Popular Lab Tests That Lack Sufficient Evidence for Routine Use

How widespread unnecessary testing is

Lab overuse isn't rare, and it isn't small. A Washington state study found more than 622,000 patients underwent an unnecessary test or procedure in a single year, at an estimated cost of $282 million. Zoom out to the national picture, and unnecessary care costs the country more than $200 billion a year. That figure covers more than lab work, but blood testing is a big piece of it.

Canada gives a cleaner, lab-specific number. A review there found roughly 22% of blood tests were likely unnecessary. Halfway around the world, a cardiology clinic in Saudi Arabia found 21% of tests ordered over six months didn't meet criteria for being ordered at all. Two different health systems, two different populations, and the number lands in almost the same place. The number landing in almost the same place across two different health systems and populations points to something structural. It means the problem isn't a handful of overzealous doctors in one country, but something structural that produces the same pattern across the whole system.

So why does it keep happening? The research points to a few recurring drivers: unclear or outdated guidelines, heavy workloads that discourage case-by-case judgment, and fear of missing a diagnosis. In emergency medicine specifically, fear of medicolegal consequences was the single most-cited reason physicians gave for ordering tests they didn't think were clinically necessary, at 70.9% in one survey. Add in patients arriving already convinced a specific test matters, because they read about it or a friend mentioned it, and there's no clear stopping point once repeat testing starts.

That last one is the one nobody talks about enough. A Canadian advisory panel on minimum retesting intervals, working with CADTH and Choosing Wisely Canada, flagged repeat ordering of the same test without a new clinical reason as one of the biggest contributors to lab overuse. Not first-time ordering. Re-ordering. A test gets drawn once, comes back fine or ambiguous, and then gets drawn again three months later out of habit rather than need. If you're staring at a chart trying to figure out why you're being tested again, that's the question worth answering, not whether the test happened at all.

None of this means the tests themselves are useless, or that the physicians ordering them are careless. It means the fact that a test got ordered isn't proof it was warranted. Knowing what actually earns a test its place, and what doesn't, is a useful skill for anyone paying for their own labs or trying to make sense of a results page full of flagged values.

Routine vitamin D screening in healthy adults

Most vitamin D testing in this country shouldn't be happening, and the reason traces back to how the reference ranges were built rather than to any real epidemic of deficiency. Depending on which lab or guideline you're reading, "normal" ran anywhere from 12.5 to 60 ng/mL, and plenty of guidelines set the bar for sufficiency at 20 or 30 ng/mL. Set the cutoff that high, and "deficiency" appears in almost every result you look at. That's arithmetic dressed up as pathology.

The Endocrine Society revisited the evidence and, in the August 2024 issue of The Journal of Clinical Endocrinology and Metabolism, published updated guidelines. The changes weren't small tweaks. The Society dropped "sufficiency" and "insufficiency" as clinical categories altogether and recommended against routine 25(OH)D screening in healthy people of any age.

That's a statement against testing before supplementing, not a statement against supplementing. For certain groups, the Society recommends supplementing without a blood draw at all: children aged 1 to 18, adults 75 and older, pregnant women, and people with prediabetes. Harvard Health's reporting on the new guidelines put the recommended dose for these groups at roughly 1,000 IU per day, low and simple. If someone's going to be told to take vitamin D regardless of what the number says, drawing the blood first doesn't change the decision. It just adds a cost and a data point nobody acts on differently.

That doesn't mean the guidance closed the book on the subject. Dark-skinned individuals, preterm infants, and people with obesity are populations where some researchers think a blanket "skip the test" stance moves faster than the evidence allows. One published critique called the absence of any recommendation for preterm and term infants "baffling." An unresolved causal question drives all of it: does low vitamin D actually drive cardiovascular disease, cancer, autoimmune conditions, and infections, or is it just a marker riding along with other risk factors? Associations appear throughout the literature. Causation hasn't been nailed down, and that gap is why testing ahead of the evidence counts as premature, not cautious.

Chronic kidney disease, Crohn's disease, and other conditions that interfere with vitamin D absorption still justify testing. The shift concerns universal screening in otherwise healthy adults. Cases that still call for the test keep it available.

Thyroid panels ordered beyond what symptoms and evidence support

TSH testing, done for the right reason, is genuinely useful. Nobody's arguing otherwise here. The trouble is how often it gets run on people with no symptoms at all, and what gets tacked onto it once the tube is already drawn.

A 2025 Therapeutics Letter from the Therapeutics Initiative at the University of British Columbia lays out a TSH-centered approach and specifically discourages routine thyroid screening in people without symptoms, citing a lack of demonstrated benefit and a real risk of overdiagnosis and overtreatment.

