The most common question in modern medicine
"My blood tests came back normal — why do I still feel terrible?"
It's the question every functional medicine clinic, naturopath, and online health forum hears constantly. It's also the question your primary care physician hears, and often answers with a shrug, a "your labs are fine," and a follow-up appointment scheduled for six months out.
Here's the thing: your doctor isn't lying to you. Your labs probably are inside the reference range. But "inside the reference range" is not the same as "no relevant information here" — and that gap is the source of an enormous amount of frustration, dismissed symptoms, and missed early signals.
This article explains why. It's based on the research, not on suspicion. We'll walk through how reference ranges actually get built, where they fail, what "subclinical" means, and the specific markers most often labeled "normal" while actually telling you something important.
Where reference ranges actually come from
A laboratory reference range is typically defined as the central 95% of values observed in a "reference population" — usually healthy-appearing adults sampled at that specific lab[1]. The bottom 2.5% and the top 2.5% are flagged as "out of range." Everything in between is "normal."
This sounds reasonable until you ask: healthy according to whom? The reference population isn't a perfectly healthy cohort. It's the people who happened to give blood at that lab and weren't already known to have a serious illness. That includes people who were tired, overweight, mildly inflamed, mildly nutrient-deficient, or in early stages of conditions that hadn't yet been diagnosed. The statistical center of that distribution becomes "normal."
The Clinical and Laboratory Standards Institute, which publishes the international guidelines for setting reference ranges, has acknowledged this limitation explicitly: reference intervals describe what's statistically usual in the reference population, not what's biologically optimal[2].
Detection thresholds vs functional thresholds
Standard medicine is built around detecting disease. Reference ranges are calibrated to flag values that suggest active pathology — anemia, hypothyroidism, diabetes, kidney failure. They are not calibrated to flag values that suggest suboptimal function — fatigue, brain fog, mood changes, exercise intolerance, hair thinning.
This is by design. A reference range that flagged every suboptimal value would be flooded with "abnormal" results and lose its clinical utility for detecting actual disease. So the threshold for "abnormal" is intentionally set high enough that most of the people inside the range are functioning more or less acceptably — not flagged for follow-up.
The result: a wide gap exists between "diagnosable disease" and "actually feels great." Someone with a vitamin D level of 22 ng/mL, a ferritin of 18 ng/mL, a TSH of 3.5 mIU/L, and a fasting insulin of 11 µIU/mL has every value "inside the reference range" by their lab's standards. They are also a person who is statistically very likely to feel tired, foggy, and physically off — and the research supports it.
The markers most commonly "normal" while actually telling you something
Here are the markers where the gap between "normal" and "research-supported optimal" is widest, and the symptom patterns research has linked to the suboptimal ranges.
Ferritin (iron stores)
Most labs flag ferritin as "low" only below 12-15 ng/mL, the threshold for iron deficiency anemia. But peer-reviewed research has consistently shown that ferritin in the 15-50 ng/mL range — formally "normal" — is associated with fatigue, hair loss, restless legs, and poor exercise tolerance, even with a normal hemoglobin and complete blood count[3][4]. A 2017 systematic review found that iron supplementation in non-anemic women with low ferritin (below 50 ng/mL) significantly improved fatigue scores compared to placebo[5]. The optimal range cited in functional medicine literature is typically 50-100 ng/mL for women and 70-200 ng/mL for men.
Vitamin D (25-OH)
Most labs label vitamin D as "sufficient" at 20 ng/mL or above, based on the threshold for bone health. The Endocrine Society's clinical practice guideline distinguishes deficiency (<20 ng/mL), insufficiency (20-29 ng/mL), and sufficiency (≥30 ng/mL), and recommends targeting 30-50 ng/mL for general health[6]. Multiple subsequent meta-analyses have linked vitamin D levels in the 20-30 ng/mL range — labeled "normal" or "sufficient" by many labs — to elevated risk of fatigue, depression, cardiovascular events, and immune dysfunction[7][8].
Vitamin B12
Standard labs flag B12 as deficient only below 200 pg/mL. But research has documented that the 200-400 pg/mL range — formally "normal" — is associated with neurological symptoms including paresthesias, cognitive changes, and depression in a substantial fraction of patients[9]. This led to recommendations from multiple groups to consider 400-500 pg/mL as a more functional minimum, with methylmalonic acid (MMA) as an additional marker to detect functional B12 deficiency[10].
TSH (thyroid)
Most US labs use a TSH reference range of approximately 0.45-4.5 mIU/L. The American Association of Clinical Endocrinologists has previously recommended a tighter functional range of 0.3-3.0 mIU/L for many patients, particularly those with symptoms[11]. Patients with TSH values in the 3.0-4.5 mIU/L range — technically normal — sometimes have early or subclinical hypothyroidism, particularly if they also have elevated thyroid antibodies or symptoms.
Free T3 and Free T4 (active thyroid)
Even when TSH is normal, the active thyroid hormones (free T3 and free T4) can sit in the lower end of their reference ranges and produce hypothyroid-like symptoms. This is sometimes called "low T3 syndrome" or "non-thyroidal illness syndrome." A 2017 review found that low-normal free T3 is associated with metabolic and cardiovascular outcomes that wouldn't be detected by TSH alone[12].
