Health care practitioners’ confidence assessing lupus-related rashes in patients of color

Abstract

Background: Patients with skin of color (P-SOC) are disproportionately burdened by lupus and often have worse disease outcomes than white patients. This is partly because educational materials underrepresent P-SOC, thereby promoting unconscious bias and clinical deficiencies among practitioners.

Purpose: We sought to measure providers' confidence in diagnosing the cutaneous manifestations of lupus (i.e., lupus-related rashes) in P-SOC and to assess which factors influenced their confidence.

Research Design: We created and distributed a survey that gathered information about participants' personal characteristics, clinical specialty, training, and current practice as well as measuring their confidence assessing lupus-related rashes in various skin tones.

Study Sample: Practitioners from the fields of rheumatology, dermatology, and internal medicine in the greater St. Louis area (Missouri, USA) participated in the survey.

Analysis: We compared practitioners' mean confidence levels assessing lupus-related rashes in patients with fair skin and P-SOC with a linear mixed effects model and used univariate and multivariate linear regression models to determine if the aforementioned factors correlated with confidence.

Results: Participants' mean confidence in diagnosing lupus-related rashes in P-SOC was significantly lower than assessing such findings in patients with fair skin (p = .009). Several factors correlated with confidence level at a univariate level; however, the multivariate model revealed experience as the only factor significantly associated with confidence (p = .001).

Conclusions: Providers report significantly less confidence assessing lupus-related rashes in P-SOC than in patients with fair skin. Our analysis demonstrates that experience positively correlates with confidence and suggests that interventions which enhance practitioners' exposure to and experience with these rashes in P-SOC can improve clinical confidence as well as patient outcomes.

Keywords

cutaneous lupus discoid lupus systemic lupus erythematosus health equity medical education

Introduction

Lupus is a systemic condition fraught with health disparities. In the United States, patients with skin of color (P-SOC) experience lower quality health care and worse health outcomes compared to white patients.¹ Lupus occurs two-to-three times more often among women of color, and minorities develop symptoms at younger ages with more serious complications.² African-American ethnicity, malar rashes, and discoid lesions can predict long-term organ damage and mortality,³ making it essential that clinicians accurately diagnose and promptly treat lupus in P-SOC.

Medical education contributes to the health disparities experienced among P-SOC with lupus. Educational materials underrepresent P-SOC,⁴ a bias extending to resources specific to lupus,⁵ and training programs contain few curricula dedicated to cutaneous findings in P-SOC.⁶ These resources and learning experiences prepare practitioners to care for patients, and they impart unconscious biases while creating deficiencies in clinical skills when they exclude P-SOC.^7,8 This can delay diagnosis for patients, can expose them to misdiagnosis, and can promote their distrust in the health care system.⁹

Given the importance of recognizing lupus among P-SOC and the consequences of biases in education, we investigated practitioners’ confidence assessing lupus-related rashes in P-SOC. Secondarily, we analyzed which predictors are associated with greater clinical confidence and whether practitioners would like more teaching about these rashes in P-SOC.

Methods

We developed an anonymous survey to measure providers’ confidence assessing lupus-related rashes in P-SOC as well as practitioners’ characteristics that might influence confidence. Questions addressed career specialty, type and duration of practice, and practitioner race and ethnicity as well as the frequency of patients’ race and ethnicity encountered in clinical practice. We included factors that previous studies indicated may contribute to provider confidence when caring for P-SOC (9, 10, and 11). On a four-point Likert scale (1 (“none”), 2 (“very little”), 3 (“moderate”), and 4 (“a lot”)), providers ranked their confidence, experience, and previous training assessing any rash, any lupus-related rash, lupus-related rashes in patients with fair skin, and lupus-related rashes in P-SOC. We used the term “lupus-related rash” to reflect the variety of skin manifestations within the domain of cutaneous lupus as well as cutaneous findings associated with lupus not specifically classified as cutaneous lupus (e.g. urticaria). We also intended “lupus-related rash” to be a descriptive term for participants who may not be well-versed in the study of cutaneous lupus. We defined training as “education (such as nursing, physician assistant, or medical school, residency, fellowship, or continuing medical education) that included tools such as reading, image reviews, or lectures that assisted practitioners in diagnosing lupus-related rashes in P-SOC.” Our final question asked if providers would like more education in assessing lupus-related rashes.

