The MRCEM SBA questions that cost you marks are rarely the ones where you did not know the condition. They are the ones where you knew it perfectly well, looked at the options, and found two that you would genuinely do for that patient. Both are defensible. Both are things you have done in a real resus bay. The exam wants one, and it wants it because of sequence rather than because the other is wrong. Learning to discriminate between two reasonable actions is a specific skill, it is separable from clinical knowledge, and it can be trained.
Key takeaways
- The best distractors in emergency medicine are not wrong actions, they are correctly timed actions offered at the wrong moment.
- Anchor every decision on the immediate physiological threat: what is going to kill this patient first?
- Stabilisation precedes diagnosis, and diagnosis precedes definitive treatment, unless the definitive treatment is the stabilisation.
- Time-critical interventions come before confirmatory investigations, even when the investigation is entirely appropriate.
- Age, pregnancy and comorbidity are not background detail: they are frequently the discriminator the whole question turns on.
Ask what kills first
When two options both look right, stop evaluating the options and go back to the patient. Ask a single question: what is the thing most likely to kill this person in the next few minutes?
That question resolves a surprising proportion of these items, because emergency medicine has a hierarchy and the exam respects it. An airway problem outranks a breathing problem, which outranks a circulatory problem, which outranks the diagnosis. A patient with a tension pneumothorax and a fractured femur does not get analgesia first, however humane that instinct is, because one of those things will kill them within minutes and the other will not.
So before you choose, name the immediate threat explicitly. If your chosen option does not address it, you have almost certainly picked the wrong one, however clinically sound it is.
Stabilisation, then diagnosis, then definitive care
Most emergency questions sit somewhere on a sequence, and the answer options are usually scattered along it.
Stabilisation buys time: oxygen, access, fluids, decompression, adrenaline, glucose, pacing. It does not require you to know the final diagnosis.
Diagnosis narrows the possibilities: the investigation, the imaging, the bedside test.
Definitive care resolves the underlying problem: the theatre, the thrombolysis, the antidote, the specific antibiotic.
The exam frequently offers you an option from each and asks which comes next. If the patient is unstable, the stabilisation option almost always wins, even when a candidate can see exactly what the diagnosis will turn out to be. Knowing the answer is not the same as being allowed to skip to it.
The important exception, and the one that catches thoughtful candidates, is when the definitive treatment is the stabilisation. Needle decompression of a tension pneumothorax is both. Adrenaline in anaphylaxis is both. Glucose in hypoglycaemia is both. When the specific treatment reverses the immediate physiological threat, it does not wait behind a confirmatory test.
Treatment before confirmation, when time is the constraint
This is the single most reliable pattern in the exam and it produces the most confident wrong answers.
A candidate reads a vignette of a patient with a clinically evident, time-critical condition, sees an investigation among the options that they would absolutely order in real life, and chooses it. It is not a wrong investigation. It is a wrongly sequenced one.
The test to apply is simple: would waiting for this result change what I do in the next ten minutes, and can this patient afford to wait for it? If the answer to the second question is no, then the investigation is not the next step, whatever its merits. Suspected meningococcal sepsis gets antibiotics before it gets a lumbar puncture. A tension pneumothorax is decompressed before it is imaged. Anaphylaxis is treated before tryptase is sent.
The modifiers are the question
The second great source of near-miss errors is treating the patient's particulars as scene-setting rather than as the point.
Age changes the answer. A drug dose, a fluid volume, a resuscitation algorithm and a differential diagnosis all move when the patient is three rather than thirty. Paediatric physiology compensates and then decompensates abruptly, which changes both the urgency and the interpretation of apparently reassuring observations.
Pregnancy changes the answer, and not only in obstetric presentations: it changes the differential, the imaging you can use, the drugs you can give, and the positioning of a shocked patient.
Comorbidity and current medication change the answer. The anticoagulated head injury, the beta-blocked patient who cannot mount a tachycardia, the immunosuppressed patient with a soft abdomen, and the renally impaired patient facing a standard drug dose are all questions where the comorbidity is the discriminator, and the candidate who read past it will choose the option that would be right for an otherwise well thirty-year-old.
When two options both look reasonable, go back and reread the patient's specifics. The detail that seems incidental is usually the one the item was written around.
Audit the option you nearly chose
The most valuable review habit for these items takes two minutes. For every question where you hesitated between two options, write down both, and then write the single sentence that separates them: not why the right one is right, but when the other one would have been right.
"Chest X-ray would have been correct if he were haemodynamically stable." "Fluid would have been correct before the adrenaline if this were septic rather than anaphylactic shock." "CT head first would have been correct if he were not anticoagulated."
That sentence is the transferable rule, and it is what the next question will test. The answer itself is worth one mark, once. The rule is worth marks repeatedly.
Where iatroX fits
iatroX's MRCEM SBA bank is built around emergency decision-making rather than isolated recall, with explanations grounded in current guidance so the reasoning behind the sequence sits alongside the answer rather than being left implicit. Missed questions can be opened in the Socratic Tutor, which asks what must happen first and why before it explains, which is precisely the reasoning step that mistimed answers skip. The adaptive engine then returns the same prioritisation principle in a different presentation, which tests whether the rule genuinely transferred rather than whether you remembered an item. Try it with free sample questions at iatroX. For preparing for the exam's structure, see the two-paper format.
Frequently asked questions
Why do I get MRCEM questions wrong when I knew the diagnosis? Because the best distractors are correct actions offered at the wrong point in the sequence. Knowing the condition is not the same as knowing what comes next, and the exam tests the latter.
How do I decide between two reasonable emergency treatments? Identify the immediate physiological threat and ask which option addresses it. Stabilisation precedes diagnosis, and diagnosis precedes definitive care, unless the definitive treatment is itself the stabilisation, as with adrenaline in anaphylaxis.
When should treatment come before investigation? When the diagnosis is clinically evident and time-critical, and waiting for the result would not change the next ten minutes or the patient cannot afford the delay. The investigation may be entirely appropriate and still be the wrong next step.
What details in the stem most often change the answer? Age, pregnancy, comorbidity and current medication. These are frequently the discriminator the item was written around, and candidates who read them as background will pick the option that would suit an otherwise well adult.