Databases and search terms

The literature review utilised PUBMED and CINAHL for searches between March and April 2023. The following search terms were used to find dermatology-focused peer-reviewed articles published in the last 10 years:

(radiotherapy [Title] OR skin reaction [Title] OR dermatolog* [Title] OR erythema [Title] OR desquamation [Title]) AND (diversity [Title] OR inclusiv*[Title] OR col?r [Title] OR black [Title] OR brown [Title] OR dark* [Title] OR white* [Title])

Inclusion and exclusion criteria

Inclusion criteria (and justification):

• Articles from the past 10 years from peer-reviewed journals (to get a picture of current dermatology issues)

• Articles that cover the key themes of the dissertation (language bias, assessment bias, management/treatment issues (due to delays in diagnosing), educational bias, other biases related due to lack of skin-tone inclusive advice)

• Relevant grey literature (lack of peer-reviewed articles means the review may benefit from the literature in which the primary focus is not peer-reviewed publications)

• The literature covering skin reactions such as erythema or desquamation—common radiotherapy reactions (not enough literature directly related to radiotherapy in this area)

Exclusion criteria (and justification)

• The literature linked to dermatology which cannot be extrapolated to radiotherapy-specific issues (is beyond the scope of this review)

• Non-English (lack of money, time and resources for translation services)

• Articles around the workforce population = beyond scope of the literature review

Data collection

A data collection table was created to record studies, information and identified themes that were explored in the discussion section. The PRISMA tool was used to detail the number of articles included and excluded in the literature review. The SIGN tool was utilised as part of the critical appraisal of the articles included in the study (Figure 1).

Figure 1. Prisma flow diagram detailing search results for the literature review.

Inclusion and exclusion criteria for PILs

Inclusion criteria (and justification)

• Authors departmental patient information leaflets aimed at providing skin care advice (the detailed methodology can be used by radiographers in other departments to analyse theirs)

• English language leaflets

• The most up-to-date version

Exclusion criteria (and justification)

• Leaflets not produced by the author’s department

• Duplicate text

• Working documents that are not yet released (may still be a work in progress and therefore not in a state ready for analysis and dissemination of results)

• Non-English translations of the leaflets

Data collection

The author obtained permission from their NHS Trust and departmental manager to use the leaflets as part of the study. The relevant PILs were found electronically and copied and pasted into a Word document for content analysis. The analysis included highlighting relevant text with green for inclusive and yellow for less inclusive and utilising the comments feature on Microsoft Word for the author’s comments and codes. The content analysis evaluated the sentence structure, language, and order of ideas regarding skin tone using a code book (see Figure 2).

Figure 2. Table showing code book author used for PILs content analysis.

Codes

The author used priori coding, so the content analysis methods produced transparent and reproducible findings. Initial coding was the first step so the authors could familiarise themselves with the patient information leaflets. Any additions, subtractions or changes in codes are made at this stage. Afterwards, line-by-line coding is used to extract all relevant information. Information was condensed for coding, categorised and grouped into themes. Finally, conclusions were drawn based on the outcomes.

Educational bias

Foreman et al. (2022) ^{Reference Foreman, Zappas and Lavell4} focused on nurse practitioner knowledge of dermatology for all pigmentations. The study noted that for those with dark skin tones, erythema was not routinely ‘red’ as it is frequently described. If visual changes were to be seen, it is much more likely that hyperpigmentation would be present. The article states, ‘…recognition [of racial bias] alone cannot resolve biases’.

Hijab et al. (2022) ^{Reference Hijab, Baranowski, MacKelfresh and Krueger5} conducted a thematic content assessment of a one-week dermatology course for medical students at a single institution. The authors aimed for 35% of the educational images were skin of colour (Fitzpatrick skin type FST 4–6) to reflect their local population. Two trained reviewers assessed the images and classified them into non-skin of colour (FST 1–3) or skin of colour (FST 4–6). The study found that 21.2% of images were classified as skin of colour, failing to reach the 35% target. Additionally, an optional pre- and post-lecture (on cultural humility and diseases that disproportionally affect dark skin) survey was handed out to the medical students. 88/107 first-year and 49/107 third-year students completed the pre- and post-surveys. Results were that the course covered diversity, inclusion and cultural competency (p > 0.001), and students who attended were more likely to agree that dermatology needs emphasis on cultural competency (p < 0.001) and to agree on the importance of the role of culture in dermatology (p = 0.0022). The article provided a blueprint or ‘roadmap’ on how to analyse the diversity of curriculums.

