Article summary
What should you know about Diagnosis, Investigation or Management??
A practical AKP framework for identifying whether a question asks for diagnosis, investigation, management or interpretation, so you answer at the right clinical level.
Why candidates answer the clinical case rather than the question asked
You recognise the diagnosis.
You know the guideline.
You understand what will eventually happen to the child.
And you still choose the wrong answer.
This is one of the more frustrating ways to lose marks in AKP, because the problem is not necessarily missing knowledge. It is often a failure to identify which clinical decision the question is asking you to make.
The RCPCH describes AKP as an examination of clinical decision-making and management. Its case histories are specifically designed to test diagnosis and the planning of investigations and management.
Those are three related but fundamentally different tasks.
A candidate who treats every stem as "work out what disease this child has" will therefore solve some questions beautifully - and answer others at entirely the wrong level.
First decide what job you have been given
Consider three questions built around exactly the same child.
A 7-year-old presents with progressive thirst, polyuria and weight loss.
Question A: What is the most likely diagnosis?
Your task is diagnostic.
Question B: Which investigation should be performed immediately?
Now recognising diabetes mellitus is only the first step. The mark depends on knowing which test matters at this point in the pathway.
Question C: The child is now drowsy, tachypnoeic and poorly perfused. What is the most appropriate immediate management?
The diagnosis may be unchanged, but the task has changed completely.
This sounds obvious when written out. Under examination pressure, it is remarkably easy to forget.
A familiar disease activates everything you know about it: diagnostic tests, complications, specialist referral, definitive treatment, follow-up. One of those facts then appears among the options and feels reassuringly correct.
But AKP does not award marks for selecting something that is true about the disease.
It awards the mark for answering this question, at this point in the case.
Diagnosis asks: what best explains the pattern?
A diagnosis question requires synthesis.
You are asking:
Which condition explains the important findings with the fewest contradictions?
The useful information is usually phenotype, chronology, examination findings and discriminating investigation results.
You are not yet asking what you would order next or how you would treat the child.
This distinction becomes particularly important when one option represents a common diagnosis and another represents a more specific complication.
For example, a child with sickle cell disease, fever and chest pain may clearly have an underlying sickling disorder. But if the question asks what explains a new infiltrate, hypoxia and respiratory deterioration, "sickle cell disease" is no longer the level at which the decision is being tested.
The question is asking you to name what is happening within that disease.
Good clinical questions frequently focus on these critical decision points. The key-feature approach to medical assessment was developed around precisely this idea: some steps in resolving a clinical problem are substantially more important and difficult than others.
In AKP, the diagnosis may therefore be the beginning of the reasoning rather than the end.
Investigation asks: what uncertainty needs resolving now?
Investigation questions produce a different error.
Candidates often choose the best test for the disease instead of the best next investigation for this patient.
Those are not always the same thing.
Imagine a child in whom a particular diagnosis would ultimately be established by MRI. If the child is currently unstable, has a rapidly reversible metabolic abnormality or requires another bedside assessment first, "MRI" may be diagnostically impressive but temporally wrong.
Investigation questions should trigger four checks:
What do I already know? Do not investigate something the stem has already established.
What remains uncertain? The next test should resolve a question that matters.
Will the result change management? An elegant investigation that does not alter the immediate decision may not be the best next step.
Is the child stable enough for it? Definitive investigations come after stabilisation when instability changes priority.
This is why words such as initial, next, most appropriate and confirmatory are not decorative.
They define where you are standing in the pathway.
"Which investigation confirms the diagnosis?" and "Which investigation should be performed next?" can have different answers.
Management asks: what should happen now?
Management questions are where otherwise excellent candidates often become victims of their own knowledge.
The more you know about a condition, the more correct treatments you can imagine.
But treatment has sequence.
A child may need antibiotics, surgery, specialist referral, imaging and long-term preventive therapy. All five may be clinically appropriate.
Only one may answer:
What is the most appropriate immediate management?
The current AKP format deliberately retains more time per question than FOP and TAS because RCPCH analysis found AKP questions to be more complex and to take candidates longer on average.
Part of that complexity is that management requires prioritisation rather than simple recognition.
