Health equity assessment of machine learning performance (HEAL): a framework and dermatology AI model case study.

Background: Artificial intelligence (AI) has repeatedly been shown to encode historical inequities in healthcare. We aimed to develop a framework to quantitatively assess the performance equity of health AI technologies and to illustrate its utility via a case study. Methods: Here, we propose a methodology to assess whether health AI technologies prioritise performance for patient populations experiencing worse outcomes, that is complementary to existing fairness metrics. We developed the Health Equity Assessment of machine Learning performance (HEAL) framework designed to quantitatively assess the performance equity of health AI technologies via a four-step interdisciplinary process to understand and quantify domain-specific criteria, and the resulting HEAL metric. As an illustrative case study (analysis conducted between October 2022 and January 2023), we applied the HEAL framework to a dermatology AI model. A set of 5420 teledermatology cases (store-and-forward cases from patients of 20 years or older, submitted from primary care providers in the USA and skin cancer clinics in Australia), enriched for diversity in age, sex and race/ethnicity, was used to retrospectively evaluate the AI model's HEAL metric, defined as the likelihood that the AI model performs better for subpopulations with worse average health outcomes as compared to others. The likelihood that AI performance was anticorrelated to pre-existing health outcomes was estimated using bootstrap methods as the probability that the negated Spearman's rank correlation coefficient (i.e., "R") was greater than zero. Positive values of R suggest that subpopulations with poorer health outcomes have better AI model performance. Thus, the HEAL metric, defined as p (R >0), measures how likely the AI technology is to prioritise performance for subpopulations with worse average health outcomes as compared to others (presented as a percentage below). Health outcomes were quantified as disability-adjusted life years (DALYs) when grouping by sex and age, and years of life lost (YLLs) when grouping by race/ethnicity. AI performance was measured as top-3 agreement with the reference diagnosis from a panel of 3 dermatologists per case. Findings: Across all dermatologic conditions, the HEAL metric was 80.5% for prioritizing AI performance of racial/ethnic subpopulations based on YLLs, and 92.1% and 0.0% respectively for prioritizing AI performance of sex and age subpopulations based on DALYs. Certain dermatologic conditions were significantly associated with greater AI model performance compared to a reference category of less common conditions. For skin cancer conditions, the HEAL metric was 73.8% for prioritizing AI performance of age subpopulations based on DALYs. Interpretation: Analysis using the proposed HEAL framework showed that the dermatology AI model prioritised performance for race/ethnicity, sex (all conditions) and age (cancer conditions) subpopulations with respect to pre-existing health disparities. More work is needed to investigate ways of promoting equitable AI performance across age for non-cancer conditions and to better understand how AI models can contribute towards improving equity in health outcomes. Funding: Google LLC.

Qualitative exploration of melanoma awareness in black people in the USA.

OBJECTIVE: Although black patients are more likely to have advanced melanomas at diagnosis, with a 5-year survival rate among black patients of 70% compared with 92% for white patients, black people are generally not the focus of melanoma public health campaigns. We sought to explore awareness and perspectives of melanoma among black people to inform the development of relevant and valued public health messages to promote early detection of melanoma. DESIGN: Inductive thematic analysis of in-depth semistructured interviews. SETTING: Interviews were conducted with participants via video software or telephone in the USA. PARTICIPANTS: Participants were adults from the USA who self-identified as African American or black. Recruitment flyers were posted around the San Francisco Bay Area and shared on our team Facebook page, with further participants identified through snowball sampling. RESULTS: We interviewed 26 participants from 10 different states. Overall, 12 were men and 14 were women, with a mean age of 43 years (range 18-85). We identified five key themes regarding melanoma awareness in black people: (1) lack of understanding of term 'melanoma' and features of skin cancer; (2) do not feel at risk of melanoma skin cancer; (3) surprise that melanoma can occur on palms, soles and nails; (4) skin cancer awareness messages do not apply to or include black people; and (5) Importance of relationship with healthcare and habits of utilisation. CONCLUSIONS: Analysis of these in-depth semistructured interviews illuminate the pressing need for health information on melanoma designed specifically for black people. We highlight two key points for focused public health messaging: (1) melanoma skin cancer does occur in black people and (2) high-risk sites for melanoma in black people include the palms, soles and nail beds. Therefore, public health messages for black people and their healthcare providers may involve productively checking these body surface areas.

Skin Tone Analysis for Representation in Educational Materials (STAR-ED) using machine learning.

Images depicting dark skin tones are significantly underrepresented in the educational materials used to teach primary care physicians and dermatologists to recognize skin diseases. This could contribute to disparities in skin disease diagnosis across different racial groups. Previously, domain experts have manually assessed textbooks to estimate the diversity in skin images. Manual assessment does not scale to many educational materials and introduces human errors. To automate this process, we present the Skin Tone Analysis for Representation in EDucational materials (STAR-ED) framework, which assesses skin tone representation in medical education materials using machine learning. Given a document (e.g., a textbook in .pdf), STAR-ED applies content parsing to extract text, images, and table entities in a structured format. Next, it identifies images containing skin, segments the skin-containing portions of those images, and estimates the skin tone using machine learning. STAR-ED was developed using the Fitzpatrick17k dataset. We then externally tested STAR-ED on four commonly used medical textbooks. Results show strong performance in detecting skin images (0.96 ± 0.02 AUROC and 0.90 ± 0.06 F1 score) and classifying skin tones (0.87 ± 0.01 AUROC and 0.91 ± 0.00 F1 score). STAR-ED quantifies the imbalanced representation of skin tones in four medical textbooks: brown and black skin tones (Fitzpatrick V-VI) images constitute only 10.5% of all skin images. We envision this technology as a tool for medical educators, publishers, and practitioners to assess skin tone diversity in their educational materials.

Teletriage for provision of dermatologic care: a pilot program in the Department of Veterans Affairs.

BACKGROUND: The scarcity of dermatologists has prompted the creation of innovative methods for delivering dermatologic care. METHODS: Teletriage is a method used in teledermatology to efficiently assess skin complaints in patients who do not have prompt access to conventional dermatologic care. Their primary care clinicians are provided with management recommendations, reassured that the lesion of concern is benign, or recommended to send their patient for a face-to-face dermatology visit. The Providence VA Medical Center conducted a pilot program testing the utility of teletriage for patients with skin complaints from June 2011 to August 2011. RESULTS: The pilot program revealed that with the teletriage protocol, face-to-face visits were reduced by 38%. This program suggests that teletriage could be a useful tool for providing efficient dermatologic care, and has led to broader implementation. CONCLUSION: Teletriage is a potentially useful approach for efficiently addressing specific dermatologic complaints and improving access to care for those complaints.