Racial Differences in Work Relative Value Units for Outpatient Pediatric Dermatology Visits

Abstract

Measures of clinical productivity may tilt financial incentives toward providing care to specific patient populations, reinforcing inequitable care. Race, sex, and age have been shown to influence differences in work relative value units (wRVUs) generated during adult dermatology encounters. We sought to identify the association of patient race and sex with wRVUs generated by outpatient pediatric dermatology encounters. A total of 22,434 encounters among 10,063 unique patients were included; most encounters were with female (12,639 [56.3%]) and White (13,036 [58.1%]) individuals. The mean (SD) wRVUs per encounter was 1.57 (0.76). Visits with Asian patients, Black patients, and patients of other races generated 0.21 (95% confidence interval [CI] = 0.16–0.25), 0.19 (95% CI = 0.16–0.22), and 0.14 (95% CI = 0.10–0.17) fewer wRVUs per encounter, respectively, than pediatric outpatient dermatology visits with White patients (P < .001 for all). There was no significant association between sex and wRVUs per encounter. Laser treatment for vascular lesions was a mediator for the observed differences in wRVUs by race (60.5% [95% CI = 45.8–75.1%] for Asian race, 62.4% [95% CI = 52.9–71.9%] for Black race, and 53.3% [95% CI = 31.7–74.9%] for other races). Variations in wRVUs by patient race may create inadvertent incentives at the health systems level to prioritize certain groups.

Introduction

The current medical reimbursement system was created in 1989 when Congress established relative value units (RVUs) as the mechanism to standardize payments made by Medicare for medical services rendered across the United States (Quinton et al, 2024). Every Current Procedural Terminology (CPT) code is assigned a total RVU value. The total RVU is the sum of 3 components: the physician work RVU (wRVU), the practice expense RVU, and the malpractice RVU. The wRVU captures the effort that is involved with the service itself—the amount of time it takes to carry out, the technical skills required, and the level of mental effort and judgement necessary to deliver it—whereas the practice expense RVU and malpractice RVU account for the environment in which the service was provided and the liability risk undertaken by the clinician, respectively (Quinton et al, 2024). The Specialty Society Relative Value Scale Update Committee was formed by the American Medical Association to advise the Centers for Medicare and Medicaid Services on changes to the relative RVU weights of current and new CPT codes. The Centers for Medicare and Medicaid Services accepts the vast majority of the Relative Value Scale Update Committee's recommendations with wide-ranging effects (Laugesen et al, 2012). The weights used by Medicare have a substantial influence over the rates paid by private insurers (Clemens and Gottlieb, 2017). At the clinician level, wRVUs are often used as a measure of productivity and determinants of financial incentives in employment contracts (Nurok and Gewertz, 2019).

With such sweeping implications, decisions made by the Relative Value Scale Update Committee can create significant financial incentives for specific CPT codes. Higher RVU weights for services with differential utilization based on patient characteristics, such as sex or race, could reinforce—or even incentivize—structural inequities within health care. These may contribute to the well-documented racial inequities that continue to burden the United States’ healthcare system: racial minorities experience decreased access to care, worse outcomes, and increased healthcare costs (Smedley et al, 2003). These differences extend to the field of pediatric dermatology (Kuo et al, 2021). Beyond the considerations of access and direct cost to patients, race also plays a role in treatment selection (Barbieri et al, 2020; Fleischer et al, 2003). A myriad of policy changes has been proposed or implemented to mitigate these disparities, but few have done so by targeting the inequities in reimbursement for patient visits (DeMeester et al, 2017). Differences in wRVU generation have previously been described in the setting of adult dermatology. Outpatients visits with patients who were younger, female, or from racial minority groups generated fewer wRVUs than visits with older White men; the effect was mediated by procedures related to the diagnosis and treatment of malignant and premalignant lesions (Orenstein et al, 2021; Supapannachart et al, 2022a).

