• by Sarah Mahoney  @mahoney_sarah, April 19, 2022

Is it skin cancer or psoriasis? Dermatitis or diaper rash? Skin conditions are often missed -- or misdiagnosed --in white people, and the problem is much worse for people of color.

Vaseline is introducing See My Skin, a database aimed at helping people with darker skin -- often entirely invisible in web searches -- find and identify the conditions they are concerned about.

It's part of Vaseline's Equitable Skincare for All Program, in partnership with Hued, a digital health company focused on improving care for people of color, and Visual Dx, an augmented thinking and visualization company.

The project is based on the American Academy of Dermatology's findings that fewer than 6% of image-based search results show conditions on skin of color.

The See My Skin consumer-facing database hopes to solve that problem, connecting people of color with images that can help them find the care they need. The goal is to create 15 million more equitable skincare experiences by 2025.

"People of color are negatively impacted by healthcare inequities that can lead to worse health outcomes, including caring for their skin," says Kevin Tolson, brand director of U.S/ skincare at Unilever, in its release. "Whether it's looking for answers in search, on social media, or even in textbooks -- the results that resemble their skin are scarce, which can leave us feeling unseen and underrepresented."

To raise awareness, the company is also creating limited-edition Vaseline Healing Jelly artwork on some packages, including a QR code that leads people to the See My Skin site. The consumer-facing website steers users to a database of board-certified dermatologists, as well, and offers educational resources. And people can upload their photos to expand the database.

The Unilever-owned brand is also trying to help the medical community, partnering with Medscape to equip dermatologists and medical practitioners to better treat, diagnose and care for skin of color.

Dermatologists have long been aware of the problem, but it has gotten more attention in recent years, much of it sparked by the research of a University of California, San Francisco dermatology professor who was unable to find a single image of what a COVID-19 rash looked like on Black skin.

Another analysis of dermatology textbooks found that only 4% and 18% of images included Black skin.

It's also a problem technology may be intensifying rather than solving. Some research has indicated that machine learning on skin cancer, based predominantly on images of fair skin, may disproportionately misdiagnose skin cancer in people of color.

Google, which says it gets about 10 billion searches each year related to skin, hair and nails questions, developed an AI-powered diagnostic tool that looks for 288 skin conditions. A peer-reviewed study showed results that are just as accurate as those developed by human dermatologists.

Google claims it has fine-tuned the tool with images that take age, sex, race and skin types into account.