Belliscoso et al. (2021) ^{Reference Bellicoso, Quick and Ayoo6} audited their curriculum to assess the diversity of skin tones represented in educational materials. Two authors independently reviewed the images used in the university, classifying FST 1–3 as white skin and FST 4–6 as skin of colour (SoC). The audit identified that only 19 out of 513 images from teaching materials and 12 out of 198 images from online learning materials depicted SoC. In additional to the audit, a survey was conducted among first- and third-year medical students to evaluate their self-rated diagnostic confidence in identifying dermatological conditions on different skin tones. Out of 269 first-year students, 101 responded, and out of 261 third-year students, 76 responded. The survey aimed to determine if there was a difference in diagnostic confidence between identifying conditions on white skin and SoC. The results indicated that first-year students had a mean confidence score of 18.75 for diagnosing white skin and 17/75 for SoC (on a scale of 25). Third-year students had a mean score of 17.75 for white skin and 15.79 for SoC. This significant difference in confidence levels (p = 0.0002) suggests that students felt less confident diagnosing conditions on SoC compared to white skin, highlighting the need for a more diverse representation of skin tones in medical education.

A 2021 study by Gregersen and Elsner ^{Reference Gregersen and Elsner7} carried out a systematic review (of 17 textbooks) to see how diverse German dermatology textbooks were and how important diversity was. They found out of 5,354 images the FST 1/2 = 4,892.11, FST 3 = 344.10, FST 4 = 51.33, FST5 = 23.5 and FST 6 = 1. The authors concluded that there is an educational bias in Germany towards dark tones, which is increasingly becoming problematic with their diversifying population.

Publication bias

A 2020 systematic review by Bray et al. ^{Reference Bray, McMichael, Huang and Feldman8} focused on the diversity in dermatology publications between January 2016 and July 2019. Twenty-five publications focusing on diversity were identified. A single-factor ANOVA and two-group t-tests were used. The mean number of dermatology publications was 6.25, with results dropping as low as 1.75 for general surgery topics (p=0.02 dermatology vs general surgery). The authors concluded that diversity in clinical trials and research articles needed to be improved to adequately represent or inform medical decisions for patients with dark skin tones.

An article by Wilson et al. ^{Reference Wilson, Sun and Ashbaugh9} reviewed 50 dermatology-based journals between January 2018 and October 2020. The authors developed ‘the first-ever prespecified criteria for assessing SoC and diversity in the dermatologic literature’. The authors created multiple criteria, including ‘Title specifically addresses issues regarding diversity and inclusion within dermatology’, to assess if the literature focused on diversity issues. The mean percentage of articles focusing on diversity and SoC was 16.8% (2.04–61.81%). The results for Western journals, such as those in Europe, dropped below 5%. Rates seen in Western publications must accurately reflect the population these journals serve. The Toronto study by Bellicoso ^{Reference Bellicoso, Quick and Ayoo6} 2021 showed their curriculum was not diverse even though Toronto is one of the most multicultural cities in the world, with nearly half of the people identifying as non-white.

Diagnostic bias

Diagnostic bias was seen for dark skin tones in all areas of the aforementioned themes (educational/publication bias). A 2021 UK-based study by Ooi, Lim, Ooi & Bennett ^{Reference Ooi, Lim, Ooi and Bennett10} addresses it directly. The authors utilised a mixed model methodology with a paired t-test, showing skin tone diverse online learning resources improved participants’ confidence with dark skin tone management (1.76 vs the post-course mean of 2.82, p = 0.0001). A feedback form of 50 doctors found that nearly all (42) reported improved knowledge of skin issues for dark skin patients. A focus group of 10 doctors found that 70% expressed a ‘further need for BAME representation in dermatological teaching’. The authors concluded that a lack of exposure to dark-skinned in training results in diagnostic uncertainty, with a lack of appropriate diagnostic tools for dark skin tones leading to poorer outcomes. The authors said that some skin issues are underdiagnosed in dark skin as most diagnostic tools rely on visual aids.

Language bias

Language bias was identified as a theme and is mainly covered by the content analysis and discussion sections.