of "common is common". Adrenaline adds time, and Multimedia brings images, audio and other exams. How each mode is unlocked is explained on the NEXO economy page.

Save, discuss and research

When you notice you fell into a shortcut, the Anki button creates the card straight from NEXO and opens it in the Anki app. The Group tab of the ranking, which gathers invited colleagues, lets you compare with your class on which clue each person got right and which heuristic led each one to the answer. If you coordinate a medical school, a league or a residency and want to use cases like these in teaching, check out NEXO Closer. Researchers interested in heuristics and clinical decision-making can propose studies through NEXO Researcher, and physicians who want to review cases in their specialty can take part through NEXO Curator. Those who produce medical content can check out NEXO Creators.

Frequently asked questions

What are heuristics in clinical reasoning?

They are simple decision strategies that use few relevant data and ignore part of the available information (Marewski and Gigerenzer, 2012). In daily clinical practice, they appear as pattern recognition, quick rules and teaching phrases.

What are the most common heuristics in medicine?

A review pointed to availability, anchoring and confirmation as the most common, followed by representativeness, overconfidence and herd effect (Whelehan et al., 2020).

Is heuristic the same thing as bias?

They are different things. The heuristic is the shortcut; the bias is the systematic error it can produce in certain situations. Heuristics are economical and generally effective, but they can lead to predictable errors (Tversky and Kahneman, 1974).

What is Occam's razor in medicine?

It is the rule of looking for a single diagnosis that explains all the findings. Hickam's dictum is the opposite: the patient can have several diseases. Experienced nephrologists reported using the first in routine cases and the second in complex patients (Ferrara et al., 2026).

Can you correct the error of a heuristic?

The experiments suggest so, with structured reflection on the case findings. In the study with residents exposed to an entry about a disease, deliberate reflection brought accuracy back to the level before the bias (Schmidt et al., 2014).

Do simple decision rules work?

In several situations, yes. In a rural hospital, residents who only knew a decision tool reached its ideal performance without doing the math (Green and Mehr, 1997).

How does NEXO help train heuristics?

On NEXO, each case arrives in up to six clues, and the proximity of the guess in ICD-10 shows whether your first hypothesis was close or far, which helps you notice when to revise the anchor. To learn about the game, read what NEXO is.

Checklist for using heuristics safely

1. Use the shortcut when the picture is typical and the risk of being wrong is low.
2. Name the heuristic that led you to the first hypothesis.
3. Remember the frequency before thinking about the rare.
4. Consider the serious cause even when it is less likely.
5. Be wary of the recent striking case.
6. Review the inherited diagnosis before repeating it.
7. Ask whether one disease explains everything or whether there is more than one.
8. Slow down when something does not fit and review the findings one by one.
9. Reflect on the errors and note the shortcut that failed.
10. Train with short cases to recognize the shortcuts as they happen.

Want to watch your heuristics in action today? Play the case of the day on NEXO.

References

1. Tversky A, Kahneman D. Judgment under Uncertainty: Heuristics and Biases. Science. 1974. https://pubmed.ncbi.nlm.nih.gov/17835457/
2. Marewski JN, Gigerenzer G. Heuristic decision making in medicine. Dialogues Clin Neurosci. 2012. https://pubmed.ncbi.nlm.nih.gov/22577307/
3. Redelmeier DA. Improving patient care. The cognitive psychology of missed diagnoses. Ann Intern Med. 2005. https://pubmed.ncbi.nlm.nih.gov/15657159/
4. Whelehan DF, Conlon KC, Ridgway PF. Medicine and heuristics: cognitive biases and medical decision-making. Ir J Med Sci. 2020. https://pubmed.ncbi.nlm.nih.gov/32409947/
5. Schmidt HG, Mamede S, van den Berge K, et al. Exposure to media information about a disease can cause doctors to misdiagnose similar-looking clinical cases. Acad Med. 2014. https://pubmed.ncbi.nlm.nih.gov/24362387/
6. Mamede S, van Gog T, van den Berge K, et al. Effect of availability bias and reflective reasoning on diagnostic accuracy among internal medicine residents. JAMA. 2010. https://pubmed.ncbi.nlm.nih.gov/20841533/
7. Ferrara A, Mason L, Ruberto P, D'Mello K, Kirpalani A. When the rules break down: how nephrologists navigate diagnostic complexity in real-world practice. J Nephrol. 2026. https://pubmed.ncbi.nlm.nih.gov/42544765/
8. Green L, Mehr DR. What alters physicians' decisions to admit to the coronary care unit? J Fam Pract. 1997. https://pubmed.ncbi.nlm.nih.gov/9300001/