In February and March 2020, we invited health professionals in the fields of rheumatology, dermatology, and internal medicine from the greater St. Louis area (Missouri, United States) to complete the survey. We shared surveys electronically via email listservs of rheumatology fellows and providers, dermatology residents and practitioners, primary care providers, and internal medicine residents affiliated with Washington University School of Medicine and its community network as well as in-person at rheumatology continuing medical education conferences. We instructed practitioners to take the survey once if they received multiple invitations. We planned to invite additional providers in primary and hospitalist care, but the onset of the COVID-19 pandemic ended recruitment. Of the 480 invited providers, 132 (27.5%) answered the survey. This value underestimates the response rate as the anonymity of our study precluded reviewing the names of invitees and correcting for duplicate requests. Responses were stored in the Qualtrics survey platform (Seattle, Washington). The study protocol was in compliance with the World Medical Association Declaration of Helsinki, and the Washington University School of Medicine Institutional Review Board approved our study with waived informed consent (approval 202001144).

A linear mixed effects model compared practitioners’ mean confidence levels among any rash, any lupus-related rash, lupus-related rashes in fair skin, and lupus-related rashes in P-SOC. This model accounted for correlated ratings of confidence from the same practitioner and assumed missing data at random; incomplete data was used in the mixed model analysis. Linear contrasts compared mean confidence level pairwise. A univariate linear regression model assessed the association between confidence assessing lupus-related rashes in various skin tones and the other potential predictors. Predictors with p-values smaller than 0.2 in the univariate analysis were incorporated in the multivariate linear regression model analysis. We used analysis of covariance (ANCOVA) with a backward selection method to construct the final model. We examined for multicollinearity with variance inflation factors and condition indices. Last, we determined least squares confidence means for each Likert scale level of experience (i.e., “none,” “very little,” “moderate,” and “a lot”) in diagnosing lupus-related rashes in P-SOC and calculated the difference of least squares confidence means between pairwise comparisons of experience levels. All the statistical tests were two-sided, and analyses were performed with SAS 9.4 (SAS Inc, Cary, NC).

Results

Internal medicine was the most-represented clinical specialty (43.2%), with internal medicine resident trainees comprising 28.0% of all respondents (Table 1). Over half of practitioners (54.5%) reported fewer than 5 years in clinical practice, and most participants (71.2%) identified as white.

Table 1.

Practitioners’ personal and practice characteristics.

Characteristics	Number of responses
Practitioner titles
Nurse practitioner/physician assistant	12
Resident or fellow	47
Attending physician	67
Other^a	6
Years in practice^b
0 to 5	72
6 to 10	13
11 to 15	4
16 to 20	8
21+	34
Practice environment
Private practice	39
Academic center	82
VA hospital	20
Other	9
Urban	94
Suburban	49
Rural	11
Inpatient only	6
Outpatient only	55
Combined Inpatient and Outpatient	75
Practitioners identifying as Hispanic, Latino, or of Spanish origin	6
Practitioner race^c
Asian	29
Black or African American	5
White	94
Other	3
American Indian or Alaskan Native	0
Native Hawaiian or other Pacific Islander	0
Patient race or ethnicity most often encountered in practice
Black or African American	46
White	84

^a Those that elected to fill in OTHER include Family Medicine (5), Medicine-Pediatrics (1), Orthopedics (1), Surgery (1).

^b One participant left this answer blank.

^c One participant left this answer blank.

