When a Blood Test Looks Reassuring: Five Caveats for the Clinical Conversation

Blood tests are among the most familiar tools of UK primary care, and among the easiest to over-read. A result that sits within a reference range can still be less decisive than it looks, because the number is shaped by the sample, the patient's preparation, the laboratory method and the question being asked. In 2026, testing literacy is part of translational care: it is how evidence about analytical variation becomes a sensible conversation in a ten-minute consultation.

1. Haemolysis: when the sample, not the patient, speaks

Haemolysis occurs when red blood cells break during collection, transport or processing. It can raise potassium, lactate dehydrogenase and aspartate aminotransferase, and it can make other results harder to interpret. UK laboratories usually flag haemolysed samples, but the flag is easily lost in a busy consulting room. A single unexpected potassium in a well patient should prompt a check on sample quality before it prompts a diagnosis.

The clinical skill is not simply recognising the flag. It is deciding when the result is unusable, when it needs repeating and when it can be interpreted with caution. A haemolysed sample does not always invalidate every analyte, but it weakens the confidence that can be placed in those most affected.

Consultation script: This sample was haemolysed, which can make some results, particularly potassium, look higher than they really are. I would rather repeat it than act on a number that may be an artefact.

2. Fasting status: the variable that goes missing

Fasting is not required for every test, but when it is required the request should say so and the result should be read in that light. A non-fasting lipid profile or glucose result may still be interpretable, but it is not interchangeable with a fasting result. The problem is rarely that clinicians do not know this. It is that fasting status is often unrecorded, assumed or lost between request and report.

A pragmatic approach is to document what is known and to be explicit when it is not. If fasting status is unknown and the decision depends on it, the safer course may be to repeat the test under controlled conditions rather than to treat the first result as definitive.

Consultation script: We did not record whether you had fasted. That does not make the test useless, but it changes how firmly I can interpret this figure. If it matters for the decision, we should repeat it fasting.

3. Reference ranges drift between laboratories and over time

Reference intervals are method-dependent and population-dependent. A result that is normal on one laboratory platform may be flagged on another, and ranges can change when an assay or analyser changes. Patients who move between practices, or whose care moves between hospital and community laboratories, should not have results compared as if the numbers came from a single universal ruler.

This does not mean that every mildly abnormal result is meaningless. It means that trend interpretation requires like-for-like comparison. A small difference between two laboratories may reflect the method rather than the patient.

Consultation script: The normal range is not a universal constant; it depends on the laboratory and the method. This result is only mildly outside the range used here, and your previous results were from a different laboratory, so I would not treat the difference as a real change without checking.

4. Incidental findings: not every abnormal number is a diagnosis

Incidental abnormalities are common: a mildly raised liver enzyme, a borderline thyroid-stimulating hormone, a low vitamin D level, a slightly elevated inflammatory marker. Some are clinically important. Many are not, or not yet. The question is whether the result explains symptoms, changes management or merely starts a cascade of repeat tests, referrals and anxiety.

Good testing literacy includes the discipline of asking what the result is for. If it will not change the decision, it may not need to be tested. If it has been tested and is mildly abnormal, the response should be proportionate: repeat, review symptoms, consider confounders and set a clear safety-net rather than rushing to investigation.

Consultation script: This is a mild abnormality that is common and often resolves. It is not an emergency, but it is worth repeating in a few weeks and checking for specific symptoms. If those symptoms develop, we should review sooner.

5. Repeat testing: a threshold, not a reflex

Repeating a test can resolve analytical noise, confirm a trend or rule out a transient abnormality. It can also waste time, create anxiety and generate further incidental findings. The useful question is not whether to repeat, but under what threshold, at what interval and for what purpose. A repeat should have a defined trigger and a defined next step.

A repeat taken too soon may not reflect a meaningful change. A repeat taken too late may delay necessary care. The threshold should be set before the first result is filed, and it should be explained to the patient in plain language.

Consultation script: Because this result is unexpected and not extreme, the next step is a repeat in a few weeks rather than immediate treatment. If the repeat is similar, or if you develop specific symptoms, we will move to the next step.

Putting uncertainty into plain English

Uncertainty is not a reason to withhold testing. It is a reason to describe what a test can and cannot settle. Useful phrases include: this result makes a condition more likely, but it does not rule it out; this number is affected by how the sample was taken; and we are testing to change a decision, not to collect data.

For patients, the practical questions are straightforward: What is this test for? Does preparation matter? Which laboratory is processing it? What happens next if the result is borderline? For clinicians, the equivalent discipline is to document sample quality, fasting status, laboratory and previous result, and to give a repeat threshold and safety-net advice.

Testing literacy is translational care in miniature. It turns the analytical reality of the laboratory into a conversation that patients can understand and clinicians can act on. A normal-looking result deserves neither blind reassurance nor unnecessary alarm. It deserves an interpretation that is honest about its limits.