Antimicrobial resistance and the AWaRe logic

At the end of a long shift, a doctor sees a thirty-eight-year-old woman with a sore throat, runny nose and cough for three days. The patient has no high fever, no shortness of breath and a normal lung exam. She says she has a trip planned and asks whether it would not be better to take an antibiotic, just to be safe. The doctor knows the picture suggests a viral infection, but the queue is long, the patient insists and the prescription would take a minute. The scene is fictional and was written for teaching. This text uses that moment as a clinical reasoning scenario: what goes into the decision to prescribe an antibiotic or not, how the World Health Organization's AWaRe classification organizes that choice and why "just in case" has a cost.

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

This text is not a treatment guide. The choice of each antibiotic follows local guidelines, the region's resistance pattern and the judgment of the clinician. The interest here is the reasoning: how the decision to prescribe can be thought of as a hypothesis that is tested and revised, not as an automatic gesture. If you play the case of the day in NEXO, you will recognize the logic: each new clue calls for a revision of the previous guess.

Key points

• Antimicrobial resistance is a global health problem, with a greater burden in places with fewer resources.
• A significant share of antibiotic prescriptions in outpatient care has no clear indication.
• The AWaRe classification divides antibiotics into three groups, Access, Watch and Reserve, and guides the use of the narrowest spectrum that solves the problem.
• Shorter courses work as well as long ones in several common infections, and each extra day adds risk of adverse effects.
• The decision to prescribe can be revised when cultures and the clinical course arrive, and that revision is part of the reasoning.

Why antibiotic prescribing is a reasoning scenario

Antibiotic prescribing looks like a simple decision, but it brings several questions together in one. Is there a bacterial infection? Where in the body? What is the most likely microorganism? What is the risk if I do not treat now? What is the risk if I do treat? Each of these questions carries uncertainty, and the final answer is usually reached in a few minutes.

A decision with two patients

Every antibiotic prescription has one effect on the patient in front of you and another on the community. Laxminarayan and colleagues describe antimicrobial resistance as a global problem that calls for solutions built jointly by countries (Laxminarayan and colleagues, 2013). Today's patient receives the direct benefit or harm, and the effect on resistance spreads to patients the doctor will never see.

What the scene teaches

In the opening scene, the doctor has a reasonable hypothesis, a viral infection, and pressure to act. The antibiotic would not change the course of a viral infection, but it would bring risks for the patient and contribute to resistance. The post on clinical decision-making shows how to weigh the benefits and risks of an approach before choosing.

The size of the problem

The collaboration on antimicrobial resistance published in the Lancet estimated that, in 2019, 4.95 million deaths were associated with resistant bacteria and 1.27 million were directly attributable to resistance (Antimicrobial Resistance Collaborators, 2022). The same study found the greatest burden in regions with fewer resources and identified lower respiratory infections as responsible for more than 1.5 million of the associated deaths.

Prescriptions without indication

Fleming-Dutra and colleagues analyzed outpatient visits in the United States and estimated 506 antibiotic prescriptions per 1000 inhabitants per year, of which 353 would be appropriate (Fleming-Dutra and colleagues, 2016). In acute respiratory conditions, the difference was greater: 221 prescriptions per 1000 inhabitants per year, against 111 considered appropriate, according to the same study.

The harm to the patient

The antibiotic also harms the person who receives it. Tamma and colleagues followed hospitalized patients and found at least one adverse event in 298 of 1488 patients who received antibiotics (Tamma and colleagues, 2017). In the same study, each additional 10 days of antibiotics increased the risk of an adverse event by 3 in every 100, and 56 adverse events occurred in patients who received antibiotics without a clinical indication.

What the AWaRe classification is

The World Health Organization created the AWaRe classification in 2017, through the expert committee of the essential medicines list, and updates it every two years (WHO, AWaRe classification 2023). Sharland and colleagues presented the proposal to group antibiotics into three categories to guide use and surveillance (Sharland and colleagues, 2018).