While malignant lesions are less common in pediatric dermatology, laser treatment is an analogous procedure that might be more commonly performed in individuals with lighter skin (Burkhart, 2020). Lasers are a mainstay in the management of port wine stains; they are also used for infantile hemangiomas and other vascular lesions (Loh et al, 2021). Laser treatment is not equally effective for all patients; those with darker skin types do not respond as well to pulsed-dye laser therapy and are more likely to experience adverse effects due to absorption of the laser by melanin (Sabeti et al, 2021). The epidemiology of the conditions for which laser is appropriate can also play a role in which patients receive laser treatment. Hemangiomas are more common in White and female patients, and although they are most often managed medically, the volume of hemangioma diagnoses could lead to a significant impact on the finances of a general pediatric dermatology practice (Burkhart, 2020; Sabeti et al, 2021; Stier et al, 2008; Supapannachart et al, 2022a).

In this work, we aimed to explore whether the racial and sex differences in wRVUs per encounter seen in adults are present in the pediatric dermatology population. If they do exist, could these differences be explained by laser destruction of cutaneous vascular lesions as an analogous procedure to (pre)malignant destruction in adults?

Results

The study sample included 22,434 outpatient pediatric encounters with a dermatology provider at Duke University between September 1, 2016 and December 31, 2020 that met the criteria for inclusion in this study (Figure 1). Patient and visit characteristics for outpatient encounters, stratified by race, are summarized in Table 1. Most encounters were with White patients (n=13,036; 58.1%), followed by Black patients (n=5176; 23.1%). Most of the patients were female (n=12,639; 56.3%), and the average age was 9.0 ± 5.8 years. Black patients had a lower percentage of commercial medical insurance (n=1819 [35.1%] vs n=10,084 [77.4%], P < .001) and a higher percentage of Medicaid insurance (n=3209 [62%] vs n=2395 [18.4%], P < .001) than White patients.

Figure 1.

Encounter eligibility flowsheet. CPT, Current Procedural Terminology; wRVU, work relative value unit.

Table 1.

Patient and Visit Characteristics for Outpatient Encounters in a General Pediatric Dermatologic Practice

Characteristics	Asian (n = 1143; 5.1%)	Black (n = 5176; 23.1%)	Othera (n = 3079; 13.7%)	White (n = 13,036; 58.1%)	Total (n = 22,434)
Sex
Female	662 (57.9%)	3027 (58.5%)	1703 (55.3%)	7247 (55.6%)	12,639 (56.3%)
Male	481 (42.1%)	2149 (41.5%)	1376 (44.7%)	5789 (44.4%)	9795 (43.7%)
Age, y; mean (SD)	9.0 (5.9)	8.6 (5.6)	8.3 (5.7)	9.3 (5.9)	9.0 (5.8)
Total RVUs charged
Mean (SD)	1.43 (0.45)	1.46 (0.46)	1.52 (0.61)	1.64 (0.89)	1.57 (0.76)
Median (Q1, Q3)	1.42 (0.97, 1.50)	1.50 (0.97, 1.88)	1.50 (0.97, 1.88)	1.50 (0.97, 1.88)	1.50 (0.97, 1.88)
Range	(0.35–5.19)	(0.48–5.29)	(0.48–6.21)	(0.18–12.28)	(0.18–12.28)
Insurance
Commercial	893 (78.1%)	1819 (35.1%)	784 (25.5%)	10,084 (77.4%)	13,580 (60.5%)
Medicare	5 (0.4%)	4 (0.1%)	0 (0.0%)	1 (0.0%)	10 (0.0%)
Medicaid	226 (19.8%)	3209 (62.0%)	2135 (69.3%)	2395 (18.4%)	7965 (35.5%)
Self-pay	20 (1.7%)	49 (0.9%)	74 (2.4%)	175 (1.3%)	318 (1.4%)
Otherb	15 (1.3%)	141 (2.7%)	114 (3.7%)	496 (3.8%)	766 (3.4%)
Hemangioma	86 (7.5%)	148 (2.9%)	371 (12.0%)	2247 (17.2%)	2852 (12.7%)
Other congenital vascular malformations	4 (0.3%)	21 (0.4%)	24 (0.8%)	239 (1.8%)	288 (1.3%)
Neoplasm of uncertain behavior	9 (0.8%)	28 (0.5%)	27 (0.9%)	328 (2.5%)	392 (1.7%)
Any skin cancer (melanoma + other malignant neoplasm)	0 (0.0%)	0 (0.0%)	0 (0.0%)	6 (0.0%)	6 (0.0%)
Laser treatment	7 (0.6%)	6 (0.1%)	68 (2.2%)	640 (4.9%)	721 (3.2%)
Premalignant lesion destruction	0 (0.0%)	1 (0.0%)	3 (0.1%)	23 (0.2%)	27 (0.1%)
Biopsy	11 (1.0%)	56 (1.1%)	30 (1.0%)	191 (1.5%)	288 (1.3%)
Malignant lesion destruction	0 (0.0%)	0 (0.0%)	0 (0.0%)	1 (0.0%)	1 (0.0%)
Benign excision	38 (3.3%)	181 (3.5%)	225 (7.3%)	1283 (9.8%)	1727 (7.7%)

