A Task Force Flagged This in 1995. The Definition Still Leads With Redness.

This one is not about beauty, and I am putting it here anyway, because it is the clearest example in medicine of an instrument built around a sign that some skin does not show.

A pressure injury — a bedsore — is staged by what it looks like. Stage 1, the one you want to catch, is defined as intact skin with a localized area of non-blanchable erythema.

Redness that does not go pale when you press it.

The clause

The current definition, from the 2016 revision, reads in full: intact skin with a localized area of non-blanchable erythema, which may appear differently in darkly pigmented skin. Presence of blanchable erythema or changes in sensation, temperature, or firmness may precede visual changes. Color changes do not include purple or maroon discoloration; these may indicate deep tissue pressure injury.

Eight words. Which may appear differently in darkly pigmented skin.

It does not say the criterion under-detects. It does not quantify by how much. It does not offer an alternative criterion. The non-visual signs it names — sensation, temperature, firmness — are described as things that may precede visual changes, which is to say adjuncts to looking, not substitutes for it.

Now the date. A task force was appointed specifically to review the adequacy of the Stage 1 definition in people with darkly pigmented skin, and reported in 1995. A draft replacement definition was published in 1997.

The revision that produced the current wording came out in 2016. Twenty-one years after that task force reported, Stage 1 is still defined by non-blanchable erythema, and the pigmented-skin problem is still handled by a subordinate clause.

There is also a precise record of the panel arguing about it. Buried in the consensus conference results is a vote on removing the color-description clause — the one about purple and maroon discoloration — from the Stage 1 description. Yes 44 percent, no 56 percent, 339 votes cast, consensus not achieved.

That is a professional body, on the record, failing to agree on the exact sentence that is hardest to apply on dark skin.

Somebody checked whether the descriptors hold

A five-year retrospective study looked at 274 Stage 1 and suspected deep tissue injuries across 96 patients of African American, Asian, Hispanic, American Indian, Alaska Native, Native Hawaiian or Pacific Islander descent, using the documentation of nine wound care nurses.

Contrary to the guidelines, the blanching effect was commonly present. Stage 1 injuries presented with normal color in seventeen instances. And suspected deep tissue injuries — which the definition describes as occurring under intact skin — had a break in the skin in 26 of them, 14 percent.

The authors’ conclusion is that the definition may benefit from revision to improve accuracy among persons with darkly pigmented skin. Their denominators as published do not reconcile cleanly, and the full text is not open, so I am reporting the counts they give rather than recomputing percentages from them.

Eleven years earlier, a spectroscopy study of 76 patients and 95 site pairs had already found that erythema in subjects with dark skin was more likely to be non-blanching, with significant spectrographic differences across pigmentation.

Its recommendation, in 2003: clinicians should use persistence of erythema rather than blanching status to judge incipient pressure ulcers.

The 2016 definition leads with blanching status.

The disparity, and the explanation that turns out to be incomplete

The numbers are not in doubt. A study of 2.1 million white and 346,808 Black residents across 12,473 US nursing homes found pressure ulcer rates of 16.8 percent for Black residents against 11.4 for white residents in 2003, and 14.6 against 9.6 in 2008.

Both rates fell. The gap did not — the trend in the disparity was not significant across six years of national quality improvement.

Then the finding that complicates the whole story. In nursing homes where at least 35 percent of residents were Black, the white residents had an adjusted odds ratio of 1.33 for pressure ulcers compared with white residents in homes serving primarily white populations.

White residents. In the same buildings. A third more likely.

That is not a skin-visibility effect. It is a site-of-care effect, and a New York state analysis of 59,740 residents found the same thing more sharply: Black residents were more likely to have pressure ulcers overall, and that effect vanished entirely once facility fixed effects were added. Their conclusion is that the excess may not result from differential within-facility treatment, but from Black residents being more likely to reside in facilities with poorer care quality.