The three categories

The name comes from the English initials of three groups. The Access group brings together first-choice antibiotics for common infections, generally with a narrower spectrum. The Watch group brings together antibiotics with a greater potential to select for resistance, which require more careful use. The Reserve group brings together last-resort antibiotics, kept for infections by multidrug-resistant microorganisms.

A target for the whole

The same WHO publication proposes that at least 60 in every 100 antibiotics consumed in a country belong to the Access group (WHO, AWaRe classification 2023). The target does not say what to give to a specific patient, but it provides a reference for looking at the prescribing pattern of a service or a country.

The antibiotic book

In 2022, the WHO published the AWaRe antibiotic book, with guidance on choosing the antibiotic, the dose and the duration for common infections, and on situations in which the antibiotic is not necessary. Sharland and colleagues presented the book as guidance for appropriate antibiotic use and for policy formulation (Sharland and colleagues, 2022).

The table below summarizes, in a didactic way, the logic of each group. It does not replace the official list or local guidelines.

Group	Role in reasoning	Question that helps
Access	First choice for common infections, generally with a narrower spectrum	Does an antibiotic from this group solve the likely problem?
Watch	Greater potential to select for resistance, for more specific indications	What justifies moving away from the Access group in this patient?
Reserve	Last resort, for confirmed or highly likely multidrug-resistant microorganisms	Is there culture data or context that makes this necessary?

What consumption data show

Klein and colleagues analyzed antibiotic sales data from several countries and found a 90.9 in every 100 increase in consumption of the Watch group between 2000 and 2015, against 26.2 in every 100 in the Access group (Klein and colleagues, 2021). In the same study, the ratio between Access and Watch consumption fell from 2.6 to 1.6.

Fewer countries within the target

Klein and colleagues also report that the number of countries reaching the target of 60 in every 100 in the Access group went from 50 in 66 countries to 42 in 76 countries during the study period (Klein and colleagues, 2021). The trend shows that, overall, use has shifted toward broader-spectrum antibiotics.

What this means in the consultation

These numbers seem distant from the clinic, but they are the sum of individual decisions. Each time an antibiotic from the Watch group is chosen where one from the Access group would suffice, the service's pattern moves a little. The clinical reasoning that asks "what is the minimum spectrum that solves it" is the individual version of the population target.

The "just in case" bias

Linder and colleagues analyzed 21,867 visits for acute respiratory infection and found antibiotic prescriptions in 44 in every 100 of them (Linder and colleagues, 2014). In the same study, the chance of prescription increased over the course of the shift, with an odds ratio of 1.14 in the third hour and 1.26 in the fourth hour, compared with the first.

Decision fatigue

The authors interpret the finding as a sign of decision fatigue: when tired, the doctor tends to choose the option that seems easy and safe, and the antibiotic prescription fills that role (Linder et al., 2014). The article cites reasons such as fear of complications and the wish to do something for the patient. The opening scene takes place precisely at the end of a long shift.

Commission bias

The "just in case" is an example of commission bias, the tendency to prefer acting over waiting, even when the action brings no benefit. Croskerry describes this and other cognitive biases as frequent sources of diagnostic and management error, and proposes strategies to recognize them (Croskerry, 2003). The opposite bias, omission bias, appears when fear of causing harm leads to not acting when action would be indicated. Both exist, and reasoning needs to consider both sides.

Other forces at play

Beyond tiredness, several forces push toward prescribing:

• The patient's expectation, real or presumed, of leaving with a prescription.
• The short consultation time, which makes explaining more laborious than prescribing.
• Fear of a rare complication and of responsibility for it.
• Difficulty with follow-up, when the patient cannot easily return.
• The feeling that the antibiotic is harmless.

The post on examples of heuristics in clinical reasoning shows how mental shortcuts help most of the time and fail in predictable situations.

Minimum effective spectrum

The logic of AWaRe starts from a simple question: what is the narrowest-spectrum antibiotic that treats the probable infection? Answering requires having a hypothesis about the site of infection and the most likely microorganism, which makes prescribing an exercise in diagnosis.