Abbreviations: Q1, quartile 1; Q3, quartile 3; RVU, relative work unit.

Data are presented as count (percentage) unless otherwise noted. Percentages may not add to 100% owing to rounding.

^aOther includes American Indian or Alaskan native, native Hawaiian or other Pacific, 2 or more races, multiracial, or other race.

^bOther includes liability generic payor, special programs, worker’s comp, other government, and other payor classes.

In univariable analysis, the average wRVUs per encounter was significantly lower for Asian patients (−0.21, 95% confidence interval [CI] = −0.25 to −0.16), Black patients (−0.18, 95% CI = −0.20 to −0.15), and patients of other races (−0.12, 95% CI = −0.15 to −0.09; all P < .001) than for White patients (Table 2). After adjusting for sex, age, and insurance type, multivariable regression revealed similar results: the average wRVUs per encounter was significantly lower for Asian patients (−0.21, 95% CI = −0.25 to −0.16), Black patients (−0.19, 95% CI = −0.22 to −0.16), and patients of other races (−0.14, 95% CI = −0.17 to −0.10; all P < .001) than for White patients. A significant relationship (P < .001) was determined between age and wRVUs per encounter in univariable analysis, with each additional year resulting in fewer wRVUs (−0.013, 95% CI = −0.014 to −0.011); multivariable regression demonstrated a similar relationship. No significant difference in wRVUs per encounter was found between sexes. Likewise, no significant difference in wRVUs per encounter was found across insurance types in multivariable analysis.

Table 2.

Factors associated with wRVUs of outpatient pediatric dermatology encounters

Variable	RVUs per Encounter Mean (SD)	Univariable		Multivariable
		β (95% CI)	P-Value	β (95% CI)	P-Value
Age per year	NA	−0.013 (−0.014 to −0.011)	<.001	−0.014 (−0.015 to −0.012)	<.001
Sex
Male	1.58 (0.76)	Reference		Reference
Female	1.56 (0.75)	−0.01 (−0.03 to 0.01)	.30	−0.01 (−0.03 to 0.01)	.51
Race
Asian	1.43 (0.45)	−0.21 (−0.25 to −0.16)	<.001	−0.21 (−0.25 to −0.16)	<.001
Black	1.46 (0.46)	−0.18 (−0.20 to −0.15)	<.001	−0.19 (−0.22 to −0.16)	<.001
Other	1.52 (0.61)	−0.12 (−0.15 to −0.09)	<.001	−0.14 (−0.17 to −0.10)	<.001
White	1.64 (0.89)	Reference		Reference
Insurance type
Commercial	1.58 (0.83)	0.04 (0.01–0.06)	<.001	0.06 (−0.06 to 0.18)	.33
Medicaid	1.54 (0.61)	−0.04 (−0.07 to −0.02)	<.001	0.07 (−0.05 to 0.19)	.25
Self-pay	1.56 (0.68)	0.00 (−0.09 to 0.08)	.92	0.03 (−0.08 to 0.14)	.56
Other	1.64 (0.80)	0.08 (0.02–0.13)	.006	0.11 (−0.02 to 0.24)	.085

Abbreviations: CI, confidence interval; NA, not available; wRVU, work relative value unit.

In mediation analysis, laser treatment of cutaneous lesions accounted for 60.5% of racial differences (95% CI = 45.8–75.1%) for Asian race, 62.4% (95% CI = 52.9–71.9%) for Black race, and 53.3% (95% CI = 31.7–74.9%) for other races. Further mediation analysis was performed to identify the association of laser treatment with race differences in RVUs after adjusting for patient age, sex, insurance, and diagnoses of hemangioma and congenital malformations. After these adjustments, laser treatment accounted for 28.9% (95% CI = 13.5–44.3%) of racial differences for Asian race and 28.8% (95% CI = 19.6–38.0%) for Black race.

