The Eczema Severity Score Is Mostly Measuring Redness

Two numbers from the Project Viva cohort, a prospective pre-birth study in eastern Massachusetts, and the gap between them is the point.

Incidence of early-childhood atopic dermatitis, 1,437 children. Adjusted odds compared with non-Hispanic white children: non-Hispanic Black children, 2.71 — confidence interval 1.75 to 4.19. Other non-Hispanic, 1.80. Hispanic children, 1.10, which was not statistically significant.

Then persistence into mid-childhood, among the 392 children with follow-up data. Non-Hispanic Black children: adjusted odds ratio 6.26, confidence interval 2.32 to 16.88. Hispanic children: 6.42, confidence interval 1.93 to 21.41.

So the disparity in getting eczema is roughly two-and-a-half fold. The disparity in still having it years later is more than six.

Notice that Hispanic children were not significantly more likely to develop it and were more than six times as likely to still have it. Whatever is driving persistence is not the same thing that drives onset, and the coverage that says “eczema is more common in Black children” collapses both into one sentence and loses the more alarming half.

What SCORAD and EASI actually weigh

Eczema severity is measured with one of two instruments, and both were built around the way the disease looks in fair skin.

SCORAD scores six intensity items, one of which is erythema. EASI scores erythema alongside induration, excoriation and lichenification across four body regions. In both, redness is a weighted component of the number that decides whether a child’s eczema is mild, moderate or severe. I have not been able to put a figure on that weight for either instrument, and I am not going to estimate one — the sibling piece on psoriasis can say one third because the index publishes it; these two do not.

Erythema in richly pigmented skin does not read as red. It reads as violaceous, dusky, gray-brown, or simply as darker skin — the same optical masking that hides rosacea, the same mechanism, a different disease.

A published review in the field makes exactly this argument: erythema affects outcome measures in atopic dermatitis in patients with skin of color. It exists, by that name, making that case.

The workaround in circulation is an instruction to add one point to the erythema score in darker skin. That adjustment is explicitly unvalidated. Somebody noticed the instrument under-reads and proposed a constant to correct it, and nobody has established that a constant is the right shape of correction, or that one point is the right size, or that it behaves the same across the range.

Why an under-reading score is not an academic problem

Because severity thresholds are gates.

Moderate-to-severe atopic dermatitis is the indication for the treatments that changed this disease — dupilumab, the JAK inhibitors. Getting to them requires a severity score, an insurer, and frequently a prior authorization that names a number.

A child whose erythema cannot be seen scores lower. Lower is milder. Milder is topical steroids and come back in six months.

The disparity literature is catching up to this. There are published analyses of racial and ethnic disparities in dupilumab access for pediatric atopic dermatitis, and of disparities in access to advanced therapies generally, and of immune signaling differences by self-reported racial group that connect directly to how patients respond. The odds ratios in several of them I could not reach, so they are absent from this page rather than approximated. What is established is that multiple independent groups have published on the existence of the gap.

And there is a phenotype problem sitting on top of the scoring problem. Atopic dermatitis in Black and Asian children frequently presents with follicular accentuation — small bumps centered on hair follicles — with a papular rather than weeping morphology, and more lichenification, the thickened leathery skin that comes from chronic scratching. The classic description a clinician learned is red, weeping, oozing patches. That is a different-looking disease.

The numbers that are not on the scale

Bring those instead. Nights of disturbed sleep per week. School or nursery days missed. How long the flare lasted. Whether there is bleeding from scratching. Whether a topical steroid course cleared it and how fast it came back.

Those are patient-reported and duration-based, and they do not pass through the melanin filter. They are also, frankly, better measures of whether a child’s life is being wrecked by this than a grader’s estimate of how red a patch looks.

Photograph flares in daylight, at their worst, not at the appointment three weeks later when it has settled. Eczema has excellent timing that way.

And if you are told it is mild while your child is awake scratching four nights a week, those two statements are in conflict and the second one is the measurement that was actually taken.

The same instrument problem has now been documented in psoriasis, where the severity index weights redness as a third of the score — and in the one trial designed around skin of color, the pigment left behind tracked quality of life better than the disease score did. The genetics underneath eczema diverge too: the standard barrier-gene panel tests four European variants, and sequencing African American children turned up nine others.

This article is educational and is not a substitute for care from someone who can look at your skin. Anything new, changing, painful or spreading should be assessed rather than researched.