The hypothesis comes before the prescription

When the doctor does not know where the infection is, the tendency is to choose a broad-spectrum antibiotic to cover all possibilities. That choice may be necessary in a severely ill patient, but in someone who is stable it usually reflects a hypothesis that is still vague. The post on differential diagnosis shows how to organize hypotheses before acting on them.

What probability changes

In upper respiratory infections, the probability of a bacterial cause is usually low, and so is the probability of benefit from an antibiotic. The post on pretest probability explains how the expected frequency of a condition changes the value of each clue. In a picture like the one in the scene, the question is not only "could it be bacterial?" but "what is the chance, and what changes if I wait?".

Delayed prescribing

Spurling et al. gathered trials on delayed prescribing, in which the patient receives the prescription with guidance to use it only if symptoms worsen or do not improve. Antibiotic use was 31 in every 100 patients with delayed prescribing, versus 93 in every 100 with immediate prescribing, with similar satisfaction (Spurling et al., 2017). The strategy is a way of handling uncertainty without ignoring the patient's concern.

Short duration

Antibiotic duration is another decision in which habit weighs more than data. Spellberg and Rice report that more than 45 randomized clinical trials found that short courses are not inferior to long ones in several common infections (Spellberg and Rice, 2019). The authors cite, for example, that 3 to 5 days of treatment are as effective as 7 to 14 days in community-acquired pneumonia.

Where the week comes from

Spellberg and Rice recall that the seven-day week, used as a reference for so many antibiotic courses, has a historical origin in the Roman emperor Constantine, not in studies of infections (Spellberg and Rice, 2019). The authors comment on the study by Vaughn et al., conducted with patients hospitalized for pneumonia, in which each day of antibiotic beyond what was needed increased by 5 in every 100, in relative terms, the chance of adverse effects reported by patients after discharge (Vaughn et al., 2019).

Stop when the patient improves

Uranga and colleagues randomized 312 inpatients with community-acquired pneumonia. In the intervention group, the antibiotic was stopped after a minimum of 5 days when the patient met clinical stability criteria, and the outcome was non-inferior to usual treatment (Uranga and colleagues, 2016). The study shows that duration can be a decision based on the patient's course, not a fixed number.

Reviewing the approach: the antibiotic as a hypothesis

An empirical prescription is a bet made with the information available at that moment. The CDC guide for hospital antibiotic stewardship programs describes the antibiotic pause as a reassessment made by the team itself, when the clinical picture is clearer and there is more diagnostic information, especially cultures and rapid tests, and cites a study in which a pause between 48 and 72 hours of treatment improved the choice of antibiotic (CDC, 2019). The same guide notes that the ideal timing of this review has not yet been defined. The NICE guideline on antimicrobial stewardship, focused on intravenous prescriptions, also proposes reviewing them between 48 and 72 hours, looking at the response to treatment and the microbiology results (NICE, 2015). Adjusting the spectrum, switching the route of administration or stopping the antibiotic are decisions as important as the initial choice.

The antibiotic pause

Thom and colleagues studied a structured pause to review the antibiotic in 3470 treatment courses in hospitals. The pause did not change total antibiotic use, but it reduced inadequate therapy from 45 to 31 per 100 courses (Thom and colleagues, 2019). The authors conclude that the pause alone, without the support of an antimicrobial stewardship team, is not enough.

Programs that work

Davey and colleagues gathered studies on interventions to improve antibiotic prescribing in inpatients. With the interventions, treatment in line with local policy rose from 43 to 58 per 100 patients, and duration fell by 1.95 days, from a mean of 11.0 days (Davey and colleagues, 2017). In the same study, mortality was similar in both groups, about 11 per 100, and length of stay fell by 1.12 days.

