Method in the KFP madness

6 minute read

“I’m terribly sorry, sir – you appear to be suffering from a KFP…”


If you want to make your registrar start trembling, sweating, gibbering and then run off to a corner to either cry or start madly studying, creep up behind them in the corridor and shout “KFP!”

Rightly or wrongly, the Key Feature Problem component of the FRACGP exam has become the most feared. The relatively high failure rates, an apparent randomness of responses, and the common belief that candidates need to read the examiner’s mind all contribute to this.

But while the KFP is a tough exam, and the questions are not always easy to follow, there is some method to the apparent madness.


The FRACGP exam has three components. Its overall aim is to set a standard at which a doctor is deemed competent to practise unsupervised in the general practice setting, anywhere in Australia.

The Applied Knowledge Test assesses clinical knowledge and the OSCE (Objective Structured Clinical Examination) tests the application of knowledge and skills in a “near-real” scenario.

The KFP exam is designed to test clinical reasoning and decision-making. Reasoning is much more difficult to assess than knowledge, however it appears to be one of the key skills that is missing in underperforming doctors. While the KFP is not used internationally as an exam, there is a body of literature behind its use. That said, the KFP is a difficult exam, with an overall pass rate of about 50%. Candidates on the independent pathway have a higher failure rate that those on the AGPT program.

Even more worrying is the rapid decline in the pass rate for those having multiple attempts, leading to a pass rate of around 16% for candidates who have four or more attempts. The message for candidates is to study hard and effectively, to give themselves every chance of passing first time.


The KFP is a written exam with a mixture of short menu and write-in questions. A short menu question gives a variety of options for the candidate to choose, whereas the write-in questions ask for short free-text answers. There are 26 clinical case scenarios, each with equal weighting, in a 3.5 hour exam. This gives a candidate eight minutes per case.


Aiden Barnes, a 10 year old boy presents with a persistent cough which started two weeks ago. Approximately three weeks ago, Aiden has a mild “cold”, experiencing a few days of runny nose, but this seems to have cleared up. Aiden attends the local primary school with his five year old sister. His mother is 36 weeks pregnant. Aiden and his sister are fully immunised.

What features on further history would lead you to suspect a diagnosis of pertussis? List up to four (4).

This question above is designed to test the ability of the candidate to think. It is asking about which features of the history are needed to make this diagnosis, i.e. the “key features” rather than every feature. Candidates should reflect on this question as if it were an actual clinical scenario and consider what they would do in practice. Most importantly, they should think about their clinical reasoning strategies, remembering that the idea of clinical reasoning is to “rule in” or “rule out” a diagnosis.

If a question is a reasonable one to ask as part of history taking, such as “Are his growth centiles normal?”, but this doesn’t help to focus the diagnosis in on pertussis, then this will not score.

Asking if the child is fully immunised against pertussis is important, as unimmunised children are at higher risk of pertussis.

The risk, however, is that the candidate may inadvertently make it hard for the examiner to score them well. The question asks for four reasons and there are four boxes.

What happens if in the first box the answer is given as “history of cough, wheeze, and shortness of breath”? Unfortunately this is over-coding, as three different symptoms have been listed in a single box, and this incurs penalty marks for over-coding.

Over-coding is not to be confused with clarification. Candidates can also run into trouble by not supplying enough information. For example, if one of the four answers was listed as simply “immunisations”, this may not be enough to get any marks, or certainly not full marks.

The examiner has to “assume” that “immunisation” means that the candidate is thinking about unimmunised children being at risk of pertussis, but the aim is not to make the examiner work, or assume. Make it obvious what the reasoning process was. For example, “Immunisation history to check if immunised for pertussis” would give a good level of clarification.

Another controversial area is the use of tests, and whether they count as a bedside test or an investigation. The feedback from the RACGP is that where there is a possibility of misinterpretation, these questions are more likely to appear as short-menu rather than a write-in answer.

As you can see even from these examples, this is not a simple, cut-and-dried exam. However, the RACGP exam report assures candidates there is no set proportion of candidates who will pass the exam; rather it is about meeting a standard. So, in an ideal world, the standard remains high and a high proportion of candidates pass.


If you are a GP supervisor with a registrar who is nervous about the KFP, what can you do? The following list of tips is by no means exhaustive, but hopefully it is a start.

1. Read the clinical scenario carefully and identify the key features. This may include gender, age, travel history, symptoms etc.
2. Read all the questions before answering the first one. You will need to click forwards and backwards through screens to do this
3. Read the questions carefully.
4. Note any qualifying statements in the question (see table above).
5. Don’t over-code.
6. Give enough detail.
7. Use Murtagh’s PROMPT strategy to decide on probability diagnosis, red flags, often-missed, masquerades and what the patient is telling you.
8. Murtagh’s triads are useful for diagnoses and key history and examination items.
9. Consider doing a KFP course. The RACGP offers a number of courses, as do private providers.
10. Make sure you do the RACGP sample questions.
11. Be aware of questions floating around that candidates have memorised or that have been written by registrars. The quality is highly variable and can give a false impression of the exam. That said, learning to write KFP questions is a valuable tool and it helps to think through the process of writing a stem and then trying to extract certain answers. See if your study group gets the answers that you think are useful.

Enrol in this free sample exam course to see the complete KFP question discussed in this article:

Thanks to the team of medical educators from Medcast for contributing to these materials
Dr Stephen Barnett is a GP Supervisor, Medical Educator, GP academic and Medical Director of Medcast

End of content

No more pages to load

Log In Register ×