Representation of skin color in dermatology-related Google image searches

Kurtti, Alana; Austin, Evan; Jagdeo, Jared

doi:10.1016/j.jaad.2021.03.036

Cited by 11 publications

(6 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Google, during SSE prior to seeking medical care.2 Previous studies searching Google Images for various dermatologic conditions found only 5.7% of total search results displaying darker skin tones. 3 This study complements previous findings by doubling the images analyzed and focusing on skin cancers. Additionally, this study incorporated the 5-tone Pantone swatch 4 instead of the Fitzpatrick scale, allowing for a more targeted categorization of darker skin tones rather than misuse of a grading system intended to measure skin response to UV light.…”

Section: Introduction Ssupporting

confidence: 81%

The Representation of Skin Tones in Google Images of Skin Cancers

Nasseri,

Sadur,

McCormick

et al. 2024

JDD

View full text Add to dashboard Cite

algorithm bias. Each reviewer assessed the first 100 images for the following searches: "Skin Cancer" (SC), "Melanoma" , "Squamous Cell Carcinoma" (SCC), and "Basal Cell Carcinoma" (BCC). Each image was designated as "Lighter Skin Tone" , "Darker Skin Tone" , "Light and Dark Skin Tone" , or "Inconclusive/ Other" based on the published 5-tone Pantone swatch. 4 "Lighter Skin Tones" included Pantones A-B or lighter while "Darker Skin Tones" included Pantones C-E or darker. Images where skin tone could not be differentiated were deemed inconclusive. Pantone shade for each image of darker skin tone was documented, along with the image number in the search; final results were compiled and averaged. RESULTSDarker skin tones were infrequently encountered in the first 100 Google Image results for each search term. 1200 pictures were evaluated, of which 95% (n=1140) depicted only lighter skin tones, 1.25% (n=15) displayed only darker skin tones, and

show abstract

Section: Introduction Ssupporting

confidence: 81%

The Representation of Skin Tones in Google Images of Skin Cancers

Nasseri,

Sadur,

McCormick

et al. 2024

JDD

View full text Add to dashboard Cite

show abstract

“…This patient's history was consistent with DRESS syndrome with allopurinol being the likely trigger; her diagnosis was further supported by meeting Bouquet's criteria 1 . The lack of diversity in skin tones in dermatology images has been well documented and is thought to contribute to health care disparities 2 . The images of this patient present an opportunity for clinicians to identify DRESS syndrome in a patient with brown skin and prevent the delayed diagnosis of an emergent dermatologic condition in future patients.…”

Section: Diagnosismentioning

confidence: 57%

“…1 The lack of diversity in skin tones in dermatology images has been well documented and is thought to contribute to health care disparities. 2 The images of this patient present an opportunity for clinicians to identify DRESS syndrome in a patient with brown skin and prevent the delayed diagnosis of an emergent dermatologic condition in future patients.…”

Section: Diagnosismentioning

confidence: 98%

Maculopapular rash in brown skin

Allen,

Co,

Kellar

et al. 2023

JACEP Open

View full text Add to dashboard Cite

“…To the Editor: We read with interest the recent paper, ''Representation of skin color in dermatology-related Google image searches,'' published by Kurtti et al 1 The authors analyzed the breakdown of skin type in Google image search results for myriad skin conditions and terms. Their characterization of the underrepresentation of skin of color in top search results online reinforces the relative lesser ability of dermatologists to diagnose conditions in darker-skinned patients.…”

mentioning

confidence: 99%