OSCE preparation: make the checklist automatic so you can think

The station is a few minutes long, an examiner is marking against a list, and a standardized patient is waiting for you to be a human being. Doing both at once is the whole skill — and it's a memory problem before it's a clinical one.

Why automaticity is the point

In a short observed station you are running two processes at once: working through a structured sequence, and attending to a person in front of you — their cues, their questions, the thing they say that isn't on your list. Working memory cannot do both if the sequence is still effortful. Every scrap of attention spent retrieving "what comes after ICE?" is attention not spent noticing that the patient just flinched.

So the goal is not to know the checklist. It's for the checklist to cost you nothing. That is a different standard, and it takes a different kind of practice than reading it over.

Re-reading the mark scheme is the trap

Reading a station checklist repeatedly produces a strong feeling of mastery and very little durable recall. In the testing-effect studies (Roediger & Karpicke), material that people kept retrieving stayed near 80% a week later, while material they only restudied dropped to the mid-30s. Confidence and actual recall are famously poorly correlated — which is why so many candidates walk in feeling ready and blank at station three.

Retrieval is the mechanism. Cover the list and produce it from memory, out loud, in order. That single change does more than another hour of reading.

Climb down the cues

Cold recall on day one fails, because there's nothing to retrieve yet. The method proven on newly-learned material is diminishing cues (Fiechter & Benjamin, 2018): full text in front of you, then key words blanked, then first letters only, then nothing at all. Each rung forces retrieval while guaranteeing enough success to be worth doing. In that study, with weak initial learning plain testing gave no benefit at all — the fading ladder did.

For a station, that looks like: read the sequence; recall it with the middle steps blanked; recall it from initials; then run it cold, out loud, in the order you'd actually say it to a patient.

Space it across days, not into one evening

Spacing beats massing, and the useful rule of thumb from Cepeda et al. (2008) is that the gap between sessions should run roughly 10–30% of the time until the exam. Two weeks out, that means short daily passes, not a weekend of cramming. Sleep matters more than most candidates think: Mazza et al. (2016) found the same two sessions split by a night's sleep took about half the relearning effort and held up much better months later.

And don't retire a station because you got it right once. That's precisely the condition that produced the collapse to the mid-30s. Keep finished stations in a light rotation.

Drill the openings and the transitions hardest

Memory for sequences is U-shaped — beginnings and endings survive, middles blur. Worse, in a sequence the cue for step four is the completion of step three, so if you only ever practise top-to-bottom, your transitions get the least practice while carrying the most load. Two cheap fixes: start run-throughs from a random point in the station, and drill the seams directly — given the last thing you said, produce the next move.

This is also your recovery skill. If you blank mid-station, you don't need the whole list — you need the next step from where you are.

Say it out loud, in the words you'd use with a patient

Speaking aloud beats silent review by 10–20% (the production effect), and it's the modality you'll be examined in. There's a second benefit specific to OSCEs: rehearsing in clinical shorthand and then having to translate live is a common failure. Practise the actual sentence — "I'd like to ask what you think might be going on" — not the abbreviation.

Practise with someone being slightly difficult

Stations rarely go as scripted. The patient goes off-topic, gets upset, asks the question you were saving for later, or an examiner's face gives you nothing. Practising under mild evaluative stress inoculates against choking (Beilock & Carr, 2001) — the arousal stops being novel.

So put yourself under a little heat: a timer running, a phone recording, a study partner told to be unhelpful. LobsterClass's roleplay can play the difficult patient and grade you against your own sequence, so the first awkward station you handle isn't a real one.

Get the sequence to the point where it runs itself, and the exam stops being a memory test. That's the whole game — free up the attention, then spend it on the person.