Mean confidence was lowest for diagnosing lupus-related rashes in P-SOC (Table 2). Among all study participants, linear mixed effects model analysis demonstrated that the estimated mean confidence level diagnosing any rash was significantly higher than confidence assessing lupus-related rashes in general, lupus-related rashes in patients with fair skin, and lupus-related rashes in P-SOC (p = .035, .049, < .0001, respectively). The mean confidence diagnosing lupus-related rashes in patients with fair skin was significantly higher than confidence diagnosing lupus-related rashes in P-SOC (p = .009). Practitioner confidence diagnosing lupus-related rashes in general compared to lupus-related rashes in patients with fair skin were similar (p = .88). Spearman’s rank correlation coefficients among mean confidence levels in diagnosing all four subgroups of rashes ranged from 0.59 to 0.93 (p < .0001), suggesting that providers reporting a high confidence level diagnosing any rash were likely to rate the other types with a high confidence level.

Table 2.

Practitioners’ mean Likert scale responses for confidence, training, and experience diagnosing rashes.

Characteristic	Mean ± SD
How confident are you in diagnosing
Any rash	2.61 ± 0.81
Any lupus-related rash	2.39 ± 0.84
Lupus-related rashes in fair skin	2.40 ± 0.85
Lupus-related rashes in skin of color	2.12 ± 0.91
To what extent did your training include information or tools to assist you in diagnosing
Any rash	3.04 ± 0.80
Any lupus-related rash	2.89 ± 0.85
Lupus-related rashes in fair skin	2.82 ± 0.91
Lupus-related rashes in skin of color	2.42 ± 0.98
How much experience do you have in diagnosing
Any rash	2.95 ± 0.83
Any lupus-related rash	2.59 ± 0.87
Lupus-related rashes in fair skin	2.52 ± 0.89
Lupus-related rashes in skin of color	2.23 ± 0.88

The univariate analysis between provider confidence diagnosing lupus-related rashes in P-SOC and predictive factors identified several significant associations (Table 3). Provider specialty, years in practice, extent of training, clinical experience, and treating either white or African American patients most frequently correlated with confidence level diagnosing lupus-related rashes in P-SOC (p < .05). After performing multivariate linear regression with ANCOVA, experience was the only predictor that correlated with confidence (p = < .001). Multicollinearity was not detected in the model (variance inflation factors <4, condition index <30).

Table 3.

Univariate association analysis with provider confidence assessing lupus-related rashes in patients with skin of color (P-SOC).

Factors	F or t value	p	Effect size
Specialty	28.25	<.0001	0.404^a
Years in practice	9.98	<.0001	0.245^a
Practice environment	2.39	.0709	0.048^a
Provider race or ethnicity	0.20	.8935	0.005^a
Extent of training in diagnosing lupus‐related rashes in P-SOC	38.19	<.0001	0.478^a
Experience in diagnosing lupus‐related rashes in P-SOC	82.56	<.0001	0.665^a
Black or African American patients most frequently seen in clinical practice	3.67	.0004	0.657^b
White patients most frequently seen in clinical practice	−2.47	.0148	0.468^b

P-SOC, patients with skin of color.

^a η2 reported for one-way ANOVA.

^b Cohen’s d reported for two-sample t-test.

The difference of least squares confidence means between pairwise comparisons of levels of experience showed that for every iterative increase in experience, there is an associative rise in confidence with assessing lupus-related rashes in P-SOC, indicating a dose–response relationship between experience and confidence: the larger the increase in experience, the greater the gain in confidence (supplement).

When asked about more training to diagnose lupus-related rashes in P-SOC, 93% of participants (N = 132) indicated interest in additional education.