Anyone who explains this disparity purely as an inability to see redness on dark skin is skipping the largest effect in the data, which is which building you are in.

And then the prospective cohort that saw the other thing happen

Except the visual problem is also real, and one study watched it occur.

The Pressure Ulcer Detection Study followed 270 pressure injuries across 142 racially and ethnically diverse nursing home residents, assessed weekly for sixteen weeks with a standardized wound tool.

Black and Hispanic individuals had normal skin color surrounding their pressure injuries.

A wound is present. The skin around it reads as normal-colored. That is the visual criterion failing at the bedside, observed directly rather than inferred from billing codes. Black residents had more persistent trunk injuries and more Stage 4; Black and Hispanic residents had more deep tissue injury; Black residents’ heel injuries were more often unstageable and necrotic.

And, critically for the argument above: no nursing home or prevention differences existed within this cohort.

So both mechanisms are real. Where you are cared for drives a large share of the gap, and the sign the system is built on does not appear the same way on everybody. A field that keeps arguing about which one it is has two problems, not one.

The pattern in hospital data is what you would expect from either. A 2018 New Jersey inpatient analysis of 17,781 patients with a pressure injury diagnosis found Black patients had a higher proportion of Stage 4 injuries and a lower proportion of Stage 1. More of the worst, less of the earliest. The authors’ own closing line is that clinical research examining the impact of skin tone rather than racial group is needed.

The study that finally measured skin tone

Almost nobody in this literature has measured skin tone. They have used race as a stand-in for it, which is precisely the substitution this whole section exists to object to.

A 2025 international prevalence analysis used the Monk Skin Tone Scale instead. Medium-to-dark skin tones carried a two- to fourfold relative risk of presenting with a Stage 2 or Stage 3-plus injury rather than a Stage 1.

Its authors state the problem exactly: decades of research show that individuals who are Black experience higher-stage pressure injuries, investigators primarily attribute this to difficulty visualizing early injury in dark skin tones, yet few have actually measured skin tone.

The published abstract of that paper does not state its sample size. The single most important skin-tone-measured study in the field, and the n is not in the record.

The instruments proposed to fix it

Sub-epidermal moisture scanning is the leading candidate, and the logic is good: it measures the electrical properties of tissue water rather than light, and melanin does not change capacitance.

The foundational evidence that it works in dark skin is a 2009 pilot titled for exactly that — subepidermal moisture and early pressure ulcer damage in nursing home residents with dark skin tones. It pooled data from two previous studies to look at the question after the fact.

Its cohort was 66 residents, of whom 77 percent were non-Hispanic white.

Which puts the dark-skin subgroup at fifteen people at most, and the abstract never states what it actually was. The effect sizes it reports are large — at a threshold of 150 dermal phase units, residents with dark skin tones were thirteen times more likely to present with a Stage 2 or worse injury the following week — and the authors end by saying these findings should be further evaluated in persons with dark skin tones.

A later study by the same group, in 417 residents of whom over half were minority, reported that the scanner predicted 41 percent of future skin damage against 27 percent for visual ratings, and asserted in its abstract that the method is objective and equally valid for all groups of patients. No stratified analysis by race or skin tone appears in that abstract to support the claim.

Then the registration study — blinded, prospective, twelve sites, 182 patients in the analysis. Sensitivity 87.5 percent. Specificity 32.9 percent. Area under the curve 0.6713.

Read the specificity again. Among patients who did not go on to develop a pressure injury, the scanner flagged roughly two in three anyway. The authors are careful — they say the device can complement standard assessment, not replace it, and they flag that the standard-of-care interventions given to flagged patients may have confounded the specificity, which is a fair point.

That cohort’s race, ethnicity and skin tone are not reported at all.