A pause in the reasoning

Reviewing the antibiotic has the same structure as the diagnostic pause. The post on the diagnostic pause describes how a short moment of checking protects against premature closure. In prescribing, the questions are similar:

1. Was the initial hypothesis of bacterial infection confirmed?
2. Do the cultures allow narrowing the spectrum?
3. Can the patient switch to the oral route?
4. What is the minimum adequate duration for this condition?
5. Is the antibiotic still necessary?

The conversation with the patient

The pressure to prescribe often comes from a difficult conversation. Explaining why the antibiotic does not help in a viral infection takes longer than writing the prescription, but it is part of care. The explanation works better when it acknowledges the patient's discomfort, says what to expect in the coming days and makes clear when to come back.

In the opening scene, the doctor could tell the patient that the picture looks like a viral infection, that symptoms usually improve within a few days and that the antibiotic would not shorten that path. She could also offer measures to relieve the sore throat and cough, agree on a timeframe for reassessment and explain which changes would justify coming back before the trip. The patient leaves with an explanation and a plan, and the decision remains open to review if the picture changes. This conversation turns the "no" into an approach.

Warning signs

A well-conducted conversation includes the signs that call for reassessment: fever that persists or returns, shortness of breath, worsening after an initial improvement, pain that intensifies. The patient leaves with a plan, not with the feeling of having been dismissed. The post on diagnostic uncertainty shows how to make reasonable decisions when the information is still incomplete.

Calibrating confidence

The fear of making a mistake by not prescribing is usually greater than the real risk. Knowing how much to trust your own hypothesis, no more and no less, is a skill you can train. The post on calibration of confidence in diagnosis explains how to compare the certainty we feel with how often we are right.

How NEXO helps train this reasoning

In NEXO, the case of the day is fictional, with a real clinical presentation and up to six progressive clues. You propose a diagnosis by looking for the code in ICD-10 and receive proximity feedback at five levels, which shows whether the guess landed on the exact code, the same category, the same group, the same chapter or a different chapter. This format connects with the theme of this text: placing the infection in the right chapter and category is the first step to choosing the right spectrum.

To use the case of the day as training, try three habits. Before each guess, ask where the problem is and what the most likely cause is. When a new clue appears, ask whether it changes the hypothesis or only confirms it. After the result, note which clue would have allowed you to narrow the hypothesis earlier.

Three exercises with cases

Locate before treating. Faced with a case with fever, write in one line which organ you think is involved before thinking about management.

Review on the next clue. With each new clue, ask whether the hypothesis of bacterial infection has become stronger or weaker.

Name the pressure. When you feel the urge to act just to be safe, write down what fear is behind that urge.

The post on case-based learning in clinical practice shows how cases help fix decision patterns.

The limits on using the game are described on the NEXO economy page. If you research the teaching of clinical reasoning, the page for researchers explains how to get in touch. Professionals who write or review cases find information on curation and creators, and institutions can learn about the proposal on the partnerships page.

Frequently asked questions

What is antimicrobial resistance?

It is the ability of bacteria and other microorganisms to survive medicines that used to eliminate them. The use of antibiotics, whether necessary or not, favors the selection of resistant microorganisms.

What does AWaRe mean?

It is the World Health Organization classification that divides antibiotics into three groups, Access, Watch and Reserve, according to spectrum and potential to select for resistance. It helps guide the choice and monitoring of consumption.

Do antibiotics help with a cold?

Colds and most upper respiratory infections are caused by viruses, and antibiotics act on bacteria. In these cases, treatment aims to relieve symptoms, with guidance on signs that call for reassessment.

What is a delayed prescription?

It is a strategy in which the patient receives the prescription with guidance to use it only if symptoms worsen or do not improve within an agreed period. Studies show lower antibiotic use with similar satisfaction.

Do short antibiotic courses work?

In several common infections, clinical trials have shown that shorter courses work as well as long ones. The appropriate duration depends on the infection and the patient's progress, and follows local guidelines.

Why review the antibiotic after a few days?