Discussion

Our study demonstrates that practitioners have lower confidence diagnosing lupus-related rashes in P-SOC than in patients with fair skin. Clinicians reported the least amount of confidence assessing rashes in patients with lupus and skin of color compared to those with any rash, any lupus-related rash, or lupus-related rashes in patients with fair skin. Univariate analysis demonstrated associations between several predictors and confidence, and experience with patients of color and lupus-related rashes uniquely correlated with provider confidence in the multivariate model. These findings represent the first investigation of practitioners’ confidence assessing rashes specific to lupus in P-SOC and of predictive factors that may influence provider confidence.

Our results complement prior research that highlights practitioners’ low confidence and clinical skills diagnosing cutaneous disorders in P-SOC. Dermatology providers previously reported less confidence evaluating common skin diseases among P-SOC,^8,11 and an assessment of medical students’ clinical abilities revealed less skill naming squamous cell carcinoma, urticaria, and contact dermatitis in P-SOC.¹² These findings suggest that practitioners lack both confidence and clinical skill when caring for P-SOC, highlighting provider-specific deficiencies that contribute to poor health outcomes.

The intersection of skin color and a diagnosis of lupus holds clinical importance. Our findings indicate that practitioners have the least confidence assessing rashes in patients with a diagnosis of lupus and skin of color compared to any rash or lupus-related rash in patients with fair skin. P-SOC with lupus experience delays in diagnosis, more severe complications, and worse health outcomes compared to white patients.^2,3 Therefore, it is critical that providers are exposed to learning experiences that develop their clinical confidence and acumen in recognizing the manifestations of lupus among P-SOC.

Previous research has shown that curricular modules can improve clinicians’ confidence in diagnosing rashes in P-SOC.⁷ In our study, educational tools from practitioners’ training predicted confidence in the univariate model but were not associated with confidence in the multivariate model. This implies that educational materials dispensed during an initial training period insufficiently prepare clinicians for treating lupus-related rashes in P-SOC in long-term practice.

On the other hand, we found that continued experience improved confidence. This factor represents more than years in practice or clinical specialty because neither career duration nor specialty was associated with practitioners’ confidence in the multivariate model. Instead, it may more accurately reflect the types of patient experiences encountered in clinical practice. Trainees reflect that spending time in dermatology clinics dedicated to the care of P-SOC improves their diagnostic acumen in assessing cutaneous findings in P-SOC.¹³ Additionally, clinics specialized in the care of patients with lupus also improve patients’ satisfaction, trust in health care providers, and health outcomes.^9,10,14 In combination with dedicated didactic teaching, such clinical experiences increase trainees’ exposure to less common patient populations and conditions, thereby consolidating their learning and enhancing their training. Independently practicing clinicians likely would appreciate similar opportunities in continuing medical education as more than 90% of survey respondents voiced interest in additional education about lupus-related rashes in P-SOC.

We included representatives from multiple specialties, namely, rheumatology, dermatology, and internal medicine, to strengthen the results of our study. We selected these specialties because they frequently treat patients with lupus and because it improves the generalizability of our results. Furthermore, our regression analyses measured the association of practitioner confidence with factors previous studies indicated might predict confidence in diagnosing lupus-related rashes in P-SOC, thereby grounding our methods in previous research.^9–11

Our study has several limitations. It was underpowered to detect an association between practitioner race and ethnicity with clinical confidence. We planned to recruit additional providers in hospital medicine and primary care but were limited by the COVID-19 pandemic. Additionally, internal medicine resident trainees comprised almost one-third of participants. With limited knowledge and insight into their abilities, novices can overestimate their skills,¹⁵ which may have influenced their survey responses. Last, this study evaluated confidence, which does not equate to clinical skill.

In future research, our approach should be applied to dermatologic manifestations of other autoimmune rheumatic diseases where P-SOC suffer worse outcomes than white patients. Educators should develop assessment tools to quantify practitioners’ competence assessing rashes in P-SOC to facilitate education and training. Also, medical educators should investigate which educational and clinical training interventions best prepare providers and trainees to care for P-SOC with lupus, paying close attention to clinics that specialize in the care of lupus and P-SOC where these approaches most fertilely intersect.