The result most quoted as proof it works in dark skin is a quality improvement project: 90 patients before, 50 after, eight hospital-acquired injuries falling to zero, reported as a 100 percent reduction. With fifty patients and a baseline rate of 8.9 percent you would expect about four or five events, and observing none in an uncontrolled before-and-after is entirely compatible with chance or with everyone paying more attention during a project. It is not evidence that a device eliminates bias.

And a note on the word validation. In this literature it attaches most firmly to an ex vivo porcine model and to synthetic phantoms. Both are legitimate bench work. Neither has a skin tone, and a phantom has no melanin at all.

The thermal camera has its own problem

Thermography should be the clean answer: emissivity in the long-wave infrared is essentially independent of melanin, because melanin absorbs in the visible, not at eight to fourteen micrometers.

A 2025 study inverted the usual recruitment failure — 35 adults, thirty of them at Monk 6 or above — and cooled one hip with a stone for five minutes to create a known temperature difference.

Cooling was detected across all skin tones. The difference between the lightest and darkest groups was 0.98 degrees.

The difference between the two cameras was 1.04 degrees.

The instrument-to-instrument variation exceeded the skin-tone variation. Both are the same size as the effect being hunted, which means a threshold calibrated on one device does not transfer to another, and the authors say tailored thresholds may be necessary for darker skin tones.

A companion study induced erythema by cupping and measured it three ways at once. The correlation between temperature change and erythema change was weak — a thermal camera is not a redness detector with the melanin problem removed, it is measuring something else. And the colorimeter, the objective instrument brought in to replace the human eye, consistently read less erythema in participants with dark skin tones.

The instrument inherited the problem it was hired to solve. That conclusion is worth sitting with, and it belongs alongside the wider case about instruments.

The most enthusiastic thermography result — a blinded prospective cohort in which the device flagged every deep tissue injury before it was visible — rests on four events across 131 anatomical areas, and reports no specificity, so we are not told how many of the 127 areas that never progressed were also flagged.

What the field says about itself

Its own 2025 state-of-the-art review inventories four clinical devices and three research ones, and concludes that current evidence did not support replacing standard visual inspection and palpation — only augmenting it. It also notes that there is no reimbursement available for early detection of pressure injury, which is the practical reason none of this is at the bedside regardless of the evidence.

A 2026 scoping review searched five databases across nine years for any primary study on identifying pressure injuries in dark-skin-tone patients. Ten studies met inclusion. The reviewers conducted no appraisal of them, and observed that routine bedside visual and tactile assessment — what actually happens to every patient, every shift — was notably underrepresented.

And the most direct attempt to validate a skin-assessment tool specifically in dark skin, a 2026 pilot using a modified color bar, could not test its primary aim because only two patients developed an injury and neither presented as a subtle color change. Its incidental finding is the one worth keeping: race and ethnicity were not an objective marker for determining skin tone by the tool.

Inside the pressure-injury literature, a direct empirical rebuttal of the substitution the rest of that literature is built on.

If somebody you love is in a bed

There is a non-technological approach in the literature built on four things you can check without a device: skin tone, temperature, texture, and twinge — meaning pain. No accuracy data accompanies it, and I would not present it as validated. But temperature and firmness are named in the official definition itself as signs that may precede visual change, and they cost nothing.

So: the back of your hand on the sacrum and the heels, feeling for a patch that is warmer or cooler than the skin beside it. A patch that feels boggy or unusually firm. And asking about pain at a specific spot, which in the descriptor study accompanied Stage 1 injuries that looked like nothing.

Ask what the heels are resting on. In the detection study, offloading happened with pillows for three quarters of residents and with proper heel boots for a fifth.

And know that if the answer to a concern is that the skin looks fine, that sentence is carrying the entire weight of a criterion which its own governing body could not reach consensus on, in skin its own task force flagged in 1995.

The same visual problem attaches to another chronic wound with a far thinner evidence base. Six randomized trials exist for sickle cell leg ulcers, totalling 198 people, and not one of them recorded time to healing, quality of life or harms.

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.