Data interpretation on the ward

Why this matters on day one

You will review far more results than you order, often out of context and out of hours. The skill is triage — separating the abnormal-but-expected from the abnormal-and-dangerous.

The sequence

1. 1Always ask three things. Is this a change from the previous value? Does it fit the clinical picture? Does it need action NOW, action later, or no action? A potassium of 6.9 in a haemolysed sample needs a repeat, not calcium gluconate — but you do both if the ECG is abnormal.
2. 2Know the ones that need action tonight. Potassium above 6.5 or with ECG changes. Sodium changing faster than 10 mmol/L per day. Corrected calcium above 3.0. Glucose below 4 or with ketones. Platelets below 20. Neutrophils below 0.5 with fever. Troponin rise with chest pain. INR above range with bleeding. pH below 7.2.
3. 3Interpret in structure. ABG: pH, then CO₂, then bicarbonate, then compensation, then oxygenation and the A–a difference. ECG: rate, rhythm, axis, intervals, then morphology, always against a previous. Chest film: projection, adequacy, then a system you always use.
4. 4Escalate the abnormal you cannot explain. An unexplained result is an active problem, not a filing task. Document the value, your interpretation, your action and your review time.
5. 5Do not over-investigate. Daily bloods on a stable patient generate incidental findings, anaemia and cannula infections. The content map lists avoiding over-investigation as a capability in its own right.

What goes wrong

• Reviewing a result without the previous value.
• Treating a haemolysed potassium; or dismissing a real one as haemolysis without an ECG.
• Correcting sodium too quickly — osmotic demyelination is iatrogenic and irreversible.
• Signing off a result you do not understand because it is 4am.
• Not looking at the patient. A "normal" set of bloods in someone who looks terrible means keep looking.