These findings show that a significant difference exists in provider confidence diagnosing lupus-related rashes in patients with fair skin compared to P-SOC among rheumatologists, dermatologists, and internists. Medical resources and training experiences that favor fair skin contribute to this finding as additional clinical experience with P-SOC is associated with greater confidence. Therefore, a combination of instructional teaching as well as clinical experiences with patients of color could improve practitioners’ confidence and patients’ health outcomes. In devising strategies to enhance practitioners’ experience with lupus and P-SOC, the profession of medicine can reduce the health inequities experienced by this vulnerable patient population.

Footnotes

Acknowledgments

We thank Drs. Aisling Doyle, MD, Iris Lee, MD, Muithi Mwanthi, MD, PhD, and Andrea Ramirez-Gomez, MD for helping us develop our survey. We thank Drs. Dominique Cosco, MD, Heather Jones, MD, and Prabha Ranganathan, MD, MSCI for supporting survey distribution.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Washington University in St Louis Mentors in Medicine Program [grant number 2020].

ORCID iD

Lisa Zickuhr

References

Feagin

Bennefield

. Systemic racism and U.S. health care. Soc Sci Med 2014; 103: 7–14.

Somers

Marder

Cagnoli

, et al. Population-based incidence and prevalence of systemic lupus erythematosus: the Michigan Lupus Epidemiology and Surveillance program. Arthritis Rheumatol 2014; 66: 369–378.

Babaoglu

Goldman

, et al.. Predictors of predominant Lupus Low Disease Activity State (LLDAS-50). Lupus 2019; 28: 1648–1655.

Louie

Wilkes

. Representations of race and skin tone in medical textbook imagery. Soc Sci Med 2018; 202: 38–42.

Rana

Witt

Jones

, et al.. The Representation of Skin Colors in Images of Patients with Lupus Erythematosus. Arthritis Care Res (Hoboken). DOI: 10.1002/acr.24712.Epub ahead of print 31 May 2021.

Nijhawan

Jacob

Woolery-Lloyd

. Skin of color education in dermatology residency programs: does residency training reflect the changing demographics of the United States?. J Am Acad Dermatol 2008; 59: 615–618.

Fourniquet

Garvie

Beiter

. Exposure to Dermatological Pathology on Skin of Color Increases Physician and Student Confidence in Diagnosing Pathology in Patients of Color. FASEB J 2019; 33: 606–618.

Rodrigues

Ross

Gilmore

and et al.. Australian dermatologists' perspective on skin of colour: Results of a national survey. Australas J Dermatol 2018; 59: e23–e30.

Gorbatenko-Roth

Prose

Kundu

and et al.. Assessment of Black Patients' Perception of Their Dermatology Care. JAMA Dermatol 2019; 155: 1129–1134.

10.

Arora

Nika

Trupin

, et al. Does Systemic Lupus Erythematosus Care Provided in a Lupus Clinic Result in Higher Quality of Care Than That Provided in a General Rheumatology Clinic?. Arthritis Care Res (Hoboken) 2018; 70: 1771–1777.

11.

Buster

Stevens

Elmets

. Dermatologic health disparities. Dermatol Clin 2012; 30: 53–viii.

12.

Fenton

Elliott

Shahbandi

, et al. Medical students' ability to diagnose common dermatologic conditions in skin of color. J Am Acad Dermatol 2020; 83: 957–958.

13.

Cline

Winter

Kouroush

, et al. Multiethnic training in residency: a survey of dermatology residents. Cutis 2020; 105: 310–313.

14.

Tull

Kerby

Subash

and et al.. Ethnic skin centers in the United States: Where are we in 2020?. J Am Acad Dermatol 2020; 83: 1757–1759.

15.

Kruger

Dunning

. Unskilled and unaware of it: How difficulties in recognizing one's own incompetence lead to inflated self-assessments. J Personal Soc Psychol 1999; 77: 1121–1134.