Discussion

Previous work in the adult dermatology population has demonstrated that encounters with older White males generate more wRVUs per encounter than encounters with other demographics; this result was largely driven by procedural charges for the biopsy and treatment of premalignant and malignant neoplasms (Orenstein et al, 2021). Although racial disparities in access to pediatric dermatology have been well-documented, differences in work productivity measures by patient race have, to our knowledge, been previously unreported. Specifically, the fact that Asian patients were just as, if not more, affected as Black patients compared with Whites is interesting even after adjustment for other demographics and insurance, especially because Asians are the fastest growing racial or ethnic group in the United States (Krogstad and Im, 2025). The aggregation of all Asians (East Asians, Southeast Asians, South Asians) into a single category may have contributed to this finding. Disaggregation at time of data capture would be a helpful future endeavor.

The impacts of race on wRVU compensation found in this study are consistent with those found in adults because both appear to be driven primarily by billing for procedures in their respective populations. Given that laser treatment continued to account for more than 25% of the racial differences even after adjusting for diagnoses, the effect is likely driven by more than just the epidemiology of these conditions. Notably, there were no differences in wRVUs per encounter across patient insurance types in multivariable analysis, a reassuring indicator that clinicians did not change coding—and potentially treatment practices—on the basis of real or perceived differences in approval or RVU conversion rates across insurances. Although there was also a significant relationship between patient age and wRVUs per encounter, the association was in the opposite direction as that found in adult studies (Orenstein et al, 2021). However, this was not the primary focus of our study. Moreover, although the coefficient for the age variable is statistically significant, it is not clinically meaningful.

RVUs were developed within Medicare for a specific purpose, but their use has extended beyond the original intent of standardizing payments. Private insurers, health systems, and others employers use wRVUs to access productivity and compensation with unintended implications for physician behavior, healthcare access, and patient outcomes (Armour et al, 2001). In clinical environments with a set number of daily appointments, large variation in wRVUs across patient race or sex may create inadvertent incentives at the health systems level to prioritize certain groups (Supapannachart et al, 2022b; Wang et al, 2020). When assigning wRVUs to CPT codes, the Relative Value Scale Update Committee intends to accurately capture differences in physician effort across services. Justifying whether the treatment of cutaneous vascular malformations with laser requires more expertise and effort than the management of other pediatric skin conditions—such as inflammatory skin diseases, which disproportionately impact racial minorities—is beyond the scope of this work (Shao et al, 2022). The variation in the underlying epidemiology of diseases and treatments appropriate for specific patients cannot be readily changed by clinicians. However, the presence of such differences necessitates greater diligence when balancing effort-based reimbursement for such services if we are to address the inequities present in dermatology.

It is important to note that the inequality in reimbursements across pediatric patients is smaller in magnitude than the previously reported gap in work productivity between pediatric and adult outpatient dermatology encounters (Supapannachart et al, 2022a). Given that the RVU system was originally intended for the elderly Medicare population, this is not surprising. Furthermore, pediatric-specific CPT codes are less likely to have undergone review for an RVU update (Massoumi et al, 2021). As a whole, such findings further highlight the importance of ensuring appropriate reimbursements for the unique considerations that are involved in the care of our youngest dermatology patients.

There are some limitations to this work. First, we examined data from a single academic center in the Southeast United States, which may not be representative of practices in other regions of the country, limiting the generalizability of our findings. In this dataset, we considered only outpatient pediatric encounters and excluded laser treatment visits that required more intensive sedation/anesthesia involving the use of special procedural rooms at the children’s hospital. Given the increased charges associated with such visits (greater complexity, anesthesia expenses, etc), including these encounters would likely have increased the wRVU differences by race. Although this study design demonstrated unequal generation of wRVUs, we are unable to determine whether those differences in wRVU compensation are equitable. In addition, we also limited the analysis to consideration of race only. As a result, ethnicity, defined as a group of people with a shared cultural background, was not a variable that was examined (Flanagin et al, 2021). Finally, we utilized retrospective billing data, which rely upon accurate coding of visits and can be subject to local coding practices. For example, CPT code 17999 is used in the Duke system for many different procedures, which led to the removal of visits with this code from our analysis because we could not reliably convert them to wRVU equivalents. Most of these visits were also self-pay, which may have affected the insurance-type variable in this work.