Fasting insulin (insulin resistance)
Most physicians don't routinely order fasting insulin — they order fasting glucose and HbA1c instead. But fasting insulin can rise years before glucose markers shift, signaling early insulin resistance long before diabetes is diagnosable[13]. A "normal" fasting insulin in many labs goes up to 25 µIU/mL, but research-supported optimal ranges are typically 2-7 µIU/mL. The HOMA-IR calculation derived from fasting insulin and glucose is the standard surrogate measure of insulin resistance in published research[14].
Homocysteine
Standard reference ranges typically flag homocysteine as "high" above 15 µmol/L. But peer-reviewed research has linked values in the 9-15 µmol/L range — formally normal — to elevated cardiovascular risk and cognitive decline[15]. The European Society of Cardiology and other bodies have recommended a tighter optimal range below 9-10 µmol/L based on this research.
hs-CRP (high-sensitivity C-reactive protein)
The 1-3 mg/L range is officially "average" cardiovascular risk by AHA/CDC criteria[16]. But the relationship between hs-CRP and cardiovascular events is continuous — every increment higher carries incremental risk, even within the "average" range. A "borderline" hs-CRP of 2.8 mg/L is technically normal but research-linked to meaningfully elevated risk. See our full hs-CRP article for the evidence.
The "subclinical" problem
Modern medicine has a term for the territory between "diagnosable disease" and "asymptomatic optimal function": subclinical. Subclinical hypothyroidism. Subclinical inflammation. Subclinical insulin resistance. Subclinical deficiency. The term describes the situation where a marker is technically inside the reference range but the research has documented real, meaningful associations with symptoms, longer-term outcomes, or both[17].
The medical profession's general approach to subclinical findings has been: "It's inside the range, so there's nothing to do." This is sometimes appropriate — many subclinical findings are not worth treating and would cause more harm than good if addressed aggressively. But it's not always appropriate. When a subclinical marker is associated with the patient's actual symptoms, ignoring it because it's "inside the range" is a failure of clinical reasoning, not its triumph.
Why your doctor doesn't usually do this analysis
This is not a criticism of doctors. Most physicians are working within enormous constraints:
- Time: A typical primary care visit in the US lasts 15 minutes[18]. In Europe it's often 8-12 minutes. There is no time to walk through 20 markers, explain optimal vs normal, and discuss research.
- Training: Medical school teaches reference-range-based thinking. Functional and optimal-range thinking is largely outside the standard curriculum.
- Liability: Diagnosing and treating subclinical findings carries clinical and legal risk. Saying "your labs are normal" is the safe default.
- Insurance and billing: Workups for "normal" labs aren't reimbursed in many systems. There's no billing code for "patient feels off and we're investigating."
- Workload: Most physicians see 20-30 patients a day and can't reasonably do per-patient research synthesis.
The result is a healthcare system that's good at diagnosing overt disease and not designed to address the broad space of "feeling unwell with normal labs." This is the gap Lipa was built to fill — not by replacing your doctor, but by adding the context they don't have time to provide.
What to do if your labs are "normal" but you still feel off
Practical steps the research supports:
1. Take your symptoms seriously
Symptoms are real data. They may not match a textbook diagnosis, and they may not be reflected in your standard panel, but they are a signal worth investigating. The first principle is not to dismiss what you're experiencing because a chart says you're "fine."
2. Look at your specific values within the range
Don't just look at whether your values are flagged. Look at where they sit within the reference range. A ferritin of 22 ng/mL is technically normal but is much closer to the deficiency threshold than to the upper end of normal. A TSH of 3.8 is technically normal but is in the territory where some endocrinologists would investigate further. The position within the range matters.
3. Consider whether your panel is comprehensive enough
Most routine panels test only a fraction of what could explain common symptoms. A patient with fatigue who got only a CBC and TSH has not had iron stores (ferritin), B12, vitamin D, free T3, fasting insulin, or hs-CRP measured — each of which could be the answer.
4. Look at the research for your specific values
This is what Lipa is built to do. For each of your markers, we surface the published research on people in your specific value range, weighted by evidence quality and funding source, and translate it into plain English. You don't have to be a clinician to know what the research says — you just need a tool that does the reading for you.
5. Bring this context back to your physician
An informed conversation is the goal. Bring printouts. Bring the cited research. Ask specific questions: "My ferritin is 22 — what does the research say about supplementing iron at this level for fatigue?" "My TSH is 3.5 with positive TPO antibodies — would you consider this subclinical Hashimoto's?" Most physicians appreciate engaged patients who have done the legwork. The framing matters: not "I think you missed something" but "I read this and wanted your clinical perspective."
How Lipa helps with this specific problem
Lipa was built around the gap between "normal" and "optimal." When you upload a blood test, every marker is analyzed against three things: the standard reference range, the research-supported optimal range, and the symptom and outcome literature for your specific value. We don't tell you you're sick if you're not. We tell you where the research places your specific number — and we show you the studies behind every claim.
For people who walked out of a doctor's office told they were "normal" but knew something was off, this is what they wanted in the first place: not a different doctor, but a way to look at the same labs through a wider lens. Lipa is that wider lens.