When you see a management lead-in, ask:
What can harm this child before I get to the next step?
Airway compromise outranks elegant diagnostics.
Shock outranks definitive disease classification.
Active seizures outrank investigation of their eventual cause.
A time-critical treatment should not be delayed because another option would eventually provide more information.
The opposite is also true. If the child is stable, an unnecessarily aggressive emergency intervention may be less appropriate than assessment, investigation or observation.
The word management does not automatically mean "choose the most powerful treatment".
It means choose the action appropriate to the child's state now.
The hidden fourth category: interpretation
Some AKP questions look like diagnosis questions but are actually interpretation questions.
You may be given:
- a blood gas;
- an ECG;
- growth data;
- renal indices;
- a drug level;
- a radiograph;
- or a table of serial laboratory results.
The temptation is to jump directly to the disease.
First ask what the data themselves show.
Is this metabolic or respiratory?
Obstructive or restrictive?
A single abnormal value or a worsening trend?
Appropriate physiological compensation or a second process?
The interpretation becomes the bridge between raw information and the clinical decision.
Skipping that bridge is dangerous because it encourages pattern matching: recognising one familiar abnormality and attaching the first associated diagnosis to it.
Research into diagnostic reasoning has found that errors such as unfocused data collection, premature closure and overly narrow reasoning are associated with poorer clinical assessment performance.
In exam language: do not let the existence of lots of clinical information persuade you that all of it needs to be solved simultaneously.
Use the lead-in to filter the stem
The same detail changes importance depending on what you are being asked.
Suppose an adolescent presents with fever, headache, vomiting, neck stiffness and reduced consciousness.
If the question asks for the diagnosis, the infectious pattern matters heavily.
If it asks for the immediate management, reduced consciousness may become the dominant feature.
If it asks for the next investigation, the level of consciousness may change whether a procedure is safe or appropriate.
If it asks for the likely organism, age, immunisation status and epidemiological context suddenly gain more weight.
The stem has not changed.
The hierarchy of evidence has.
That is why a useful habit is to read the lead-in and mentally complete one sentence:
"My job is to…"
"My job is to identify the disease."
"My job is to decide what to investigate next."
"My job is to treat the immediate threat."
"My job is to interpret this result."
Five words can prevent two minutes of solving the wrong problem.
Beware the answer that is correct eventually
One of the strongest AKP distractors is not medically wrong.
It is simply out of sequence.
You identify the diagnosis and see its definitive treatment among the options. It feels secure because you remember it from a guideline.
But another option belongs first.
This is particularly common in:
- emergency presentations;
- neonatal deterioration;
- endocrinological crises;
- poisoning;
- safeguarding;
- surgical problems;
- neurological emergencies;
- and severe infection.
The distinction is not:
correct versus incorrect.
It is:
correct now versus correct later.
We will deal with that trap in detail in the next article in this series.
A five-second classification before every AKP question
Before comparing the options, classify the question.
D - Diagnosis
What is happening?
Look for the pattern and the strongest discriminator.
I - Investigation
What uncertainty must I resolve next?
Choose the test appropriate to the current stage, not merely the most definitive test available.
M - Management
What should I do now?
Prioritise stability, urgency and sequence.
I - Interpretation
What do these data actually show?
Describe the physiological or pathological pattern before attaching a disease label.
Call it D - I - M - I if you need a mental prompt.
It is not an official RCPCH algorithm. It is a way of preventing a very ordinary reasoning error: giving a sophisticated answer to a question that was never asked.
Knowing the case is not enough
AKP is difficult precisely because clinical knowledge is interconnected.
Once you recognise a disease, you know its investigations. Once you know the investigations, you remember its treatment. Once you remember the treatment, you think about complications.
That interconnectedness is clinically useful.
In an SBA, however, it can make you overshoot.
The examiner does not need everything you know about the child.
They need one decision.
So when a stem feels complicated, resist the urge to solve the entire clinical journey.
Ask something simpler:
Am I being asked to name it, investigate it, interpret it or manage it?
Then answer only that question.
A surprising number of difficult AKP stems become easier once you stop trying to answer all the others.