United States’ clinical practice guidelines (CPGs) are often produced by professional societies and used worldwide in daily medical practice. However, studies in various medical specialties demonstrate underrepresentation of women and racial and ethnic minority groups in CPGs. The representation of authors by gender, race, and ethnicity of US pathology CPGs has not been previously evaluated.
To assess if women and individuals from racial and ethnic minority groups are underrepresented as authors of pathology CPGs.
The gender, race, ethnicity, and terminal degrees of authors of 18 CPGs from the College of American Pathologists were coded by using photographs and other available information online and compared to their representation in academic pathology per Association of American Medical Colleges benchmark data.
Two hundred seventy-five author positions (202 physician author positions) were analyzed. Women overall (119 of 275; 43.3%) and women physicians (65 of 202; 32.2%) held fewer positions than all men and men physicians. Women physicians were significantly underrepresented in physician author positions, while White men physicians were significantly overrepresented in all, first, senior, and corresponding authorship roles when compared to the proportion of women and White men physicians among pathology faculty, respectively. Asian men and women physicians were underrepresented as compared to their representation among pathology faculty.
Men, particularly White men physicians, are overrepresented among pathology CPG author positions, while women physicians and some physicians from racial and ethnic minority groups are underrepresented. Further research is needed to understand the impact of these findings on the careers of underrepresented physicians and the content of guidelines.
Clinical practice guidelines (CPGs) influence the daily practice of medicine. In the United States, guidelines are often produced by professional societies and used around the world. As such, they have a significant and far-reaching impact on the way that care is delivered and resources are distributed. The development of CPGs has changed over time and involves an increasingly formal process that includes experts as authors and contributors.1,2 Studies have demonstrated across many specialties that women experts,3–12 particularly women physicians,2,13,14 are underrepresented as authors of CPGs. Recent studies have also demonstrated that experts from racial and ethnic minority groups are underrepresented as authors.2,15
In a landmark study published in The Lancet in 2018, Merman et al13 studied CPGs from 28 specialties and found that female authors overall had low representation (37.5%). For women physicians, the proportion dropped to 25.0%. The authors also found that there was no improvement in representation during the study period (2012–2017). Unfortunately, this study did not include pathology. Merman and colleagues13 noted that their study should be used to encourage further investigation, and since its publication, many specialties have assessed the gender composition of author panels. For example, there are more than half a dozen reports focused on the representation of women authors for cardiac-related CPGs.9,11–14,16,17 Another study that included all CPGs published in national general medical journals from the United States, United Kingdom, Canada, and Australia between June 2014 and June 2021 assessed authors by using a binary classification (White person versus racialized person).15 For the 237 guidelines included, there were 3696 unique panel members. Of these, 38% were women, 60% were men, and 2% were unclear. Racialized panel members comprised 20%. The most common category of panel member was White men and the least common was racialized women (who were often not included). Notably, to our knowledge the specialty of pathology has not been studied to date.
There are 2 main reasons why the author composition of CPG panels has come under scrutiny. First is that there are many well-documented gaps in access to care as well as treatment itself for patients from historically or socially marginalized groups including, but not limited to, women, people who identify with racial or ethnic minority groups, individuals with disabilities, or people who identify as a sexual orientation or gender identity minority. Although the literature on whether and how the composition of the author panel affects the content of CPGs is in its infancy, this is an area of active study and is of utmost importance.2
The second reason why author composition is a concern is that authorship confers prestige and a host of benefits to invited authors. For example, there is a large body of literature documenting disparities in the promotions process for faculty from underrepresented groups in academic medicine,18 and authorship of a highly cited CPG may not only burnish one’s reputation and support future collaborations with coauthors on the panel (opportunities that may be difficult to measure but are undeniably impactful in one’s career), but also increase bibliometrics such as number of citations and Hirsch index (h-index) (the impact of which is easily measured as it pertains to authorship of a single CPG). Bibliometrics are used to measure how someone’s body of work impacts medicine and science, and thus they are used by many institutions as part of the criteria for rank promotion to assistant, associate, and full professor. For instance, it would not be unusual for a single CPG to contribute 1 point and more than 100 citations to a faculty member’s h-index and total citations. Equitable inclusion of faculty members downstream from the promotions process is an essential part of addressing gaps in promotion for talented and qualified physicians and scientists from diverse backgrounds. Notably, while most CPG authors tend to be currently affiliated with academic medicine or have been in the past, some authors come from other practice settings or may be associated with industry, the community, or another realm. Even for authors not in academic medicine, participation may confer many career or other benefits.
This novel study in the specialty of pathology builds on the previous literature in other specialties that focus on the gender, race, and ethnicity of CPG authors. In this study we assessed the characteristics of gender, race, and ethnicity among authors of pathology CPGs. We hypothesized that women and authors identified with racial or ethnic minority groups would be underrepresented. Notably, we use gender terms (ie, men/women) except when describing the literature if the original source uses sex terms (ie, male/female).
METHODS
The study sample included the complete set of CPGs (n = 18) published between 2012 and 2021 and listed on the College of American Pathologists (CAP) website under the heading “Current CAP Guidelines” (https://www.cap.org/protocols-and-guidelines/current-cap-guidelines). The guidelines were accessed in November and December 2021 and are listed in the Supplemental Table (see supplemental digital content at https://meridian.allenpress.com/aplm in the February 2024 table of contents). CPGs published before 2012 or after 2021 were not included. We also did not include CPGs that were produced by other organizations and listed on the website under the category titled “External Guidelines Endorsed by the CAP.”
The analysis focused on author positions, rather than unique authors, since every position is an opportunity for inclusion and some authors were on more than 1 CPG. All individual author positions in the byline of the CPG were included in the analysis, whereas “contributors” or members of work groups were excluded. Individual author positions were designated as “First,” “Middle,” or “Senior” as based on the order listed in the byline, with the senior author position being the last position. The corresponding author was determined by the listing in the byline as the person listed as either “corresponding author” or the individual to address for “reprints.”
Terminal degrees were obtained from the postnominal initials (eg, MD, DO, PhD, MBA) stated for each author position in the CPG’s byline. If unavailable, degrees were determined from publicly available profiles or other publications.