Subclinical hypothyroidism is where this plays out most visibly, and it's the clearest example of a definition manufacturing its own patients. It's defined as an elevated TSH with normal thyroid hormone levels, and no symptoms are required for the label to apply. In Wales, researchers identified 199,520 individuals (63.8% female) with subclinical hypothyroidism between 2000 and 2021, and 23.6% of them ended up on levothyroxine over the study period. Whether that treatment rate reflects sound clinical judgment or systematic overtreatment is still argued over, and there's a mechanical reason the debate won't die: narrow the TSH cutoff used to define the condition, and more people fall outside the normal range by definition alone. A line moves on a page, and thousands of people wake up with a diagnosis they didn't have the day before.

Then there's the issue of expanded panels, and here the evidence is unambiguous rather than murky. Free T3, reverse T3, and thyroid antibody tests get ordered constantly without a clear reason tied to symptoms. The 2025 Therapeutics Letter is specific: free T4 is generally warranted after an abnormal TSH rather than ordered alongside it by default, and total T4 has limited support for initial hypothyroidism workups. Reverse T3 and routine antibody testing in people with no symptoms lack backing from major clinical guidelines. These tests appear constantly in direct-to-consumer testing and functional medicine settings, often without anyone able to say what clinical question they're supposed to answer.

Pregnancy adds another wrinkle. Universal thyroid screening during pregnancy also lacks sufficient evidence to support routine use, and a meta-analysis of three randomized trials found no evidence that treating subclinical hypothyroidism diagnosed during pregnancy with levothyroxine improved obstetrical outcomes, neonatal outcomes, or childhood IQ and neurodevelopment. That's a striking result given how often the test gets framed as protective.

TSH earns its place when symptoms or real risk factors are present. Anything beyond that needs its own specific question behind it, something more deliberate than a reflexive add-on riding along because the tube was already drawn.

Salivary cortisol and "adrenal fatigue" panels

"Adrenal fatigue" gets marketed as a kind of chronic, low-grade adrenal exhaustion caused by ongoing stress, and it appears constantly in wellness testing panels. It shouldn't be ordered, and the Endocrine Society's position is about as direct as medical language gets: there is no test that detects adrenal fatigue, because adrenal fatigue is not a recognized medical diagnosis. A systematic review in BMC Endocrine Disorders went looking for evidence to support it anyway, examined 58 qualifying studies, and came up empty.

Why doesn't cortisol testing work here, specifically? Cortisol moves throughout the day, and it moves further in response to short-term stress, sleep, exercise, even the time you happen to spit in the tube. An endocrinologist quoted in a Medscape article said the testing "does not work because it fluctuates too much from time to time" and is "not reliable or reproducible." Salivary cortisol profiles paired with symptom-scoring questionnaires have never been validated as tools for assessing how the hypothalamic-pituitary-adrenal axis actually functions. Run the validated tests instead, serum cortisol, plasma ACTH, an ACTH stimulation test, and the results on fatigued patients usually come back normal, unless the person has already been taking steroid-containing supplements that muddy the picture.

Persistent fatigue affects a lot of people and has plenty of identifiable causes: sleep apnea, insomnia, thyroid disease, depression, anxiety, iron deficiency, B12 deficiency, and true adrenal insufficiency, better known as Addison's disease. Adrenal fatigue as a diagnosis isn't one of them. Addison's is legitimate and diagnosable through blood work and ACTH stimulation testing, run by an endocrinologist who knows what to look for.

The real risk of adrenal fatigue testing is delay, plain and simple. Chasing a diagnosis that doesn't exist pushes off the workup that would catch something that does, and in some cases exposure to steroid-containing supplements sold alongside these panels complicates the legitimate testing that comes later.

IgG food sensitivity panels

IgG food sensitivity tests get sold as a way to pin down which foods are behind fatigue, bloating, brain fog, or skin flare-ups. The pitch sounds reasonable. The biology behind it doesn't hold up: these panels measure a normal immune response and sell it back as a diagnosis.

IgG antibodies to food antigens are an expected result of eating that food, not a sign of harm. If a panel shows an elevated IgG reaction to, say, eggs or wheat, the most likely explanation is that the person eats eggs or wheat regularly. That's the entire mechanism, and it's why the major allergy and immunology organizations don't back this kind of testing. The American Academy of Allergy, Asthma, and Immunology, the European Academy of Allergy and Clinical Immunology, and the Canadian Society of Allergy and Clinical Immunology all decline to endorse it for diagnosing food allergies or sensitivities.

