How to read a clinical vignette: a method for the long, wordy questions

Here's a question that looks like a lot of work. A 45-year-old man presents with a three-day history of fever and a productive cough. He has a history of alcohol use and diabetes. On examination his temperature is 38.9 degrees, his respiratory rate is 26, and there is bronchial breathing in the right lower zone. A chest X-ray shows a right lower lobe consolidation. His blood tests show a raised white cell count. He is started on antibiotics. Which of the following is the most likely organism? And then four options.

You read the whole paragraph. You pull out the facts. You think. And a good minute later you pick an answer, which may or may not be right, and which cost you more time than a question that arrived in ten words.

A growing share of questions look like this: a short clinical story, a few investigations, and a question that is often not "what is the diagnosis". Reading them well is a skill separate from knowing the content, and it's one you can practise. This article is a method.

Why vignettes feel hard

A vignette gives you more information than you need. Most of it is either confirming the obvious or deliberately irrelevant, and part of the skill is telling which is which. The long format also tempts you to start reasoning before you know what's being asked, which means you may reason about the wrong thing.

There are really three separate skills hidden in the process: finding the question, recognising the pattern, and checking your answer against what was asked. Let's take them in order.

Step 1: Read the last line first

Before you read the story, read the question at the end. Is it asking for the diagnosis, the next step, the investigation of choice, the mechanism, the most likely complication, the drug of choice, or the organism?

This single habit changes how you read. If the question is "what is the next best step in management?", you're looking for features of instability or urgency. If it's "which organism?", you're looking for risk factors and the setting. If it's "which of the following is NOT...?", you've got a different task altogether.

It takes two seconds, and it stops you from reading the story as a puzzle with no direction.

Step 2: Read the story and mark the clues

Now read the vignette, looking for the facts that carry diagnostic weight. A useful way to sort them:

• Demographics and setting. Age, sex, occupation, geography, pregnancy. These shift probabilities.
• The key symptom or sign. The thing that brought the patient in.
• The tempo. Acute, subacute, chronic, progressive, episodic. Time course is among the strongest clues.
• Risk factors and history. Alcohol, diabetes, immunosuppression, drugs, family history.
• The positive finding that points somewhere. A specific sign, a lab value, an image finding.
• The negative that excludes something. "No history of trauma", "afebrile", "no rash".

If you have a rough sheet, write two or three words. "45M, fever, cough, alcohol + DM, RLL consolidation." That's your whole story in a line.

What to ignore

Don't treat every number as equally important. Vital signs matter when they point to instability. A raised white cell count in an infection is expected and doesn't narrow anything. A normal value is sometimes the clue. And some details are there to be distracting. Notice what changes your answer and what doesn't.

Step 3: Name the pattern before you look at the options

Try to name the diagnosis or answer before you read the choices. Doing so prevents the options from steering you. If you look at four plausible-sounding options first, each will seem reasonable, and you'll end up arguing with yourself.

In the example, the pattern is community-acquired lobar pneumonia in a patient with alcohol use and diabetes. The most common bacterial cause of community-acquired pneumonia is Streptococcus pneumoniae. But alcohol use and diabetes also raise the possibility of Klebsiella. Now you look at the options. If both are there, you need a tie-breaker in the stem, such as "currant jelly sputum", a thick, bloody, gelatinous sputum classically linked to Klebsiella. Without it, the safer answer is the most common cause.

Having formed your own answer, you can look at the options and see whether yours is among them. If it is, great. If it isn't, you've probably misread something, so reread.

Step 4: Answer exactly what was asked

This is where many marks are lost, even by people who got the diagnosis right. You identified the disease correctly, and the question wanted something else.

• "Most likely diagnosis" wants a diagnosis.
• "Investigation of choice" wants the best test, not the first one, and not every useful one.
• "Next best step" often wants the immediate action: stabilise, give a first-line drug, get an urgent scan. It's frequently not the final definitive treatment.
• "Most likely complication" wants the commonest or best-recognised one.
• "Mechanism of action" wants the molecular or physiological explanation.
• "Drug of choice" wants first-line therapy for that situation.

When a question has an emergency flavour, the answer to "what next?" is usually the thing that deals with the threat to life first. Airway, breathing, circulation. Fluids before insulin in DKA. Adrenaline in anaphylaxis. Thiamine with glucose in a suspected alcohol-related encephalopathy.

Step 5: Check the options against the stem

Now go through the options and test each against the facts, not against your feelings. For each option, ask: does anything in the stem contradict this? Is there a feature that fits this better than my answer?

Treat words like "always", "never", "only" and "all" with suspicion, since medicine has few absolutes. And watch for options that are true statements that don't answer the question. A fact can be correct and irrelevant.

Three worked examples

Example one: the next step

A 19-year-old with type 1 diabetes presents with vomiting, abdominal pain and rapid deep breathing. His breath smells fruity. Blood glucose is 450 mg/dL. After securing the airway and starting monitoring, what is the most important next step?