Gender was determined by gender specifications (eg, man/men, woman/women) or pronouns (ie, he/she/they, him/her/their, his/hers/theirs) stated in public websites or videos. Public information included professional profiles, curriculum vitae, interviews, news, and social media profiles. If the gender or pronouns were unavailable, gender was coded by examining online author photos. If an author’s pronoun or picture was not publicly available, gender was determined by first name, using an online gender application programming interface (API) tool from the gender-API website.19 Because all findings reported gender as binary (ie, he/she) and the coders did not identify any instances of nonbinary descriptions (eg, they), we report gender as binary for the remainder of this study.
Race and ethnicity were designated by authors’ publicly available photographic images, taking into consideration other publicly available data such as birthplace; elementary, undergraduate, and postgraduate education location; languages spoken; organizational affiliation(s); and to a lesser extent first or last names, with the caveat that some authors may be multiracial or have adopted their partner’s surname. In accordance with US Census Bureau guidelines,20 race was designated as “Asian,” “African American/Black,” or “Caucasian/White.” Specifically, individuals with origins from North Africa, the Middle East, or Europe were considered “Caucasian/White,” whereas individuals from the Indian subcontinent, Southeast Asia, and the Far East were considered “Asian.” Ethnicity was designated as “Hispanic/Latino” or “Non-Hispanic/Non-Latino” independent of race. “Hispanic/Latino” individuals were those from Spain, Portugal, and Spanish- and Portuguese-speaking countries in the Americas. To simplify, for the remainder of this report where our analysis is described we will use the terms White, Asian, Black, Non-Hispanic, and Hispanic (capitalizing all terms for consistency and standardization). This method for coding race and ethnicity has been used in previously published studies,2 and important limitations include the inability to identify individuals belonging to other minority groups (or subgroups) and those of mixed race.
Two authors (A.A.M., S.C.W.) independently collected data and coded author gender, race, and ethnicity. Interrater reliability was assessed by using Cohen κ. Coding reconciliation was performed by 2 different authors (J.K.S., G.S.B.). Notably, when researchers code race and ethnicity, they are identifying the perceived race and ethnicity of an individual rather than their true or self-identified race and ethnicity. The specific issue of perception by others of someone’s race and ethnicity as it pertains to research methodology has been discussed in previous literature,2,21 and the perception by others is believed to be an important factor in bias and discrimination. In this study, it is acknowledged that race and ethnicity are based on the coders’ perceptions.
The authors’ gender was compared to the US census’ 1-year estimates for the years 2012, 2014, 2015, and 2021.22–25 The years 2012 and 2021 were used because the CPGs included in this study were published within that timeframe. The years 2014 and 2015 were also used because the faculty benchmarks used for comparison included those years, as well. Races and ethnicities of the authors were compared to the US population per the census data from 2010 and 2020.26–29
For comparison purposes, the gender benchmarks used included Association of American Medical Colleges (AAMC) data from 2021: (1) all faculty (ie, inclusive of PhDs and other nonphysician faculty) in clinical science pathology (55.7% men and 44.3% women) and (2) faculty physicians in clinical science pathology (55.6% men and 44.4% women).30 Basic science pathology is a separate category for AAMC benchmarks, and these were not used. Notably, there are slightly more men in basic science pathology than in clinical science pathology (eg, in 2021 basic pathology had 56.9% men faculty and 56.4% men physician faculty). Race and ethnicity benchmarks came from the 202122 and 201523 AAMC data sets. The 2015 data set contained only “all faculty” so this benchmark was not used for the physician analysis. AAMC benchmark data are not available for each year and have become more comprehensive over time, so for the “all faculty” analysis we used 2015 (61.1% men and 38.9% women)31 and for the physician analysis we used 2014 (63% men and 37% women).32 Since the AAMC does not publish the same reports every year, the benchmarks used in this study were selected carefully after discussions among the authors and communication with AAMC. Refer to the Limitations section for further details. Descriptive statistical analyses were performed. The Fisher exact test was used to compare the differences in the proportions of the categorical data for gender, race, and ethnicity. An α of .05 was used to construct 2-sided CIs for odds ratios (ORs). A P value < .05 was considered significant. The R version 4.0.3 (R statistical software, Vienna, Austria) software was used for the statistical analysis.
To determine whether there were trends in specific areas of pathology, we grouped the guidelines into the topic categories of surgical pathology, breast pathology, hematopathology, gastrointestinal pathology, cytopathology, urogenital pathology, and molecular pathology.
The information used in this analysis is in the public domain and there was no interaction with human subjects; therefore, institutional review board approval was not required.
RESULTS
Among the 18 CPGs analyzed, there were 275 author positions (221 unique authors, 29 of whom held a position on more than 1 CPG). Further information about the authors is described in the Table.
All Authors
Gender, Race, Ethnicity
Of the 275 total author positions, more than half were held by men (156 of 275; 56.7%) versus women (119 of 275; 43.3%). Women were overrepresented in author positions in 2015 (38.9%, 1613 of 4149; OR, 1.20; 95% CI, 0.93–1.55; P = .16) and slightly underrepresented in 2021 (44.3%, 2182 of 4922; OR, 0.96; 95% CI, 0.74–1.23; P = .76) when compared to the respective AAMC benchmarks for all faculty (inclusive of, but not limited to physicians); however, these trends were not significant (Figure 1). Nevertheless, women were significantly underrepresented when compared to the US population in 2012 (159 477 797 of 313 914 040), 2014 (161 966 955 of 318 857 056), 2015 (163 250 987 of 321 418 821) (50.8%; OR, 0.74; 95% CI, 0.58–0.94; P = .01), and in 2021 (50.5%, 167 543 042 of 331 893 745; OR, 0.75; 95% CI, 0.59–0.96; P = .02) (Figure 1). There was 100% interrater reliability for gender.
Composition of all author positions and physician author positions by gender on pathology clinical practice guidelines. Horizontal lines represent the US Census Bureau’s 1-year estimates for all females in the US population (for the years 2012, 2014, 2015, and 2021) and the Association of American Medical Colleges (AAMC) benchmarks for all women (2015 and 2021) and women physicians (2014 and 2021) among pathology (clinical sciences) medical faculty. All women (P = .01) and women physicians (P < .001) were underrepresented from the US population in all years. Men physicians were overrepresented from the AAMC 2021 pathology physician faculty data (P = .001). *“Females” instead of “women” as per the US Census Bureau’s description.
