Mail-In Blood Panels: What a Drop of Blood Can and Can't Measure

The full story

The pitch is a laboratory in an envelope. Prick a finger, blot a card, drop it in the post, get a PDF. For some analytes that pitch is entirely true. For others it is a number with error bars wide enough to drive a category change through. The box does not distinguish between the two, because the box is selling a panel while the validation happens one analyte at a time.

That is not a consumer-review complaint, it is what the dried blood spot literature says about itself. A review in Current Environmental Health Reports notes that spots are minimally invasive, cheap to collect, transport and store - and that dried blood spot assays require additional validation for accuracy and reproducibility analyte by analyte. A methods paper in Psychoneuroendocrinology makes the same argument from the other end: the decision to substitute a spot for a venous tube should turn on whether it is analytically sensible and feasible to measure that specific parameter in whole blood from a spot, because spot assays have to be adapted from serum and plasma protocols one analyte at a time.

The specimen is the weakest link

A study in the American Journal of Clinical Pathology took six successive drops from a single fingerprick in 11 donors and compared their variability with venous controls. The coefficient of variation was up to 3.4 times higher for haemoglobin, 5.7 times for total white cells, 7.7 times for granulocytes and 4 times for platelets. Fluctuations only fell within instrument variability once at least 60 to 100 microlitres of fingerprick blood had been pooled. A single drop is not a small sample of your blood. It is a sample of that drop.

Then there is the physics of filter paper. Two reviews - one in Therapeutic Drug Monitoring, one in Current Environmental Health Reports - converge on the haematocrit effect: blood viscosity changes how far a spot spreads across the paper, which changes what a punch of fixed diameter contains. Add spotted volume and punch site, and there are three sources of method uncertainty stacked up before the analyser is switched on. Volumetric microsampling devices are designed to collect a fixed dried volume and sidestep exactly this, but clinical implementation is progressing slowly and more analytical and clinical evaluation is needed.

What survives the post

Vitamin D is the success story of this field. In the Journal of Clinical Endocrinology and Metabolism, 62 participants from the Melbourne Collaborative Cohort Study gave matched plasma and dried blood spot samples, and LC-MS/MS measurements of 25-hydroxyvitamin D agreed well, with an R-squared of 0.73 regressing plasma on spot. Across 500 replicates, within-batch and between-batch intraclass correlations were 0.82 and 0.73. One detail from that paper deserves more attention than it gets: a calibration model was required to convert spot values into plasma-equivalent numbers comparable with clinical reference ranges. That conversion is invisible to the customer and essential to the result.

Vitamin D also survives the logistics. In Annals of Clinical Biochemistry, 40 outpatients with a median age of 78 gave paired capillary spot cards and venous samples that were then put through storage conditions mimicking shipment. Spot vitamin D did not differ significantly from venous plasma across those conditions; correlation was best at 21C or 35C for four days, at 0.906 and 0.903, and freeze-thaw had no significant effect.

Proteins do better than you would guess. Multiple-reaction-monitoring mass spectrometry assays for spots, reported in Molecular and Cellular Proteomics, achieved a median total coefficient of variation of 18% at the lower limit of quantification across 245 proteins and a median inter-day coefficient of variation of 8% over four days. Concentrations stayed stable at ambient laboratory temperature for up to two months for 190 of those proteins, which means cold-chain shipping can be dropped for that set. Note the two figures: 18% at the bottom of the measurable range, 8% day to day in the middle of it.

What does not travel

HbA1c is the cleanest demonstration that format beats analyte. A systematic review in Diabetic Medicine collated 30 studies of postable self-collected capillary samples against venous blood. Capillary micro or capillary tubes gave mean differences of minus 0.4 to 1.4 mmol/mol, with limits-of-agreement widths of 2.4 to 6 mmol/mol. Dried blood spots, for the same analyte through the same post, gave mean differences of minus 4.3 to 7.2 mmol/mol with limits of agreement 11.7 to 16.8 mmol/mol wide. Same molecule, same courier, agreement several times worse depending only on what the kit asked you to bleed onto. Users were content either way - 83% to 96% reported satisfaction with self-collection - which tells you that satisfaction and analytical agreement are unrelated variables.

Fingerstick point-of-care chemistry is similarly uneven within a single panel. A diagnostic-accuracy study in Cien Saude Colet compared fingertip point-of-care measurement with venipuncture laboratory reference methods in Brazilian quilombola adults. Agreement was substantial for triglycerides, poor for total cholesterol and glucose, and the Bland-Altman analysis concluded the fingertip method did not produce good measures. Three analytes, one drop of blood, one panel, one that held up.

What the evidence does not show

It does not show that mail-in testing is junk. Vitamin D by mass spectrometry from a spot is a genuinely good measurement, spot protein assays are reproducible, and capillary-tube HbA1c lands close to venous. What the evidence also cannot show is that any panel is trustworthy in general, because "in general" is not how validation works here - and nothing in this evidence base validates a bundled consumer panel as a unit. Every analyte on the requisition is a separate claim.

The regulatory floor is lower than the marketing implies, too. CMS regulates all laboratory testing except research performed on humans in the United States through the Clinical Laboratory Improvement Amendments, covering approximately 320,000 laboratory entities, and certification is required for a lab to receive Medicare or Medicaid payment. That is a statement about how a laboratory operates, not evidence that a given test predicts anything about you. The FDA reviews direct-to-consumer tests intended for moderate-to-high-risk medical purposes for analytical and clinical validity, but states that in general direct-to-consumer tests for non-medical, general wellness or low-risk medical purposes are not reviewed before they are offered, and that such tests are not a substitute for a traditional health care evaluation. A review in EJIFCC names the gap directly: advertising claims of medical usefulness and quality regulations are tightly controlled inside healthcare and essentially absent in direct-to-consumer testing, and results arrive without medical interpretation - a risk for the consumer and for the clinician later handed the printout.

Order of operations

• Pick the analyte first, then check the format. A capillary tube or a volumetric microsampling device is a different specimen from a classic filter-paper spot, and for HbA1c the difference was several-fold in limits of agreement.
• Ask whether spot results are calibrated to plasma equivalents. Without that conversion step, a spot number and a clinical reference range are two different scales.
• Ask whether that specific analyte has a validated spot assay. Not the panel. The analyte.
• Read the shipping window. Vitamin D tolerated four days at 21C or 35C and shrugged off freeze-thaw. That tolerance was demonstrated for vitamin D, not for whatever else is on your card.
• Anchor anything consequential to a venous draw. Mail-in kits are strongest as repeat measurements of analytes with validated spot assays, weakest as the only measurement of something you intend to act on.

The useful mental model is not "home tests versus real tests." It is that a blood test is a chain - specimen, volume, transport, assay, calibration, reference range - and mail-in kits swap out three links of that chain at once. For vitamin D, somebody has done the work of proving the swap holds. For a lot of what gets bundled onto a wellness panel, nobody has, and the PDF looks identical either way. Read the kit as a set of separate tests that happen to share an envelope, and you will get roughly the right amount of confidence out of each one.