General Data Interpretation

Data interpretation stations often feel more intimidating than they need to be. In reality, they reward candidates who work through the data systematically rather than jumping straight to conclusions. Below is a practical approach you can apply to any graph, chart, or table you’re given.

1. Orient yourself first

Before you say anything about what the data shows, take a moment to understand what you’re actually looking at:

• What does the title tell you the data represents?
• Where has this data come from, if a source is given? Consider whether it’s a reliable body (e.g. ONS, NHS Digital, a peer-reviewed study).
• What type of visual is it — bar chart, line graph, scatter plot, table?
• What are the axes measuring, and in what units (e.g. cases per 100,000 people, percentage, £)?
• Are there any keys, legends, or footnotes you need to factor in?

Skipping this step is one of the most common mistakes — candidates who rush straight to “the trend is…” often misinterpret the data because they haven’t checked the units or scale properly.

2. Describe the overall pattern

Once you’re confident you understand the data, zoom out and describe the big picture:

• What is the overall direction of change — is it rising, falling, staying flat, fluctuating?
• Does this match what you’d expect, given your background knowledge? If not, that’s worth flagging.
• Why might this pattern exist? You don’t need to know the definitive answer — showing you can generate a plausible hypothesis is what matters. For example: “Emergency department attendances appear to have risen sharply from 2020 onwards, which may reflect a backlog of delayed care following the pandemic.”
• If you want to impress, go one step further and suggest what could be done in response — either to reverse a negative trend or reinforce a positive one.

3. Zoom in on the smaller shifts

Very few real datasets move in one smooth, uninterrupted direction. Look closely for smaller fluctuations within the bigger trend:

• Are there periods where the pattern briefly reverses or changes pace before returning to the main trend?
• Can you offer a plausible explanation for why this smaller shift might have occurred?
• Could this smaller shift be tackled or reinforced through a specific, targeted intervention, distinct from your answer to the wider trend?

4. Flag anything unusual

Almost every dataset has at least one point that doesn’t fit neatly into the pattern. Interviewers want to see that you notice this rather than glossing over it:

• Is there a data point, or short period, that stands out from the rest?
• What could explain it — a data collection error, an unusual one-off event, a genuine outlier?
• Should this unusual point change your overall interpretation of the data, or can it reasonably be set aside?
• How might you go about checking whether it’s a genuine anomaly (e.g. repeating the measurement, checking methodology, comparing against another dataset)?

5. Bring it all together

Finish with a clear, structured summary rather than trailing off:

• Restate what the data shows in one or two sentences, referencing the overall trend, any smaller shifts, and any anomaly you identified.
• State what this means in practical terms — what does this data tell us about the health issue at hand?
• Connect it to the bigger picture of population health and why collecting this kind of data matters to the NHS. For example: “Tracking this data allows the NHS to evaluate whether current interventions are working, and to direct funding toward the services that need it most.”
• If you want to go further, mention your own enthusiasm for engaging with health data as a future clinician — for instance, contributing to national audits, or being involved in research that adds to the evidence base.

The key takeaway: don’t try to say everything about the data at once. Work through it in layers — first understand it, then describe it broadly, then narrow in on detail, then question what doesn’t fit, and only then draw your conclusions. Practising this approach repeatedly on real past-paper data sets is what will make it feel automatic under interview pressure.