Composition of all author positions and physician author positions by gender on pathology clinical practice guidelines. Horizontal lines represent the US Census Bureau’s 1-year estimates for all females in the US population (for the years 2012, 2014, 2015, and 2021) and the Association of American Medical Colleges (AAMC) benchmarks for all women (2015 and 2021) and women physicians (2014 and 2021) among pathology (clinical sciences) medical faculty. All women (P = .01) and women physicians (P < .001) were underrepresented from the US population in all years. Men physicians were overrepresented from the AAMC 2021 pathology physician faculty data (P = .001). *“Females” instead of “women” as per the US Census Bureau’s description.
Of the 275 author positions, there was 1 position where race and ethnicity was unable to be identified and therefore was excluded. The analysis revealed that of 274 author positions, the race identification was as follows: 232 White (84.7%), 33 Asian (12.0%), and 9 Black (3.3%) (Figure 2; Table). White authors were overrepresented when compared to the US population in 2010 (72.4%, 223 553 265 of 308 745 538; OR, 2.11; 95% CI, 1.51–3.00; P < .001) and in 2020 (61.6%, 2 042 777 273 of 331 449 281; OR, 3.44; 95% CI, 2.47–4.90; P < .001). Similarly, Asian authors were overrepresented when compared to the US population in 2010 (4.8%, 14 674 252 of 308 745 538; OR, 2.74; 95% CI, 1.85–3.96; P < .001) and in 2020 (6.0%, 19 886 049 of 331 449 281; OR, 2.16; 95% CI, 1.44–3.10; P < .001). However, Black authors were underrepresented when compared to the US population in 2010 (12.6%, 38 929 319 of 308 745 538; OR, 0.24; 95% CI, 0.11–0.45; P < .001) and in 2020 (12.4%, 41 104 200 of 331 449 281; OR, 0.24; 95% CI, 0.11–0.46; P < .001).
Composition of pathology clinical practice guidelines’ author positions by author type, race, and ethnicity. Horizontal lines represent the proportion of people by race and ethnicity in the United States, based on the 2010 and 2020 US Census Bureau data. *P < .01, **P < .001, §P < .05.
Composition of pathology clinical practice guidelines’ author positions by author type, race, and ethnicity. Horizontal lines represent the proportion of people by race and ethnicity in the United States, based on the 2010 and 2020 US Census Bureau data. *P < .01, **P < .001, §P < .05.
Most positions were held by White men (50.4%, 138 of 274) and White women (34.3%, 94 of 274), while Asian men, Asian women, Black men, and Black women combined accounted for 15.3% of all author positions (Figure 3; Table).
Composition of all author positions (n = 274) for which race and ethnicity could be determined on pathology clinical practice guidelines in comparison to 2015 and 2021 Association of American Medical Colleges (AAMC) benchmarks for all pathology (clinical sciences) medical school faculty. *P < .05; **P < .001.
Composition of all author positions (n = 274) for which race and ethnicity could be determined on pathology clinical practice guidelines in comparison to 2015 and 2021 Association of American Medical Colleges (AAMC) benchmarks for all pathology (clinical sciences) medical school faculty. *P < .05; **P < .001.
White men (50.4%, 138 of 274) were overrepresented when compared to 2015 (39.9%, 1654 of 4149; OR, 1.53; 95% CI, 1.19–1.97; P < .001) and 2021 (34.7%, 1710 of 4922; OR, 1.91; 95% CI, 1.48–2.45; P < .001) AAMC benchmark data. Similarly, White women (34.3%, 94 of 274) were overrepresented when compared to 2015 (23.3%, 966 of 4149; OR, 1.72; 95% CI, 1.31–2.24; P < .001) and 2021 (25.4%, 1252 of 4922; OR, 1.53; 95% CI, 1.17–1.99; P = .001) AAMC benchmarks. Asian men (6.2%, 17 of 274) were underrepresented as compared to AAMC 2015 (11.5%, 479 of 4149; OR, 0.51; 95% CI, 0.29–0.84; P = .005) and 2021 (13.6%, 669 of 4922; OR, 0.42; 95% CI, 0.23–0.69; P < .001) faculty benchmarks. Asian women (5.8%, 16 of 274) were not significantly underrepresented when compared to AAMC benchmarks in 2015 (7.7%, 318 of 4149; OR, 0.75; 95% CI, 0.42–1.26; P = .34), but were significantly underrepresented as compared to AAMC benchmarks in 2021 (12.4%, 610 of 4922; OR, 0.44; 95% CI, 0.25–0.73; P = .001). Black women (2.9%, 8 of 274) were overrepresented when compared to 2015 (0.8%, 33 of 4149; OR, 3.75; 95% CI, 1.48–8.38; P = .003) and 2021 (1.1%, 54 of 4922; OR, 2.71; 95% CI, 1.10–5.81; P = .02) AAMC faculty data. Black men (0.4%, 1 of 274) were underrepresented, though not significantly, as compared to 2015 (0.9%, 37 of 4149; OR, 2.46; 95% CI, 0.41–99.95; P = .73) and 2021 (1.1%, 52 of 4922; OR, 2.91; 95% CI, 0.50–117.70; P = .53) AAMC benchmarks. Interrater reliability was 99% for race identification.
Ethnicity for 274 total author positions demonstrated 262 Non-Hispanic (95.6%) and 12 Hispanic (4.4%) author positions. Hispanics were underrepresented when compared to the US population in 2010 (16.3%, 50 477 594 of 308 745 538; OR, 0.23; 95% CI, 0.12–0.42; P < .001) and in 2020 (18.7%, 62 080 044 of 331 449 281; OR, 0.20; 95% CI, 0.10–0.35; P < .001) (Figure 2). The gender and ethnicity analysis revealed that 9 of the positions (3.3%) were held by a Hispanic woman and 3 (1.1%) were held by a Hispanic man (Figure 3). Hispanic women were significantly overrepresented as compared to the 2015 (0.9%, 39 of 4149; OR, 3.58; 95% CI, 1.51–7.61; P = .002), but not the 2021 (2.5%, 122 of 4922; OR, 1.34; 95% CI, 0.59–2.66; P = .42), AAMC benchmarks. Hispanic men were underrepresented, but not significantly, when compared to the 2015 (2.6%, 108 of 4149; OR, 0.41; 95% CI, 0.08–1.26; P = .16) and 2021 (2.8%, 139 of 4922; OR, 0.38; 95% CI, 0.08–1.15; P = .12) AAMC benchmarks. There was 100% interrater reliability for ethnicity identification.
