12-Lead ECG Interpretation Practice

There are two ways to learn 12-lead ECGs. The first is to collect patterns: memorise what an inferior infarct looks like, what a bundle branch block looks like, what hyperkalaemia looks like, and reach for the closest match. The second is to read every trace the same way, in the same order, every time, and let the findings assemble into a conclusion.

Pattern collection is faster and it is what most of us do. It also fails in a specific and predictable way: when a trace contains two patterns, or a pattern that has been distorted by something else, the closest match is often the wrong answer. A systematic order costs you fifteen seconds on an obvious trace and saves you the ECGs that matter.

If you would rather read traces than read about them, work through timed ECG cases in the app.

A reading order you can run every time

The order below is unremarkable — most structured approaches contain the same elements. The value is not in the specific sequence but in never deviating from it, including on the traces you think you have already solved.

1. Check the technical details first. Name, date, time, calibration and paper speed. A trace at a non-standard gain or speed will make every measurement you take afterwards wrong, and comparison with a previous ECG is often more informative than the current trace alone.
2. Rate. Count it rather than estimating it. Note whether it fits the clinical picture: a rate of 105 in a patient who should be tachycardic is a different finding from the same rate in someone at rest.
3. Rhythm. Regular or irregular. If irregular, regularly so or not. Then ask whether there is atrial activity, and whether it relates to the ventricular activity.
4. Axis and intervals. PR, QRS duration and QT. A broad QRS changes how you interpret everything that follows, so establish it before you look at ST segments.
5. P waves and conduction. Present, absent, or present but dissociated.
6. QRS morphology. Q waves, R wave progression across the precordial leads, and voltage.
7. ST segments and T waves, lead by lead and in territories. Not "is there ST elevation" but "which contiguous leads, how much, and what is happening in the reciprocal leads".
8. Compare and conclude. Against the previous trace if there is one, and against the patient in front of you.

Two things about this list are commonly skipped and shouldn't be. Step one, because it feels administrative. And the second half of step seven, because reciprocal change is one of the few features that reliably separates a genuine territorial abnormality from an artefact of position or morphology.

Why the broad QRS comes before the ST segment

The single most common structural error in ECG interpretation is assessing repolarisation before establishing depolarisation. If the QRS is broad — bundle branch block, paced rhythm, a ventricular rhythm, a metabolic cause — the ST segments and T waves are no longer being generated by normal conduction, and the thresholds you memorised for a narrow-complex trace do not transfer.

That does not mean a broad-complex trace is uninterpretable. It means the question changes. Instead of "is there ST elevation beyond the threshold", the question becomes "is the repolarisation appropriate for this depolarisation, or is it discordant in a way that requires explanation". Specific criteria exist for that judgement, they are taught on advanced courses, and they are worth learning properly rather than approximately.

The practical rule for anyone still building competence: when the QRS is broad and you are unsure, say so explicitly, escalate to someone who reads more ECGs than you, and make the decision on the patient rather than on your reading. That is not a failure of interpretation. It is an accurate report of your confidence, which is clinically more useful than a confident wrong label.

When the ECG and the patient disagree

This is where most real harm happens, and it happens in both directions.

A reassuring trace in an unwell patient is not reassuring. A single ECG is a twelve-second sample. Early occlusion, intermittent conduction problems, posterior territory changes and evolving pathology can all produce a trace that looks acceptable at the moment it was recorded. If the story is convincing, the trace being normal means you repeat it, serially, and act on the trend and the clinical picture.

An alarming trace in a well patient deserves the same scepticism in reverse. Lead placement errors, movement artefact, baseline abnormalities and normal variants all generate findings that look serious. Before escalating on an isolated abnormality, check the technical quality, repeat the recording, and look for a previous trace.

Both cases resolve the same way: the ECG is one input, it is repeatable, and repeating it is cheap. Serial traces answer questions that a single trace cannot, and the interval between them is a decision you should make deliberately rather than by default.

Four practice traces worth seeking out

These are categories rather than specific ECGs. For each one, find several examples in a teaching library or a case system, and read them with the eight-step order rather than by pattern.

1. The trace with two things happening

Conduction abnormality plus an acute change. Rate abnormality plus a metabolic cause. Pattern-matching produces one answer and stops; the reading order forces you to notice the second finding because you have not yet reached the end of the list.

2. The subtle territorial change

Changes that fall below the thresholds you memorised, or sit in territories that the standard twelve leads represent poorly. The question to practise here is not "does this meet criteria" but "is this ECG consistent with the story I have been given, and what would the next trace tell me".

3. The technically poor recording

Artefact, misplaced leads, wandering baseline. Learning to recognise a trace you should not interpret is a skill in itself, and it is almost never taught explicitly. If your conclusion would change depending on whether a finding is real, repeat the ECG before concluding.

4. The serial pair

Two traces from the same patient, half an hour apart. Practise describing the difference in specific terms — which leads, which direction, how much — rather than as "worse". Precise change descriptions are what make a handover useful.

Reporting what you found

An interpretation that cannot be handed over is only half finished. A useful ECG report has four parts, and it takes about twenty seconds to say out loud.

• Rate and rhythm in plain terms.
• The specific abnormality, described by lead and direction rather than by label where you are uncertain. "ST elevation of two millimetres in the inferior leads with reciprocal depression laterally" survives disagreement; "inferior STEMI" does not.
• Your confidence, stated rather than implied. "I am confident" and "I would like this reviewed" are both clinically actionable. Silence is not.
• What you are doing about it, including when the next trace will be recorded.

The habit of separating description from conclusion is worth building early. It is the part of ECG reporting that stays useful when you are tired, when the trace is ambiguous, and when the person you are handing over to cannot see the paper.

Where to go next

Volume with feedback is the only reliable route to ECG competence, and studies of interpretation accuracy across training levels consistently point the same way: accuracy improves with deliberate practice and structured teaching rather than with time served.

Two useful next steps. Work through traces that are deliberately chosen to defeat pattern-matching — the companion article on the conditions that mimic occlusion is built for exactly that. And practise reading ECGs inside cases rather than in isolation, because the decision that follows the trace is the part that is actually being tested at the bedside.