A joint consensus report from the AAAAI and EAACI listed allergen-specific IgG and IgG4 testing among the modalities "lacking sufficient evidence to recommend their use." A 2025 American College of Gastroenterology guideline on eosinophilic esophagitis went further, recommending against using available allergy testing, serum IgG included, to guide food elimination diets. The guideline found that diets built around allergy test results did not outperform alternatives in the meta-analyses it reviewed.

Most of the underlying research on IgG testing comes from small, single-center studies with real methodological problems: no blinding, no validated reference standard, no demonstrated improvement in how patients actually feel. For someone using one of these panels to guide their diet, the practical risk isn't hypothetical. Elimination diets built on IgG results tend to cut out foods without fixing the original symptom, and they build genuine anxiety around eating with no biological basis behind it.

Homocysteine testing for cardiovascular risk

Homocysteine has a plausible enough story to make testing tempting. It's tied to endothelial dysfunction, oxidative stress, inflammation, and clot formation, mechanisms that connect logically to cardiovascular disease. The observational data back the association up: systematic reviews and meta-analyses have found that rising plasma homocysteine tracks with meaningfully higher risk of coronary heart disease. A more recent umbrella review confirmed links to several cardiovascular outcomes, with stroke among the more consistently reported findings.

So where's the catch? Lowering homocysteine with B-vitamin supplements hasn't reliably reduced actual cardiovascular events in intervention trials. The number moves. The outcome doesn't follow it down, and that gap is the whole story here: an association that looks solid in observational data doesn't automatically mean the pathway is causal in a way testing and treating can act on. Ordering this test on a routine basis doesn't make sense given what's known right now.

Insurance coverage reflects that gap in practice, and this is where it stops being an academic argument. At least one major insurer doesn't cover homocysteine testing for cardiovascular risk assessment, working from Avalon's 2025 policy criteria. A payer looked at the evidence for clinical utility and decided it doesn't clear the bar, because paying to lower a number that doesn't move outcomes is money spent on nothing.

The broader lesson applies well past homocysteine: a biomarker can be strongly associated with disease, mechanistically sensible, and still not worth testing routinely, if acting on the result doesn't move the outcome that actually matters.

Hs-CRP placement, a contrast case of selective rather than routine use

hs-CRP deserves a different conversation than everything above it. It's a real inflammatory marker with genuine risk stratification value, and it doesn't belong in the same bucket as adrenal fatigue panels or IgG food testing. What separates it is entirely who it's used on, and why.

A 2025 analysis looked at hs-CRP's role in identifying residual inflammatory risk in patients who already have established cardiovascular disease, who've hit their LDL cholesterol targets, and who remain at elevated risk anyway. That's a narrow, specific clinical question aimed at a population where the answer changes what happens next.

Run the same test as part of a general wellness panel on someone with no cardiovascular symptoms or risk factors, though, and the value collapses to near nothing. hs-CRP rises for reasons that have nothing to do with heart disease: an active infection, an autoimmune flare, obesity. An isolated elevated result on a low-risk person is hard to interpret and rarely points anywhere useful on its own.

What separates the two uses is the question behind the order, which shapes what the test result means, more than the test itself. One is a specific clinical question about residual risk in a statin-treated ASCVD patient. The other is a general hope that the results will turn up something interesting. Same test, same lab tube, completely different value depending on who's holding it and why.

Weighing whether a given test is worth ordering

Run back through the list: vitamin D screening in healthy people, expanded thyroid panels, adrenal fatigue testing, IgG food sensitivity panels, homocysteine, hs-CRP ordered without a clinical reason. Every one of them fails at the same point: clinical utility. The result doesn't reliably lead to an action that actually improves what happens to the person being tested.

A few questions cut through most of this before a test ever gets ordered. Will the result change what happens next? If the plan is to supplement, treat, or monitor regardless of the number, the test is adding a data point nobody's using. Does the marker predict something that actually matters clinically, or does it just correlate with it in observational data? Has acting on an abnormal result been shown to improve outcomes, not just move the number itself? And does the benefit of finding something outweigh the cost of chasing false positives, unnecessary treatment, and the anxiety that comes with an ambiguous flag on a lab report?

None of this argues against testing. It argues for testing with a reason behind it, decided before the blood gets drawn rather than after. The most expensive lab work is the panel ordered on a hunch, the one that comes back with a number nobody quite knows what to do with. It's the one ordered on a hunch, the one that comes back with a number nobody quite knows what to do with, and sets off three more appointments trying to figure out whether it meant anything at all.