Physician Authors
Of the 275 author positions, 202 (73.5%) were held by a physician.
Gender, Race, Ethnicity
Of the 202 total physician author positions, most were held by men (67.3%, 136 of 202) compared to women (32.7%, 66 of 202) (Figure 1). Women physicians were underrepresented when compared to the US population in 2012 (159 477 797 of 313 914 040), 2014 (161 966 955 of 318 857 056), 2015 (163 250 987 of 321 418 821) (50.8%; OR, 0.47; 95% CI, 0.34–0.64; P < .001), and 2021 (50.5%, 167 543 042 of 331 893 745; OR, 0.48; 95% CI, 0.35–0.64; P < .001). Men physicians were overrepresented, though not significantly, when compared to 2014 AAMC physician faculty data (63%, 1728 of 2743; OR, 1.21; 95% CI, 0.86–1.67; P = .23); however, they were significantly overrepresented when compared to 2021 (55.6%, 1996 of 3588; OR, 1.64; 95% CI, 1.21–2.26; P = .001) AAMC physician faculty data (Figure 1).
White physicians accounted for most physician author positions (87.1%, 176 of 202), followed by Asian physicians (11.9%, 24 of 202), and Black physicians (1.0%, 2 of 202) (Figure 2). When compared to the US population, White physicians were overrepresented in 2010 (72.4%, 223 553 265 of 308 745 538; OR, 2.58; 95% CI, 1.70–4.06; P < .001) and in 2020 (61.6%, 204 277 273 of 331 449 281; OR, 4.21; 95% CI, 2.78–6.63; P < .001), as were Asian physicians in 2010 (4.8%, 14 674 252 of 308 745 538; OR, 2.70; 95% CI, 1.69–4.15; P < .001) and in 2020 (6.0%, 19 886 049 of 331 449 281; OR, 2.11; 95% CI, 1.32–3.25; P = .002). However, Black physicians were underrepresented when compared to the US population in 2010 (12.6%, 38 929 319 of 308 745 538; OR, 0.07; 95% CI, 0.01–0.25; P < .001) and in 2020 (12.4%, 41 104 200 of 331 449 281; OR, 0.07; 95% CI, 0.01–0.26; P < .001). Race and ethnicity benchmark data for physicians from the AAMC were not available for 2014 or 2015, so the analysis focused only on 2021 and this is a limitation of the study. White men physicians held significantly more physician author positions (59.9%, 121 of 202) than the overall proportion of White men physicians among pathology faculty (35.4%, 1271 of 3588; OR, 2.72; 95% CI, 2.02–3.69; P < .001) in 2021 (Figure 4). There were also more White women physician author positions (27.2%, 55 of 202) as compared to the proportion of White women physician pathology faculty (24.6%, 883 of 3588; OR, 1.15; 95% CI, 0.82–1.59; P = .40), but this was not significant. Asian men were significantly underrepresented in author positions (6.9%, 14 of 202) as compared to the proportion of Asian men physician faculty in 2021 (12.9%, 462 of 3588; OR, 0.50; 95% CI, 0.27–0.88; P = .01). Similarly, Asian women physicians held significantly fewer author positions (5.0%, 10 of 202) than the proportion of Asian women pathology physician faculty in 2021 (13.5%, 484 of 3588; OR, 0.33; 95% CI, 0.16–0.63; P < .001) (Figure 4). Fewer author positions were held by Black men physicians (0.5%, 1 of 202) and Black women physicians (0.5%, 1 of 202) than the proportion of Black men physicians (1.2%, 42 of 3588; OR, 0.42; 95% CI, 0.01–2.50; P = .73) and Black women physicians (1.1%, 40 of 3588; OR, 0.44; 95% CI, 0.01–2.63; P = .72) among pathology faculty in 2021, though these differences were not significant.
Composition of physician author positions on pathology clinical practice guidelines in comparison to 2021 Association of American Medical Colleges (AAMC) benchmarks for pathology (clinical sciences) physician medical school faculty. *P < .05; **P < .001.
Composition of physician author positions on pathology clinical practice guidelines in comparison to 2021 Association of American Medical Colleges (AAMC) benchmarks for pathology (clinical sciences) physician medical school faculty. *P < .05; **P < .001.
Non-Hispanic physicians comprised 95.5% (193 of 202) of physician author positions, while Hispanic physicians accounted for 4.5% (9 of 202). Hispanic physicians were underrepresented from the US Hispanic population in both 2010 (16.3%, 50 477 594 of 308 745 538; OR, 0.24; 95% CI, 0.11–0.46; P < .001) and 2020 (18.7%, 62 080 044 of 269 369 237; OR, 0.20; 95% CI, 0.09–0.39; P < .001) (Figure 2). Hispanic women physicians (3.5%, 7 of 202) were overrepresented (2.7%, 98 of 3588; OR, 1.28; 95% CI, 0.49–2.79; P = .51) and Hispanic men physicians (1.0%, 2 of 202) were underrepresented (3.3%, 117 of 3588; OR, 0.30; 95% CI, 0.04–1.11; P = .09) as compared to the 2021 AAMC benchmark, though these differences were not significant (Figure 4).
Authorship Roles
First Author Positions
There were 18 first author positions held by 17 individuals, all of whom were physicians (1 White, Non-Hispanic man held a first author position on 2 CPGs). More first author positions were held by men (66.7%, 12 of 18) than women (33.3%, 6 of 18). When compared to the US female population in 2012 (159 477 797 of 313 914 040), 2014 (161 966 955 of 318 857 056), 2015 (163 250 987 of 321 418 821) (50.8%; OR, 0.48; 95% CI, 0.15–1.39; P = .16), and 2021 (50.5%, 167 543 042 of 331 893 745; OR, 0.49; 95% CI, 0.15–1.41; P = .16), women were underrepresented in first author positions, although the differences were not significant.
