Normal lab values, explained in plain English

How to read your result

A reference range is not a boundary between healthy and sick. It is the middle 95% of results from a group of healthy people, which means about 5 in 100 healthy people fall outside it on any single test, by definition. Run enough tests and that adds up fast: a 2025 analysis by Doles and colleagues in the Journal of the American Board of Family Medicine calculated that 64.2% of normal patients would have at least one abnormal result on a panel of 20 tests.

That is why a single flagged value, sitting a point or two past the edge, is so often repeated rather than acted on. Distance matters more than the flag.

The second thing to know: the range printed on your own report wins. Every lab builds its range from its own instruments and its own local population, so the same value can be flagged at one lab and normal at another. That is the reason this tool has a field for your lab's range. Enter the low and high from your report and the reading switches to yours.

And a normal result can still be worth a question. A value that has climbed from the bottom of the range to the top across three years never once got flagged, but the trend is real. Numbers also lean on each other: total calcium reads low when albumin is low, even when the active calcium is fine. Normal means common, not proven fine.

One example, start to finish

A daughter is reading her father's discharge paperwork. The CBC shows hemoglobin 11.2, flagged low. Platelets 240,000 and MCV 89 are both unflagged. She wants to know how worried to be before the follow-up on Thursday.

Hemoglobin 11.2 sits below the 13 to 18 range used for men, so it is anemia, but it is not far out and it is a common finding after a hospital stay. Repeated blood draws, IV fluids diluting the blood, and several days of eating poorly all pull it down. The MCV is what narrows things: at 89, the red cells are normal-sized, which argues against a straightforward iron shortfall (that usually drags MCV under 80) and against a B12 or folate problem (which pushes it over 100). Normal platelets alongside it say the drop is not across every cell line, which is reassuring.

What that combination does not tell her is whether he is still bleeding somewhere, or whether this is a slow decline that started before the admission. Both need the previous results for comparison. So the question she brings on Thursday is not "is 11.2 bad." It is "what was his hemoglobin six months ago, and is this new?" The guide to reading lab results walks through building that comparison from paperwork you already have. If treatment is part of the picture, low blood counts during chemo covers what those same numbers mean on a treatment schedule.

When to actually call

Most lab pages stop at "talk to your doctor," which is not useful at nine at night with a portal notification on your phone. Here is the shape of the answer, from the caregiver side of it.

• Call now if a number is far outside its range and the person has symptoms that match: confusion or new weakness with a sodium or calcium result, palpitations with potassium, chest or belly pain with a very high triglyceride level, yellowing eyes with bilirubin, or fever with a low neutrophil count during cancer treatment. Fever plus a low neutrophil count is the one that goes straight to the treating team or the emergency room, not to a message.
• Message this week if a value is clearly outside its range but the person feels fine, or if a number has moved a lot since the last draw even while staying in range. Ask two things: should this be repeated, and does anything change in the meantime.
• Bring it to the next visit if the value is a small distance outside, unchanged from last time, or one of the results that bounces for ordinary reasons. Hydration, a hard workout, a fasting sample that wasn't fasting, and a difficult draw all account for more flags than illness does.

When you do call, lead with the change rather than the number. "His creatinine went from 0.9 to 1.4 in six weeks" gets a faster, better answer than "his creatinine is 1.4." If you're staring at one specific flagged value, we have decoders for the ones people ask about most, including what a high RDW means and the BUN to creatinine ratio.

Where these ranges come from

The numeric backbone is the American Board of Internal Medicine laboratory reference ranges, revised January 2026, cross-checked against StatPearls on the NCBI Bookshelf. The plain-English explanations come from MedlinePlus, the patient-facing service of the National Library of Medicine, with each test's page linked next to its result. Everything here was checked on August 27, 2026.

Those sources do not always agree, and pretending otherwise would be dishonest. On about 10 of the 34 tests here, the two differ somewhere. The two widest gaps: the ferritin upper limit for women is 150 in one source and 307 in the other, and the TSH upper limit is 4.0 in one and 5.0 in the other. Where they disagree, this tool refuses to call you abnormal on a number reasonable experts argue about. TSH results between 4.0 and 5.0 show as a neutral borderline band with the dispute named. Ferritin uses the wider upper limit, so a woman's result between 150 and 307 reads as inside the range, with a note that one source would place it outside.

A few other decisions worth stating. All values are adult, and US conventional units, the ones printed on most reports in the United States. Much of the world uses SI units, where the same result carries a completely different number. The lipid and A1c figures are category cut points, not reference ranges, so there is no low flag on them. HDL is read the other way around, where low is the flag. And TSH reads inversely: a high TSH means an underactive thyroid.

None of this is a diagnosis.

Questions about lab results

How do I interpret my blood test results?

Start with the range printed next to each number on your own report, not a range you found online. Then look at direction and distance: is the value above or below, and by how much. A value a hair outside the range means something different from one at double the top. Last, read the numbers as a set, because a low hemoglobin next to a low MCV tells a different story than a low hemoglobin with a normal MCV.

What are the most important numbers in a blood test?

It depends on why the test was ordered, which is the honest answer most pages skip. For general health, hemoglobin, white count, platelets, creatinine, and glucose carry the most weight. During cancer treatment, the absolute neutrophil count and platelets are watched most closely because they decide infection and bleeding risk. For heart risk, LDL and triglycerides matter more than total cholesterol.

When should I worry about blood test results?

Distance from the range matters more than being outside it. A number a few points past the edge is usually repeated rather than acted on. A number far outside, a number that has moved sharply since the last draw, or any result paired with new symptoms is the one to raise. Calcium, potassium, and a very high triglyceride level are the ones clinicians tend to act on quickly.

Why do reference ranges differ between labs?

Each lab sets its range from its own instruments and its own local sample of healthy people, so two labs can print different numbers for the same test. Authorities disagree too. On the ferritin upper limit for women, one major source says 150 and another says 307. That is why this tool lets you enter the range from your own report and reads your value against that instead.

Can a result be normal and still be a problem?

Yes, and it happens more than most people expect. A value that has moved from the bottom of the range to the top over a year is still normal on every report, but the trend is real. A number can also be pulled into range by something else, the way a low albumin makes calcium read lower than it truly is. Normal means common, not proven fine.

Can AI interpret my lab results?

It can explain what a test measures and what high or low usually means, which is what this page does without any AI at all. What it cannot do is diagnose. Interpretation depends on the medications someone takes, what the last result was, why the test was ordered, and what the exam showed. Use a tool like this to arrive at the appointment with a sharper question, not to replace the answer.

Is this a diagnosis?

No. It is education, and your care team reads these numbers with the whole picture in view.