In short
- The reference range is not the line between healthy and sick. It is the span into which the middle 95% of healthy people fall.
- This has an unavoidable result: one in twenty results in a healthy person is outside the range. With a panel of 30 tests, the chance that at least one is flagged is about 79%.
- What matters is three pieces of information together: how far outside, which way it is moving over time, and whether you have symptoms.
- A few values matter even with a small deviation: potassium, calcium, glucose, platelets. There, the rule “a little outside is nothing” does not apply.
- Repeating the same test gives a different number, even if nothing has changed in you. That is not a lab error.
The most common question we get about a report is this: one value is slightly outside the range, is that bad? The answer is almost always the same. On its own, it does not tell you enough. This article explains why, and what to look at instead.
What the reference range is
The lab does not set the reference range by where disease begins. It sets it by measuring a large group of healthy people and taking the middle 95 percent of results. The lower limit is therefore the 2.5th percentile of that group, and the upper limit the 97.5th percentile. That is how the range is defined under the international standard for setting reference intervals.
This has two consequences worth understanding. First: by definition, 5 percent of perfectly healthy people are outside the range. Second: someone with a real problem can also be inside the range. Their personal value has shifted but has not yet crossed the group limit.
The limit is not the crossing between health and disease
A value 0.1 unit above the upper limit and a value 0.1 unit below it mean practically the same state. The limit is useful as a warning, not as a verdict.
Why a deviation is expected in a large panel
For one test, the chance that a healthy person is outside the range is 5 percent. So the chance of being inside is 95 percent. With several independent tests, these probabilities multiply:
| Number of tests in the panel | Chance that all are within range | Chance of at least one deviation |
|---|---|---|
| 1 | 95% | 5% |
| 10 | 60% | 40% |
| 20 | 36% | 64% |
| 30 | 21% | 79% |
| 50 | 8% | 92% |
With a panel of fifty tests, it is almost certain that at least one value will be flagged. That holds even in a completely healthy person. It is not an error and not a reason to panic. It is arithmetic that anyone ordering a large panel should know.
In reality, tests are not fully independent of each other. So the true percentage is somewhat lower than calculated. The order of magnitude holds.
The three facts that only mean something together
| Question | What it means | Example |
|---|---|---|
| How far outside? | A deviation of a few percent is a different thing from a doubled value. | ALT 42 with a limit of 40 is not the same as ALT 240. |
| Which way is it moving? | One measurement is a photo, a series is a film. The direction says more than the number. | Ferritin 28 → 22 → 16 matters more than a one-off 28. |
| Do you have symptoms? | The same number in a person with no complaints and in a person with complaints means different things. | TSH 4.6 without symptoms versus TSH 4.6 with exhaustion and feeling cold. |
Say you answer all three questions reassuringly: a little outside, no trend, no symptoms. Then the right step is almost always a repeat in three months. Not an extra test, and not a medicine.
Where “a little outside” still counts
For some tests the body is so tightly regulated that even a small shift means something. For these, the rule about a reassuring small deviation does not apply:
- Potassium. A shift of a few tenths can affect heart rhythm. Every deviation needs an explanation.
- Calcium (read together with albumin). Persistently raised calcium always has a cause that needs to be found.
- Fasting glucose. The range between 5.6 and 6.9 mmol/L is not “almost fine” but the name for prediabetes.
- Platelets. Low platelets change the risk of bleeding and feed into calculations such as FIB-4.
- Haemoglobin. Mild anaemia is often the first sign of blood loss that needs to be explained, not topped up.
Why a repeat differs even though nothing has changed
Every measured value carries two kinds of variation. Biological variation is the body's own: the same substance is not the same every day. Analytical variation is the method's. Together they mean that two measurements in the same person in the same week will differ, even if nothing has happened.
For some tests this variation is small (sodium, calcium). For others it is large (triglycerides, iron, cortisol). That is why, for iron, we look at ferritin and not serum iron. Serum iron swings from hour to hour.
A practical rule: before you conclude that something has changed between two reports, ask one question. Is the difference bigger than the usual variation of that test? If the difference is a tenth of the value, you are probably looking at noise.
How Blood Lab 360 reads it
Our app does not label values only as “within range” and “outside range”. Every marker gets an assessment that takes into account your sex and age band. It also looks at how far from the limit the value is, and at the comparison with previous reports. Where a number without context says nothing, the app says exactly that. It does not invent an assessment.
If the report has no information on sex or age, the explanation for dependent markers stays closed until you add it. We would rather say nothing than say something wrong.
Frequently asked questions
One value is slightly raised. Is that a disease?
Probably not. The reference range covers the middle 95% of healthy people, so one in twenty results in a healthy person is outside the range. It becomes important if the deviation is large, if it grows over time or if you have symptoms.
Why is something almost always flagged in a large panel?
Because the probabilities multiply. With 30 tests, the chance of at least one deviation in a healthy person is about 79%, with 50 tests around 92%.
When does it make sense to repeat a test?
Usually in three months, for the thyroid in 6–12 weeks. Before that, most values have not moved enough for a fair comparison.
Why are two reports different even though nothing has changed in me?
Every measurement carries the body's biological variation and the method's analytical variation. For triglycerides, iron and cortisol this variation is large, for sodium and calcium small.
Can I compare one lab's reference range with another's?
Carefully. Methods differ, so the limits differ too. For tracking over time, stay with the same laboratory.
Sources: CLSI/IFCC — setting reference intervals as the middle 95% of a healthy population · IFCC C-RIDL — reference intervals and comparability between laboratories · EFLM — biological variation and the smallest significant difference between two measurements · our knowledge base (26 sex/age bands).