White physicians (88.9%, 16 of 18) were significantly overrepresented when compared to the 2020 US population (61.6%, 204 277 273 of 331 449 281; OR, 4.98; 95% CI, 1.17–44.65; P = .02) (Figure 2). Although not significantly, White physicians were also overrepresented from the 2010 US population (72.4%, 223 553 265 of 308 745 538; OR, 3.05; 95% CI, 0.72–27.33; P = .18). Similarly, Asian physicians (11.1%, 2 of 18) were overrepresented when compared to the 2010 (4.8%, 14 674 252 of 308 745 538) and 2020 (6.0%, 19 886 049 of 331 449 281) US population, but the differences were not significant. In contrast, Black physicians (0%, 0 of 18) were underrepresented from the 2010 (12.6%, 38 929 319 of 308 745 538) and 2020 (12.4%, 41 104 200 of 331 449 281) US population but not significantly.
Most first author positions were held by White men (66.7%, 12 of 18) (Figure 5). Significantly more first author positions were held by White men physicians than the overall proportion of White men pathology physician faculty in 2021 (66.7% [12 of 18] versus 35.4% [1271 of 3588]; OR, 3.64; 95% CI, 1.26–11.87; P = .01). Fewer first author positions were held by Asian men physicians (0.0%, 0 of 18) than the proportion of Asian men physicians among pathology faculty (12.9%, 462 of 3588; OR, 0.00; 95% CI, 0.00–1.55; P = .15) (Figure 5) in 2021, though this was not significant. Asian women physicians among first author positions (11.1%, 2 of 18) were underrepresented, though not significantly, compared to the proportion of Asian women physician pathology faculty in 2021 (13.5%, 484 of 3588; OR, 1.25; 95% CI, 0.29–11.22; P > .99).
Physician author positions on pathology clinical practice guidelines by authorship, gender, race, and ethnicity. White men were overrepresented in total authorship roles (P < .001), first authorship roles (P = .01), and senior authorship roles (P = .02) when compared to 2021 Association of American Medical Colleges (AAMC) benchmarks for pathology (clinical sciences) physician medical school faculty.
Physician author positions on pathology clinical practice guidelines by authorship, gender, race, and ethnicity. White men were overrepresented in total authorship roles (P < .001), first authorship roles (P = .01), and senior authorship roles (P = .02) when compared to 2021 Association of American Medical Colleges (AAMC) benchmarks for pathology (clinical sciences) physician medical school faculty.
Black men physicians (0.0%, 0 of 18) and Black women physicians (0.0%, 0 of 18) held fewer first author positions than Black men faculty physicians in pathology (1.2%, 42 of 3588; OR, 0.00; 95% CI, 0.00–19.96; P > .99) and Black women faculty physicians in pathology (1.1%, 40 of 3588; OR, 0.00; 95% CI, 0.00–21.00; P > .99).
The representation of Hispanic physicians (11.1%, 2 of 18) in first authorship roles was similar to that of the US Hispanic population in 2010 (16.3%, 50 477 594 of 308 745 538) and 2020 (18.7%, 62 080 044 of 269 369 237) (Figure 2). However, Hispanic men physicians (0.0%, 0 of 18) were underrepresented in first author positions when compared to Hispanic men faculty physicians in pathology (3.3%, 117 of 3588; OR, 0.00; 95% CI, 0.00–6.84; P > .99), while Hispanic women physicians (11.1%, 2 of 18) were overrepresented in first author positions when compared to Hispanic women faculty physicians in pathology (2.7%, 98 of 3588; OR, 4.45; 95% CI, 0.49–19.33; P = .09). Neither of these trends were statistically significant.
Senior Author Positions
Eighteen senior author positions were held by 17 individuals, and 83% (15 of 18) of positions were held by physicians. One Black, Non-Hispanic, nonphysician woman was a senior author on 2 CPGs. Women (13.3%, 2 of 15) were significantly underrepresented in senior author positions when compared to the 2012 (159 477 797 of 313 914 040), 2014 (161 966 955 of 318 857 056), 2015 (163 250 987 of 321 418 821) (50.8%; OR, 0.15; 95% CI, 0.15–1.39; P = .004), and 2021 (50.5%, 167 543 042 of 331 893 745; OR, 0.15; 95% CI, 0.02–0.67; P = .004) US female population. Women physicians held significantly fewer senior author positions than the proportion of women physicians among pathology faculty in 2021 (13.3%, 2 of 15 versus 44.4%, 1592 of 3588; OR, 0.19; 95% CI, 0.02–0.85; P = .02).
Asian physicians (20%, 3 of 15) were significantly overrepresented in senior author positions as compared to the 2010 US population (4.8%, 14 674 252 of 308 745 538; OR, 5.01; 95% CI, 0.91–18.56; P = .03) (Figure 2). Additionally, Asian physicians were overrepresented from the 2020 Asian US population (6.0%, 19 886 049 of 331 449 281; OR, 3.92; 95% CI, 0.71–14.52; P = .06). White physicians (80.0%, 12 of 15) were overrepresented from the White US population in 2010 (72.4%, 223 553 265 of 308 745 538; OR, 1.52; 95% CI, 0.41–8.42; P = .77) and in 2020 (61.6%, 204 277 273 of 331 449 281; OR, 2.49; 95% CI, 0.67–13.75; P = .19), and Black physicians (0.0%, 0 of 15) were underrepresented from their respective proportions in the US population in both years; however, these differences were not statistically significant.
White women physicians (13.3%, 2 of 15) and Asian women physicians (0.0%, 0 of 15) were underrepresented among senior author positions as compared to the proportion of White (24.6%, 883 of 3588; OR, 0.47; 95% CI, 0.05–2.09; P = .55) and Asian (13.5%, 484 of 3588; OR, 0.00; 95% CI, 0.00–1.80; P = .25) women physician pathology faculty, though these differences were not significant. Most senior author physician positions were held by a White individual, and significantly more were held by White men physicians as compared to the proportion of White men pathology physician faculty in 2021 (66.7%, 10 of 15 versus 35.4%, 1271 of 3588; OR, 3.64; 95% CI, 1.13–3.62; P = .02). No Hispanic physician held a senior author position (Figure 5). Even though Hispanic physicians were underrepresented (0%, 0 of 15) from the 2010 (16.3%, 50 477 594 of 308 745 538) and 2020 (18.7%, 62 080 044 of 331 449 281) US population, this was not significant (Figure 2). Of the 18 CPGs, only 1 CPG had women in both the first and senior author positions.
