The RACP Divisional Written Examination is the gate at the end of Basic Training, and it consists of two papers sat on the same day: a Clinical Applications paper of 100 questions and a Medical Sciences paper of 70, giving 170 questions across roughly six hours. Both are single best answer with a small number of extended matching items, and there is no negative marking. The structure tells you something important that candidates consistently ignore. These are not two halves of one exam. They are two different assessments, testing two different things, and you have spent the last two years getting better at one of them while the other quietly decayed.
Key takeaways
- The DWE is 170 questions across two papers on one day: Clinical Applications, 100, and Medical Sciences, 70.
- Clinical Applications is what your ward job has been training you for. Medical Sciences is not.
- Basic science decays fast in clinical training, and the exam knows this.
- There is no negative marking, so every unanswered question is a mark surrendered for nothing.
- Six hours in a day is an endurance test as well as a knowledge test, and it needs rehearsing.
Two papers, two problems
Look honestly at where your last two years have gone.
Clinical Applications is a paper of clinical vignettes: a patient, some investigations, a decision. That is what you do every day. Your registrar job has been an extended training course for this paper, and while it is the larger of the two and demands genuine preparation, it is preparation on ground you occupy.
Medical Sciences is a different proposition. It asks about the physiology, the pharmacology, the immunology, the genetics and the pathophysiology that underlie the medicine, and it asks at a level you have not been operating at since your preclinical years. Nobody on a ward round asks you to explain receptor kinetics or the mechanism of a channelopathy. So it goes.
The consequence is predictable. Candidates prepare for the paper that feels like their job, they arrive at Medical Sciences under-prepared, and the seventy questions they were least ready for cost them the exam.
Track them separately
The measurement error that costs candidates most is a single overall percentage.
Your combined figure will be carried by the clinical questions, because there are more of them and because you are better at them. It can climb steadily while your basic science performance sits well below the line, and you will not see it.
Keep two dashboards. Set a separate target for each. Be honest about which paper you have been avoiding, because it will be the one you find less enjoyable, and the examination will not care about your preference.
Medical Sciences must be derived, not memorised
The instinct, having identified basic science as the weakness, is to go back and re-read the physiology. That is slow, passive, and produces recognition rather than the ability to reason under pressure.
The alternative is the same protocol that works for any mechanism-heavy paper. For each question, before you look at the options, state the mechanism in one line, and then predict what it should produce.
If the prediction matches the vignette, the answer is usually now obvious and you have derived it rather than recognised it. If it conflicts, that conflict is the most valuable thing in the question, because your mechanism was wrong and the direction of the conflict tells you where to look.
When you are wrong, find the exact link that broke. Did you misidentify the mechanism, mispredict its consequence, or fail to map it onto the clinical picture? Those are three different gaps with three different remedies, and only the first is a factual one.
Do not neglect the statistics
Within the science content sits a small, finite, entirely learnable body of material that candidates reliably postpone: study design, bias and confounding, sensitivity and specificity, predictive values and their dependence on prevalence, likelihood ratios, number needed to treat, and the interpretation of confidence intervals.
It recurs, it is straightforward once learned, and it decays, which means most of your competitors will half-know it on the day. Two weeks of short sessions secures it permanently.
Clinical Applications rewards decisions, not diagnoses
A word on the larger paper, because candidates over-prepare the diagnosis and under-prepare the decision.
The vignettes will usually make the diagnosis reachable. What separates candidates is what comes next: which investigation actually changes management, what to do first when several things need doing, when not to treat, and when to escalate.
So when you review a wrong answer on this paper, ask which of those failed. If you diagnosed correctly and then chose an action from the wrong stage of care, that is not a knowledge gap and revising the disease harder will not fix it.
Weight your practice towards Australian guidance, because that is what defines the correct answer: Australian therapeutic guidelines, Australian prescribing conventions, and the pathways that actually exist here. A candidate whose reflexes were built elsewhere can reason impeccably to an answer that is not the answer, and that is a calibration problem rather than a knowledge one.
Answer everything
There is no negative marking. A wrong answer costs you nothing that a blank does not, which means a blank is a mark given away for free.
If you are running short of time in either paper, ensure every question has an answer before you go back to refine the ones you flagged. Eliminate what you can and guess between the rest. Candidates still leave blanks, and it is the cheapest mistake available.
Rehearse the day, not just the content
Six hours of dense reasoning in a single day is a physical task as well as an intellectual one, and it is trainable.
Build to full-length papers: a timed 100-question paper, and a timed 70-question paper. And at least once, sit both on the same day, with a realistic interval, because that is the thing you will actually do.
Then look at your accuracy by position. If it holds across both papers, your endurance is fine and your errors are knowledge errors. If it falls away in the second paper, or in the back half of each, you have a concentration problem, and more content revision will not touch it. That pattern calls for longer practice sessions, attention to sleep, and a deliberate plan for how you will use the break.
Where iatroX fits
iatroX's RACP bank covers both the clinical applications and the basic science content as separately tracked domains, so a decaying medical sciences score cannot hide behind your clinical strength, which is exactly what a blended percentage does with it. The adaptive engine targets the domains where you are genuinely weak, spaced repetition holds the mechanisms and statistical concepts that decay fastest, and missed questions can be opened in the Socratic Tutor, which asks you to derive and reason before it explains. Try it with free sample questions at iatroX. For the derivation habit itself, see rebuilding basic science through mechanism questions.
Frequently asked questions
What is the format of the RACP Divisional Written Examination? Two papers on the same day: a Clinical Applications paper of 100 questions and a Medical Sciences paper of 70, giving 170 questions in total across roughly six hours. They are single best answer with a small number of extended matching items, and there is no negative marking.
Which paper do candidates under-prepare? Medical Sciences. Clinical Applications resembles the job you have been doing for two years, while the basic science underlying it has not been used since your preclinical years and has decayed accordingly.
How should I revise basic science for the DWE? By deriving rather than re-reading. State the mechanism before you look at the options, predict what it should produce, and compare that with the vignette. Reading physiology produces recognition rather than the ability to reason under pressure.
Should I guess if I do not know an answer? Always. There is no negative marking, so a blank guarantees zero while a guess after eliminating an option or two has a reasonable chance of scoring. Ensure every question has an answer before refining your flagged ones.