Last line first: next step. Pattern: diabetic ketoacidosis. What does the examiner want? The immediate, most important action. In DKA, that is fluid resuscitation with isotonic saline. Insulin follows once potassium is checked, and potassium is replaced as needed. The tempting wrong answers are the dramatic ones: a bicarbonate bolus, a big dose of insulin, a potassium bolus. They aren't first.

Example two: the mechanism

A child presents with a membranous grey-white film over the tonsils, a bull neck and a low-grade fever. Which of the following best describes the mechanism of action of the toxin responsible?

Last line first: mechanism of toxin. Pattern: diphtheria. The toxin ADP-ribosylates elongation factor 2, halting protein synthesis. The wrong options are the other toxins you know: cholera toxin (activates adenylate cyclase), tetanus and botulinum toxins (cleave SNARE proteins), and superantigens (toxic shock). Naming the pattern first made the right answer obvious.

Example three: the investigation

A 30-year-old woman has episodes of palpitations, sweating and headache with severe hypertension. A mass is seen in the adrenal gland. Which of the following is the best biochemical test?

Last line first: best biochemical test. Pattern: phaeochromocytoma. The classic triad of episodic headache, sweating and palpitations with hypertension points to catecholamine excess. The best biochemical test is plasma free metanephrines or 24-hour urinary fractionated metanephrines. The tempting distractors are tests for other adrenal conditions, such as an aldosterone-renin ratio or a dexamethasone suppression test, which belong to different stories.

In all three, the sequence is the same: ask, read, name, answer, check.

Traps to watch for

Anchoring on the first diagnosis. You decide early and stop reading. Then a later detail contradicts you, and you miss it. Read to the end before you commit.

Answering the diagnosis when the question wants the next step. You prove you know the disease, and lose the mark.

Over-reading rare diseases. When in doubt, think of the common cause of an uncommon presentation before the rare cause of a common one. Unless the stem hands you a pathognomonic clue, the common answer is usually right.

Being thrown by distractors. A mention of an unrelated drug, a normal investigation, a family history that doesn't matter. Ask whether it changes the answer.

Ignoring the time course. "Over three days" versus "over three months" can separate diagnoses completely.

Misreading a negative. "No" and "not" are easy to skim past.

Changing the question to fit what you know. If you notice yourself saying "but if they'd also mentioned X...", stop. Answer the question that was asked.

Practice this skill deliberately

Reading vignettes well is a skill, so practise it as one.

Practise the sequence on easy questions first. Ask, read, name, answer, check. Do it until it's automatic. Then it'll hold up under time pressure.

Do some questions with the options covered. Name the answer first, then uncover.

Time yourself on vignettes specifically. Find your average and your outliers.

Review your misses by type. When you get a vignette wrong, was it knowledge, or was it reading? Keep a tally. If reading is the problem, knowledge drills won't help.

Write a one-line summary of long questions in review. "45M, alcohol + DM, RLL consolidation, organism?" Summarising trains you to pick out what matters.

Read the explanations for the wrong options. They show you the other questions the examiner could have asked, and the specific details that rule them out.

A fourth example: the negative question

Which of the following is NOT a feature of nephrotic syndrome? Options might include heavy proteinuria, hypoalbuminaemia, oedema and haematuria with red cell casts.

With a negative question, flip your task. You're no longer looking for what fits, you're looking for what doesn't. Go through each option and ask whether it belongs. Heavy proteinuria, low albumin and oedema all belong. Red cell casts and haematuria point to a nephritic picture. So the answer is the odd one out. Many people get these wrong because they rush the first option they recognise as true. Read all four, and keep in mind that you're hunting for the exception.

Lab tables and units

When a vignette includes a table of results, check three things before you interpret anything: the units, the reference range given, and which values are actually abnormal. A haemoglobin in g/dL and one in g/L differ by a factor of ten, and a quick misread can send you to the wrong diagnosis. Look for the pattern among the abnormal values, not each in isolation. A low sodium with a low serum osmolality and concentrated urine points one way. A raised calcium with a low phosphate points another. The pattern, not the single number, tells the story.

Spending your time sensibly

Vignettes take longer, so decide in advance how you'll budget. If you know a question is a classic, answer it quickly and bank the time. If it's long and you can't see the pattern after a minute or so, flag it and move on, and return with fresh eyes. Often the answer is obvious on a second read, once the rest of the paper has relaxed you. And don't let one tangled vignette steal time from three straightforward ones.

A short summary

• Read the last line first so you know what's being asked.
• Mark the clues: setting, key finding, tempo, risk factors, positives and negatives.
• Name your answer before you read the options.
• Answer exactly what the question asked, such as the next step, the best test or the mechanism.
• Test each option against the stem and be wary of absolutes.
• Avoid anchoring, rare-disease over-reading and answering the wrong question.
• Practise reading vignettes as its own skill.

NextPrep's questions are written as vignettes where it makes sense, with a reason shown for every option, so that after each one you can see not only the right answer but which detail in the stem decided it and why the alternatives fell short. They're original practice material, still awaiting independent expert review. Use any good set of questions, and every time you get one wrong, ask the one question that matters most: did I not know it, or did I not read it?