Corresponding Author Positions
Among the 18 CPGs, 16 had an individual listed as a corresponding author. All were physicians, but 4 had an additional PhD degree. Three of the 4 corresponding authors with an added PhD degree were women. Of all corresponding author positions, 62.5% (10 of 16) were held by men and 37.5% (6 of 16) were held by women. Although not significantly, men physicians were overrepresented when compared to the proportion of males in the US population in 2012 (154 436 243 of 313 914 040), 2014 (156 890 101 of 318 857 056), 2015 (158 167 834 of 321 418 821) (49.2%; OR, 1.72; 95% CI, 0.57–5.76; P = .33), and 2021 (49.5%, 164 350 703 of 331 893 745; OR, 1.70; 95% CI, 0.56–5.69; P = .33). Women representation as corresponding authors was not significantly different when compared to the proportion of women clinical pathology physician faculty in 2014 (37.0%, 1015 of 2743) and 2021 (44.4%, 1592 of 3588) per the AAMC report.
Of all corresponding author positions, 88% (14 of 16) were held by White physicians, whereas 13% (2 of 16) were held by Asian physicians and 0% (0 of 16) by Black physicians. White physicians were significantly overrepresented when compared to the 2020 US population (61.6%, 204 277 273 of 331 449 281; OR, 4.36; 95% CI, 1.00–39.51; P = .04). White physicians were also overrepresented when compared to the 2010 US population, although not significantly (72.4%, 223 553 265 of 308 745 538; OR, 2.67; 95% CI, 0.61–24.18; P = .26). Similarly, Asian physicians were overrepresented, whereas Black physicians were underrepresented, compared to the 2010 and 2020 US population benchmarks, but these differences were not significant (Figure 6). With regard to ethnicity, 12.5% (2 of 16) of corresponding authors were Hispanic, and although Hispanic physicians were underrepresented from the 2010 and 2020 US population, the differences were not significant (Figure 2).
Women physician author positions on pathology clinical practice guidelines by subspecialty as compared to the US Census Bureau’s 1-year estimates for all females in the US population (for the years 2012, 2014, 2015, and 2021) and Association of American Medical Colleges (AAMC) benchmarks (2014 and 2021) for women pathology (clinical sciences) physician medical school faculty.
Women physician author positions on pathology clinical practice guidelines by subspecialty as compared to the US Census Bureau’s 1-year estimates for all females in the US population (for the years 2012, 2014, 2015, and 2021) and Association of American Medical Colleges (AAMC) benchmarks (2014 and 2021) for women pathology (clinical sciences) physician medical school faculty.
When broken down by gender, all men corresponding authors (100%, 10 of 10) were White, whereas 66.7% (4 of 6) of the women were White and 33.3% (2 of 6) were Asian. White men were overrepresented when compared to the AAMC 2021 faculty physicians (62.5%, 10 of 16 versus 35.4%, 1271 of 3588; OR, 3.04; 95% CI, 1.00–10.19; P = .03). However, Asian men (0%, 0 of 16 versus 12.9%, 462 of 3588; OR, 0.00; 95% CI, 0.00–1.76; P = .25), Black men (0%, 0 of 16 versus 1.2%, 42 of 3588; OR, 0.00; 95% CI, 0.00–22.74; P > .99), and Black women (0%, 0 of 16 versus 1.1%, 40 of 3588; OR, 0.00; 95% CI, 0.00–23.93; P > .99) were underrepresented from the AAMC 2021 benchmarks, although not significantly. The representation of White (25%, 4 of 16) and Asian (12.5%, 2 of 16) women physicians in corresponding author positions was similar to the White (24.6%, 883 of 3588) and Asian (13.5%, 484 of 3588) women faculty physicians per the 2021 AAMC report. These findings can be seen in Figure 5.
With regard to ethnicity, 12.5% (2 of 16) of corresponding authors were Hispanic, and although Hispanic physicians were underrepresented from the 2010 and 2020 US population, the differences were not significant (Figure 2). Hispanic men were underrepresented (0%, 0 of 16 versus 3.3%, 117 of 3588; OR, 0.00; 95% CI, 0.00–7.80; P > .99), whereas Hispanic women were overrepresented (12.5%, 2 of 16 versus 2.7%, 98 of 3588; OR, 5.08; 95% CI, 0.55–22.60; P = .07), from their AAMC 2021 benchmarks; however, these findings were not statistically significant (Figure 5).
Clinical Practice Guidelines by Topic Area
The analysis of CPGs by topic area revealed women physician author positions were most represented in the areas of breast and gastrointestinal pathology and least represented in surgical and urological pathology (Figure 6). Figures 7 and 8 illustrate the racial and ethnic composition of all author positions and physician author positions by topic area, respectively.
Race and ethnicity of all author positions on pathology clinical practice guidelines by topic.
Race and ethnicity of all author positions on pathology clinical practice guidelines by topic.
Race and ethnicity of physician author positions on pathology clinical practice guidelines by topic.
Race and ethnicity of physician author positions on pathology clinical practice guidelines by topic.
DISCUSSION
The key findings in this study demonstrate that women, inclusive of physicians and women with other degrees, were overrepresented, though not significantly, or nearly equitably represented in author positions on pathology CPGs as compared to AAMC benchmarks for women in academic pathology (Figure 1). In contrast, women physicians were significantly underrepresented as compared to AAMC benchmarks for women physicians in academic pathology (Figure 1). Furthermore, women were underrepresented in all author positions, physician author positions, and first author positions when compared to the US population (Figure 1).
The race analysis revealed that among all author positions and physician author positions, White and Asian people were overrepresented, whereas Black individuals were underrepresented, as compared to the US population (Figure 2). White individuals as first and corresponding authors were also overrepresented from the 2020 US population, and Asian individuals were overrepresented from the 2010 US population as senior authors. When broken down by gender, among all author positions, White men, White women, Black women, and Hispanic women were significantly overrepresented, while Asian men and Asian women were significantly underrepresented when compared to AAMC benchmarks (Figure 3). Among physician author positions, White men were significantly overrepresented, whereas Asian men and Asian women were significantly underrepresented when compared to 2021 AAMC benchmark data (Figure 4). White men physicians were also overrepresented among first, senior, and corresponding author positions when compared to 2021 AAMC benchmark data (Figure 5).