Because the initial prescription is made with incomplete information. Antibiotic stewardship guidelines propose a review between 48 and 72 hours of treatment. When cultures arrive and the course becomes clearer, it is possible to narrow the spectrum, change the route or stop treatment.

Final checklist

• I formed a hypothesis about the site of the infection and the most likely microorganism.
• I estimated the chance of a bacterial cause before deciding to prescribe.
• I considered the narrowest-spectrum antibiotic that solves the problem.
• I thought about the minimum appropriate duration, not just the usual number of days.
• I planned to review management when cultures and clinical progress arrive.
• I explained to the patient what to expect and when to return.

Want to practice reasoning with progressive clues? Play the case of the day.

References

1. Antimicrobial Resistance Collaborators. Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis. Lancet. 2022. PMID 35065702.
2. Centers for Disease Control and Prevention. Core elements of hospital antibiotic stewardship programs. 2019. Publication page.
3. Croskerry P. The importance of cognitive errors in diagnosis and strategies to minimize them. Acad Med. 2003. PMID 12915363.
4. Davey P, Marwick CA, Scott CL, et al. Interventions to improve antibiotic prescribing practices for hospital inpatients. Cochrane Database Syst Rev. 2017. PMID 28178770.
5. Fleming-Dutra KE, Hersh AL, Shapiro DJ, et al. Prevalence of inappropriate antibiotic prescriptions among US ambulatory care visits, 2010-2011. JAMA. 2016. PMID 27139059.
6. Klein EY, Milkowska-Shibata M, Tseng KK, et al. Assessment of WHO antibiotic consumption and access targets in 76 countries, 2000-15: an analysis of pharmaceutical sales data. Lancet Infect Dis. 2021. PMID 32717205.
7. Laxminarayan R, Duse A, Wattal C, et al. Antibiotic resistance: the need for global solutions. Lancet Infect Dis. 2013. PMID 24252483.
8. Linder JA, Doctor JN, Friedberg MW, et al. Time of day and the decision to prescribe antibiotics. JAMA Intern Med. 2014. PMID 25286067.
9. National Institute for Health and Care Excellence. Antimicrobial stewardship: systems and processes for effective antimicrobial medicine use (NG15). 2015. Recommendations.
10. Sharland M, Pulcini C, Harbarth S, et al. Classifying antibiotics in the WHO Essential Medicines List for optimal use: be AWaRe. Lancet Infect Dis. 2018. PMID 29303731.
11. Sharland M, Cappello B, Ombajo LA, et al. The WHO AWaRe Antibiotic Book: providing guidance on optimal use and informing policy. Lancet Infect Dis. 2022. PMID 36309019.
12. Spellberg B, Rice LB. Duration of antibiotic therapy: shorter is better. Ann Intern Med. 2019. PMID 31284302.
13. Spurling GK, Del Mar CB, Dooley L, et al. Delayed antibiotic prescriptions for respiratory infections. Cochrane Database Syst Rev. 2017. PMID 28881007.
14. Tamma PD, Avdic E, Li DX, Dzintars K, Cosgrove SE. Association of adverse events with antibiotic use in hospitalized patients. JAMA Intern Med. 2017. PMID 28604925.
15. Thom KA, Tamma PD, Harris AD, et al. Impact of a prescriber-driven antibiotic time-out on antibiotic use in hospitalized patients. Clin Infect Dis. 2019. PMID 30517592.
16. Uranga A, España PP, Bilbao A, et al. Duration of antibiotic treatment in community-acquired pneumonia: a multicenter randomized clinical trial. JAMA Intern Med. 2016. PMID 27455166.
17. Vaughn VM, Flanders SA, Snyder A, et al. Excess antibiotic treatment duration and adverse events in patients hospitalized with pneumonia: a multihospital cohort study. Ann Intern Med. 2019. PMID 31284301.
18. World Health Organization. AWaRe classification of antibiotics for evaluation and monitoring of use, 2023. 2023. Publication page.
19. World Health Organization. The WHO AWaRe (Access, Watch, Reserve) antibiotic book. 2022. Publication page.