Fitzpatrick Skin Type Diversity in Online Patient Education Images
Implicit bias may impact health outcomes for patients of racial and ethnic minorities. While most images in plastic surgery literature reflect white skin types, the content of online resources has not been studied (Massie et al, 2019). Especially in aesthetic surgery, where the internet is a commonly used resource for patient decision making, it is paramount to understand the extent of diversity in images available to patients online (Parmeshwar et al, 2018). The objective of this study was to determine if online before/after images of breast augmentation represent the racial and ethnic patient population in the US.
Methods/Technique
:A Google search for “boob job [city name]” was conducted for 7 US cities. The first 20 websites were analyzed for inclusion of before/after patient photos, and each case was categorized as a Fitzpatrick skin type I-VI by three reviewers. The top 10 pages of “breast augmentation before/after” images on the patient blog Realself were also analyzed to understand what patients themselves are posting. Previous studies have classified skin types I-III as white and IV-VI as non-white, however conflating race with Fitzpatrick types can be problematic and inaccurate (Massie et al, 2019; Ware et al, 2020). We analyzed images by groupings of skin types (I-II, III-IV, and V-VI), as the greatest inter-rater variability existed between the grouped categories. Additionally, since it is difficult for skilled researchers to distinguish types within the groupings while using reference pictures/definitions, we believe these categories are consistent with how patients would view similarities and differences in skin types presented.
Demographic data for each city and nationwide ASPS surgical data were also collected to compare patient representation in online images to the composition of the general population and women undergoing breast augmentation. We used descriptive data to describe percentages of skin type and their distribution. Statistical analysis was conducted using Microsoft Excel and STATA.
Results/Complications
:In total, 109 websites and 3549 images were analyzed. The most common Fitzpatrick skin types were I-II, ranging from 80.41% in Miami and 90.70% in LA. Types III-IV were second most common, ranging from 8.68% in LA to 18.43% in Miami. Atlanta had the most V-VI skin types, at only 1.65%. Of 299 Realself images analyzed, 91.64% were skin types I-II. Table 1 shows complete data.
We do not believe that Fitzpatrick types can be conflated with racial or ethnic categories, and we recognize that vast diversity exists between skin types IV-VI. However, to underscore the sparse representation of higher number skin types, and to best compare our data to existing literature and census data, we have included Table 2, grouping our data against 2019 US census data as “white” (I-III) and (“non-white”) VI-VI. Notably, census data show that most residents of the cities self-designate as a race origin other than white, ranging from 62.90% in Atlanta to 90.20% in Miami.
Table 1: Percentages of Fitzpatrick skin type by city
|
City |
I-II |
III-IV |
V-VI |
|
Baltimore |
85.85 |
13.78 |
0.37 |
|
Dallas |
88.38 |
12.01 |
0.83 |
|
Chicago |
88.38 |
10.52 |
1.10 |
|
Miami |
80.41 |
18.43 |
1.15 |
|
New York |
88.75 |
10.84 |
0.41 |
|
Los Angeles |
90.70 |
8.68 |
0.62 |
|
Atlanta |
83.71 |
14.64 |
1.65 |
|
Realself |
91.64 |
7.69 |
0.67 |
Table 2: Percentages of Fitzpatrick skin type vs. US Census data as “white” or “non-white” by city
|
City |
I-III (“White”) |
IV-VI (“Non-White”) |
Census data: White, not Hispanic or Latino |
Census data: Non-white |
|
Baltimore |
93.67 |
6.33 |
27.50 |
73.90 |
|
Dallas |
96.89 |
3.11 |
29.00 |
72.20 |
|
Chicago |
95.92 |
4.08 |
32.80 |
68.50 |
|
Miami |
93.09 |
6.91 |
10.70 |
90.20 |
|
New York |
95.09 |
4.91 |
32.10 |
71.30 |
|
Los Angeles |
97.52 |
2.48 |
28.50 |
73.60 |
|
Atlanta |
91.96 |
8.04 |
37.60 |
62.90 |
(Note: can select multiple races on census)
Conclusion
:Our study showed that online images that serve as patient education material consistently represented lower number Fitzpatrick skin types across cities in all quadrants of the US. However, census data clearly depict significant diversity in the cities studied. This leads us to believe that online images do not accurately reflect the skin type diversity of our cities nor do they represent our breast augmentation patients. These findings illuminate possible implicit bias in our image selection as healthcare providers. To work towards equitable care, it is imperative that considerable thought and effort be put into designing plastic surgery online resources so that they are inclusive and representative of skin type diversity.
Citations:
Massie JP, Cho DY, Kneib CJ, Burns JR, Crowe CS, Lane M, Shakir A, Sobol DL, Sabin J, Sousa JD, Rodriguez ED, Satterwhite T, Morrison SD. Patient Representation in Medical Literature: Are We Appropriately Depicting Diversity? Plast Reconstr Surg Glob Open. 2019 Dec 26;7(12):e2563. doi: 10.1097/GOX.0000000000002563. PMID: 32042543; PMCID: PMC6964926.
Parmeshwar N, Reid CM, Park AJ, Brandel MG, Dobke MK, Gosman AA. Evaluation of Information Sources in Plastic Surgery Decision-making. Cureus. 2018 Jun 9;10(6):e2773. doi: 10.7759/cureus.2773. PMID: 30109166; PMCID: PMC6084688.
Ware OR, Dawson JE, Shinohara MM, Taylor SC. Racial limitations of fitzpatrick skin type. Cutis. 2020 Feb;105(2):77-80. PMID: 32186531.