Our analysis of pathology CPG authors demonstrated relatively equitable proportions of women authors overall, yet women physicians were significantly underrepresented. This latter finding is consistent with some other studies that have found women physicians are more likely to be underrepresented.2,3,13 Additionally, our findings parallel those reported by Verduzco-Gutierrez et al2 with regard to the same woman being included as an author on more than 1 CPG (the Table demonstrates that 16 women and 13 men were included as authors on more than 1 CPG), and while this may improve the appearance of gender equity, it can also exacerbate disparities. For example, of the 16 women included more than once, 10 were not physicians. This likely contributed to the underrepresentation of women physicians in author slots—there was a missed opportunity for multiple qualified women pathology physicians to be on a CPG committee in their field of practice.
There are few prior studies that have examined race and ethnicity of CPG authors.2,15 Similar to our results, they demonstrated disparities for people from racial and ethnic minority groups. Women who are also racial or ethnic minorities may face increased barriers to inclusion.2 Persaud et al15 demonstrated the disproportionately low representation of “racialized” women in multiple different countries. While our work overall agrees that White individuals, especially White men, are overrepresented, the inclusion of various racial and ethnic groups among CPG authors deserves further scientific inquiry.
Whether the diversity of authors affects the content of CPGs has not been well studied. However, it is reasonable to consider that increasing diversity among CPG committee members may lead to improved content and recommendations. For example, in a study analyzing the gender associated with CPGs produced by the World Health Organization (WHO), Bohren and colleagues4 wrote, “When people from diverse backgrounds—including different genders, cultures, ethnicities, and religions—join forces, they bring with them their own experiential knowledge that enriches discussions and promotes equality.” By including authors with varied backgrounds, an assortment of perspectives and ideas can be more thoroughly evaluated by the committee, with the aim to avoid bias that might otherwise be introduced into the recommendations. Additionally, by including authors of varied genders, race, and ethnic backgrounds, clinicians and patients who identify with these groups may feel more confident in accepting the recommendations made by the CPG committees.33
Diversity among authors of CPGs is important, as participation in this type of scholarly work is part of the academic rank promotions process. According to the AAMC, in 2018 women represented 51% of anatomic and clinical pathology residents; however, they only represented 21.7% of clinical department chairs.34,35 Additional AAMC data from 2020 demonstrate that the drop in percentage of women can be seen throughout academic ranking, with women encompassing only 48% and 51% of assistant professor positions, but 30% and 33% of full professor positions in basic science and clinical science pathology, respectively.35 In addition to this numeric data, Lipscomb et al36 demonstrated that the Association of Pathology Chairs experienced a number of challenges before promotion related to lack of time or support, as well as experiences of gender bias (89%), racial/ethnic bias (19%), and sexual harassment (58%). The promotional challenges that women face are likely multifactorial; however, publications and bibliometrics related to them (eg, number of citations, h-index) are known to be contributory factors. CPGs represent an opportunity to reflect the field and provide authorship opportunities for a variety of experts.
Although the representation of women experts on pathology-related CPGs had not previously been reported, there is evolving evidence of gender disparities in the specialty, as well as concern that the content of a CPG may be affected by the composition of the writing team. For example, Khoshpouri and colleagues37 demonstrated a statistically significant disparity between men and women in leadership positions, as well as significant differences in academic productivity, within pathology programs in North America. The discrepancy between male and female authorship has worsened in the setting of the SARS-CoV-2 pandemic, with an increasing gender gap among medRxiv publications.38 Notably, author position and sequence may influence academic hiring, promotion, and funding, illustrating the importance of both first and senior (last) author positions.39 Additional studies that have focused on transfusion medicine journal authorship demonstrated a decrease in female first authors, but not senior authors, during the pandemic, potentially indicating a decrease in academic activity among younger researchers.40 Moreover, the award conferral rate within transfusion medicine demonstrates significant gender inequities,41 and there are also documented disparities among the field’s journal editors. Furthermore, in considering all medical specialties, most women in higher academic positions are White, demonstrating an even more significant decrease in involvement of racial and ethnic minority groups at higher ranking positions42 ; this holds true for pathology departments, with more than half of faculty identifying as White.43
Although we did not study causality, workforce disparities are typically multifactorial. As discussed by Spector and Overholser,44 gender disparities in publications, speakers at national conferences, and salary have all been previously demonstrated. They additionally suggest that narrative feedbacks often used for promotions have different language to describe women and men physicians that may be detrimental to women in promotional tracks. Conscious (explicit) and unconscious (implicit) bias may be implicated in limiting the advancement of women physicians, and identifying and reframing the requirements for promotion and advancement will be required. CPGs have a particular space in promotion considerations given the national and international attention they receive, as well as their potential influence on an individual’s h-index. Physicians who were identified as visible minorities likely face similar struggles and would benefit from revised advancement criteria by, for example, factoring in clinical and educational achievements.
Initial ideas to improve the disparities among authors of CPGs include acknowledgment that women and racial and ethnic minorities may be underrepresented as compared to those in practice.2 This important step could be accomplished by a statement, followed by updating the reports and recommended guidelines that CPG committees rely on for direction to specifically address the need for diverse author teams to support the visibility and participation of underrepresented pathologists. However, documentation and awareness of an issue, while necessary, may not be sufficient to drive change. For example, a re-analysis of gender disparities among transfusion medicine journal editors found that despite this being a relatively easy problem to fix, some journals did not address the problem.45 Many CPGs are convened by medical societies, and those produced in the United States are often recognized for their authoritative guidance, and thus distributed worldwide. As such, both medical societies and their affiliated journals have a responsibility to ensure that author panels are appropriately diverse. The term interorganizational structural discrimination describes a problem that occurs when 2 organizations (in this case a medical society and a journal) collaborate despite one of the organizations having a remediable issue related to structural discrimination (ie, policies or practices that create inequities).46 Avoiding interorganizational structural discrimination is an important step forward in addressing many inequities, including but not limited to the development of CPGs, since organizations must hold each other accountable. This is similar to the social movement referred to as “Fair Trade” that encourages buyers to not support producers that have unethical practices such as child labor or unfair wages for workers.47 Therefore, the initiatives to work toward for inclusion of women as well as racial and ethnic minorities need to be a targeted and conscious effort for both CAP and the Archives of Pathology & Laboratory Medicine; for instance, collaboration with the CAP Council on Membership and Professional Development to review, monitor, and improve upon diversity, equity, and inclusion metrics related to CPG authorship, and the journal can require a review of authorship diversity before publication.
