Department of Surgery, SUNY Downstate Health Sciences University, College of Medicine, Brooklyn, New York.
Department of Surgery, Maimonides Medical Center, Division of General Surgery, Brooklyn, New York.
J Surg Educ. 2024 Nov;81(11):1772-1777. doi: 10.1016/j.jsurg.2024.07.029. Epub 2024 Sep 20.
Disparity in healthcare is an important and timely topic. One example of such disparity appears to be the disproportionate use of lighter skin tone illustrations in medical and surgical educational material. Visual representation of pathology in instructional material is one key element that informs decision making in surgical disease and could contribute to disparity in outcomes in underrepresented tonal groups. Our hypothesis is that visual examples (illustrations) of clinical pathology in major surgical texts are biased in that they are heavily weighted to Caucasian skin tones and therefore fail to accurately represent the racial distribution of patients in the United States population.
Images from 4 commonly used general surgery textbooks were screened independently by 2 reviewers from Maimonides Medical Center and SUNY (State University of New York) Downstate College of Medicine. Human photographic and cartoon images (where skin tone could be determined), with adequate skin shown, were included. These images were assigned a Fitzpatrick skin photo type (FP) score (1-6). The distribution of images among the 6 FP categories were compared to the expected distribution of images in the United States population, as described from a previous National Health and Nutrition Examination Survey (NHANES). Differences in distribution were compared using a chi-squared test, with p-value <0.05 considered as statistically significant.
There were 556 total images deemed adequate for assessment from the 4 textbooks chosen. 169 from Sabiston, 61 from Mulholland and Greenfield, 177 from Schwartz, and 149 from ACS. About 539 of these images (96.9%) were depictions of patients with light skin tone (FP scores 1-3.) while less than 4% of images were of dark-skinned individuals (FP score between 4 and 6.) An FP score 1 accounted for most images, comprising 477 images (86.1%). There was a 1.8% analytical discrepancy noted between the textbook reviewers. The distribution on the general US population (NHANES) is: FP score 1: 1.6%, FP score 2: 33.1%, FP score 3: 47.8%, FP score 4: 4.9%, FP score 5: 3.6%, FP score 6: 9.0%.
Screening of commonly used general surgery textbooks showed a significant lack of diversity in image-based skin tone representation when compared to the United States population at large. The overwhelming majority of images were of light skin tones. Improving diversity of imagery in educational material, such as basic textbooks, might help reduce observed disparities in outcomes among surgical patients in the future.
医疗保健方面的差异是一个重要且及时的话题。这种差异的一个例子似乎是在医学和外科教育材料中不成比例地使用浅色肤色插图。教学材料中病理学的视觉表现是影响外科疾病决策的一个关键因素，并且可能导致代表性不足的肤色群体的结果存在差异。我们的假设是，主要外科教科书中的临床病理学的视觉示例（插图）存在偏差，因为它们严重偏向于白种人肤色，因此未能准确代表美国人口中患者的种族分布。
我们的研究人员从 Maimonides 医疗中心和纽约州立大学（State University of New York）Downstate 医学院的 2 名审阅者分别对 4 本常用普通外科教科书的图像进行了筛查。纳入了具有足够皮肤显示的人类摄影和卡通图像（可以确定肤色）。这些图像被分配了 Fitzpatrick 皮肤照片类型（FP）评分（1-6）。将图像在 6 个 FP 类别中的分布与先前的国家健康和营养检查调查（NHANES）中描述的美国人口中的图像预期分布进行比较。使用卡方检验比较分布差异，p 值<0.05 被认为具有统计学意义。
从选择的 4 本教科书中评估了 556 张总图像。Sabiston 有 169 张，Mulholland 和 Greenfield 有 61 张，Schwartz 有 177 张，ACS 有 149 张。大约 539 张（96.9%）图像显示的是浅色肤色的患者（FP 评分 1-3），而不到 4%的图像显示的是深色皮肤的个体（FP 评分在 4 到 6 之间）。FP 评分 1 占了大多数图像，共 477 张（86.1%）。教科书审阅者之间存在 1.8%的分析差异。美国普通人群（NHANES）的分布为：FP 评分 1：1.6%，FP 评分 2：33.1%，FP 评分 3：47.8%，FP 评分 4：4.9%，FP 评分 5：3.6%，FP 评分 6：9.0%。
对常用普通外科教科书的筛选显示，与美国普通人群相比，基于图像的肤色代表性严重不足。绝大多数图像都是浅色肤色。在基础教科书等教育材料中提高图像的多样性，可能有助于减少未来外科患者结果中观察到的差异。
