Reference range or optimal range: what a "normal" result does and doesn't tell you

A "normal" result means the number falls inside the laboratory's reference interval: the range that covers most of the healthy people the lab measured. It tells you the result is common. It doesn't tell you the result is ideal for this person, or that nothing is changing. An optimal range is a narrower target inside it, and it's only as good as the evidence behind it, which for some markers is strong and for many is thin.

I build Sera, a lab-review workspace for practitioners, and the examples below come from its published range set, where every band is linked to its sources. Read the last section with that in mind.

How a reference interval is built

A laboratory measures a marker in a group of reference people, usually adults screened as healthy, and takes the range that covers the middle of their results. By convention that's the central 95%. The top 2.5% and the bottom 2.5% fall outside by design.

That has a consequence people often miss. A 2024 review of reference intervals in the British Journal of Biomedical Science points out that about 5% of disease-free people will fall outside any given interval, and that the more tests you run, the more likely at least one result is flagged. MedlinePlus says the same thing in plainer words: it's common for healthy people to sometimes have results outside the reference range.

Not every lab builds its own interval from volunteers. The same review describes direct methods, where a screened healthy group is measured, indirect methods, where intervals are estimated from large sets of routine results, and labs verifying a manufacturer's interval rather than setting one from scratch.

Why the interval differs from lab to lab

The same review lists the reasons: method bias, the population the interval came from, and the age, sex and ethnicity mix of that population. In practice it means two labs can print different limits for the same marker. MedlinePlus goes as far as saying you can't compare test results from different labs.

TSH shows how much this moves. The traditional reference range is 0.5 to 4.5 mIU/L, as the American Thyroid Association describes it. Mayo Clinic Laboratories prints 0.3 to 4.2. Labcorp prints 0.45 to 4.5. A UK NHS laboratory prints 0.38 to 5.33 (TSH sources). A TSH of 4.8 is high at one of those labs and normal at another. Neither lab is wrong. Each interval belongs to its own assay and its own reference group.

What a "normal" result doesn't tell you

Where in the range it sits. A fasting glucose of 72 and one of 98 mg/dL are both inside Labcorp's interval of 70 to 99. They aren't the same result.

Which way it's moving. A marker can drift from the bottom of the interval to the top over three years and never be flagged. One result says very little about a trend. Several results from the same lab, under the same conditions, say a lot more.

What it means alongside other markers. Ferritin is the usual example: inflammation raises it regardless of iron stores, so a normal-looking ferritin can sit beside iron deficiency (ferritin sources).

Whether the reference group looks like your client. An adult interval may not suit an older client, a pregnant client or a teenager, and some intervals aren't split by sex or age.

This is the gap clients feel when they hear "your labs are normal" and still don't feel well. It's a real gap. It isn't proof that the result is abnormal.

What a narrower target adds

An optimal range makes a claim the reference interval doesn't make: that results in one part of the range go with better health or lower risk. When there's outcome data behind that claim, it's useful. When there isn't, it's a preference.

Fasting glucose has both kinds of edge (fasting glucose sources):

• Sera's sufficient band stops at 99 mg/dL (5.5 mmol/L), just below the American Diabetes Association's threshold for impaired fasting glucose at 100 mg/dL. That edge is a guideline figure.
• The line isn't the same everywhere. NICE in the UK counts 5.5 to 6.9 mmol/L as high risk, and Diabetes Canada doesn't call fasting glucose impaired until 6.1 mmol/L.
• The lower edge of 70 mg/dL is the usual laboratory lower limit and the conventional definition of low blood sugar.
• The optimal band ending at 85 mg/dL is a practitioner target. No current guideline or recent cohort study sets it.

HbA1c is a useful contrast. Its optimal band ending at 5.4% follows pooled European and US cohort data in which 5.0 to under 5.5% was the lowest-risk group (HbA1c sources). So a narrower target can rest on data. The point is to know which ones do.

Where the evidence for narrower targets is thin

TSH is the clearest case. Sera's optimal band for TSH ends at 2.0 mIU/L, a target common in functional practice, and there's no guideline behind it. In pooled data on more than 134,000 adults, cardiovascular and mortality risk was lowest at a TSH of roughly 1.9 to 2.9, and the lowest fifth of TSH carried higher all-cause mortality. That data doesn't support a target that stops at 2. Sera labels the band as a functional target and says so in the note on the TSH page.

That doesn't make narrower targets useless. A target can be a sensible place to look more closely, to retest, or to ask more questions. It becomes a problem when it's presented as a diagnosis, or when a result just outside it is treated as a finding in itself.

How to explain the difference to a client

Plain wording works best:

• "Normal means your result is in the range most healthy people tested at this lab fall into."
• "Your result is in that range, but near the top. I'd like to see where it is next time."
• "The narrower range I use is a target, not a diagnosis. Here's where it comes from."

If a result is outside the lab's interval, or the picture doesn't fit, that's the point to refer, not to reinterpret.

Where Sera fits

Every band Sera uses is published with its sources and an evidence grade, so you can see whether an edge is a guideline threshold, a published reference interval or a functional target before you rely on it. A client's results are shown as optimal, sufficient or needs attention, over years and across labs, with up to three markers on one chart. Sera doesn't store or show the lab's own reference interval, so keep the original report for that. It keeps every result linked to its page in the report, and it doesn't diagnose or suggest treatment.

It's free for three clients, with no card. There's a 3-minute tour if you'd rather watch first.

Alina Radu is the founder of Sera and a software engineer. The idea started with her own Hashimoto's tests, scattered across years without the big picture.