One such metric could be the development of an authorship diversity score similar to the h-index, wherein authors that select to identify demographic information could be coded and an output variable designed to see how diverse the author line is. This theoretical model would help the academic community move away from “mappers” and credit those studies that not only contain interesting material but also have purposely advanced the careers of all physicians and scientists. Furthermore, if this were publicly available, it would help would-be members to decide if the organization itself is reflective of the field and represents the composition they seek to engage with. This is important because CAP and the 2 other largest pathology professional societies have been shown to have large and troubling award conferral rate discrepancies by gender.48 Additionally, those who are invited to serve on these committees should raise concerns if the committee and author list does not reflect the true composition of gender and minorities within their field of practice. Encouraging members of committees to bring forth their concerns in a positive environment is also a critical step in committing to bring forward authors from diverse backgrounds.
Even though it is imperative that journals actively measure authorship diversity pre publication, a greater surveillance of publications ex post facto would further improve author diversity. Specific measures for this include monitoring the degree of authors’ self-citation, as men have been found to have higher self-citations than women.49–51 Given the higher number of views, downloads, and citations within CPGs, there is a compounding effect in h-indices, academic promotion, and career advancement if the authors of said CPGs engage in higher self-citation rates. A greater emphasis should also be placed on alternative metric scoring for academic advancement. Authors using nontraditional means of article broadcasting (ie, social media) should be recognized and rewarded. Regardless of what individuals in academia view social media, it is here to stay. Social media in academics takes time but is purposeful and engaging for physicians and the lay public; the same cannot be said of traditional biomedical publishing, which suffers from paywalls and exorbitant subscription fees. Therefore, incorporating social media and other alternative metric scoring as part of the decision for promotion would ensure that generations that are digitally connected are rewarded for their time and energies.
Interestingly, the Archives of Pathology & Laboratory Medicine has a very large journal masthead,52 but it lacks a specified assistant or associate journal editor who is dedicated to diversity, equity, and inclusion. Such a dedicated journal editor is becoming more common within medical journals, including high-impact medical journals like the Journal of the American Medical Association.53 Having a dedicated editor of diversity advocacy within pathology journals who monitors prepublication and postpublication surveillance would further facilitate diversity within pathology journals.
Beyond diverse author committees, diversity metrics, and dedicated journal editors for equity, diversity, and inclusion, changes aimed at addressing gaps in access to care as well as social determinants of health (in the content itself) are needed. This is illustrated by the known gap in the evidence as it applies to cardiac disease in women (versus men), and Sardar et al14 noted that adding sex-specific recommendations to CPGs would highlight where data are lacking as well as provide clinicians with information about the available evidence that may be applied to patient care rather than reliance on anecdotal experience. Regarding the latter, Norris et al54 found that it was “highly feasible to put in place a structure process among guideline developers to better inform sex-specific assessments of the quality of evidence and strength of recommendations for cardiovascular CPGs.” They recommended appointing a champion or expert who was knowledgeable about sex differences to facilitate oversight and decision-making. Recent guidance statements from the American Academy of Physical Medicine and Rehabilitation have used a model that incorporates the inclusion of content related to social determinants of health, with narrative examples in the body of the report as well as a health equity table—ensuring that readers understand the importance of equitable access to care and interventions.55 In the aforementioned WHO study, Bohren et al4 noted that gender parity (ie, 50% women) is one strategy used to encourage gender equity. They also recommended transparency in the selection process.
LIMITATIONS
Limitations exist for our study, including the designations of gender, race, and ethnicity. Similar to this study, in a previous report, Bertolero et al21 wrote they were not seeking or claiming to determine the “true” race of any given authors, but rather they were using “a flawed approach for assessing something as personal, complex, and societally defined as race.” They went on to explain, “To the extent that these models can accurately measure one or both of these characteristics, they can nontrivially capture the effects of such biases on authors of color.”
Although it seems like a simple solution to survey the CPG authors and ask them to self-identify their gender, race, and ethnicity, in reality this would lead to a smaller sample size (and therefore be subject to statistical errors), since it is unlikely that all authors would respond to a survey. Current established methodologies for identifying identity characteristics in a given sample have limitations, yet there is an urgent need to better identify workforce disparities and their impact on patient care. Thus, while the methodology in this study has limitations, it nevertheless accounts for the perception others have of potential authors’ gender, race, and ethnic identities, which in turn may influence whether they are included in sentinel papers, such as the CPGs analyzed. Even though we analyzed the perceived gender of CPG authors, some of our benchmark data were based on sex. We acknowledge that gender and sex are not always congruent and ideally should not be used interchangeably, though in this study, some benchmark data based on gender were not available (eg, US population). As noted in the Methods section, the AAMC benchmarks used were publicly available, and we contacted AAMC for guidance regarding the available data. Benchmarks for gender date back further than for race and ethnicity. Benchmarks for other identity characteristics such as disability or sexual orientation or gender identity were not available. An important limitation of this study is the inability to account for other identity characteristics. We were unable to account for the timeframe regarding when the authors were selected versus when the CPG was published. We used benchmarks from pathology medical faculty (surrogate for academic pathologists), although CPG author panels often include individuals from other work settings. However, CPG authors are typically currently in academia, have an affiliation with academia, or have previously been associated with academia. Importantly, we did not study causality for the disparities documented, and this is an important area of future research.
CONCLUSIONS
In conclusion, though women overall are not significantly underrepresented among pathology CPG author positions, women physicians are significantly underrepresented as compared to AAMC benchmarks for women physicians in academic pathology. Conversely, White men physicians are significantly overrepresented in all author roles, while physicians from racial and ethnic groups are variably represented. Additional research is required to understand the implications of these findings on individual careers and pathology guidelines themselves.