Although this work highlights differences in visit productivity measures that appear to be driven by procedural work, it is important to recognize that the study period extended only through 2020. In 2021, the Centers for Medicare and Medicaid Services updated the Physician Fee Schedule to increase wRVU compensation for evaluation and management services. These changes may have impacted the proportion of procedural visits relative to that of nonprocedural ones and may have altered the racial differences we found. Future work should re-examine this analysis once sufficient post-2021 billing data are available.

Our findings of racial differences in outpatient pediatric dermatology wRVU allocation further support the growing literature demonstrating structural factors that may disincentivize the provision of dermatologic care to specific demographic groups and, thereby, contribute to disparities in access (Barry and Hawryluk, 2022; Kuo et al, 2021; Prindaville et al, 2015). Our results build upon previous work suggesting that dermatologic procedures that are less likely to be used in racial minorities, and yet command higher wRVUs, may drive racial differences in reimbursement of dermatology care. Although patient characteristics may be a necessary consideration when pursuing certain treatments, higher compensation for such services may reinforce structural inequities in the field of pediatric dermatology. Further research should investigate whether these differences persist after the 2021 changes to the Physician Fee Schedule and the correlation between clinician effort and current RVU value.

Nonetheless, if higher compensation is tied to procedural services, perhaps this represents an incentive to find ways that lasers can be more effective in patients with darker skin, as may be the case with certain Asian and Black patients. Other potential methods to spur innovation could include a voucher program—similar to the rare pediatric disease priority review voucher program for pharmaceuticals—for expedited review of novel devices addressing underserved diagnoses and populations or a limited-time CPT modifier code for specific new devices to account for the additional education and time investment required for clinicians utilizing the latest technologies (U.S. Food and Drug Administration, 2024). Caring for pediatric patients involves thoughtful analysis of many unique factors; effort-based reimbursement systems should account for these in a way that appropriately incentivizes the delivery of high-quality dermatologic care to all patients.

Materials and Methods

Study population and design

Billing and demographic data were collected for all pediatric outpatient encounters with a dermatology provider at Duke University from September 1, 2016 to December 31, 2020. Exclusion criteria were patients aged >18 years; patients missing data on age, race, or sex; nursing, pathology, research, scheduling, laser hair removal, postoperative, cosmetic, video, and telephone visits; visits with a catch-all CPT code; and visits generating ≤0 wRVUs. Visits with the catch-all CPT code (17999) were excluded owing to a Duke coding practice that assigned this code to a range of different procedures—from laser hair removal to syringoma removal—many of which were self-paid, making systematic conversion to a wRVU equivalent difficult (Figure 1). Of these excluded encounters, the vast majority were with White patients (94; 84.7%) with the remainder being with patients of other (14; 12.6%) and Asian (3; 2.7%) races.

The primary outcome was wRVUs generated per encounter. Information on race was drawn from the electronic medical record. Race was self reported and categorized as Asian, Black, other, and White. The other race category included patients whose selected race in the electronic medical record was one of the following: American Indian or Alaskan native, native Hawaiian or other Pacific, 2 or more races, multiracial, or other race. Insurance type was categorized as self-pay, commercial, Medicare, Medicaid, or other. Procedures were identified with CPT codes, and diagnoses were identified using the International Classification of Diseases, Tenth Revision and Ninth Revision.

Statistical analysis

Univariable and multivariable fixed effects linear regression models were utilized to study the association between total RVUs per encounter and race, age, sex, and insurance type. Individual patients may have more than 1 insurance; therefore, each insurance type was analyzed as a dichotomous variable (yes vs no). Regression models are presented as the regression coefficient (β) with a 95% CI.

Mediation analysis through the difference method was performed to identify the association of laser treatment of cutaneous vascular lesions (CPT codes 17106–17108) with the observed race differences in wRVUs after adjusting for age, sex, and insurance; 95% CIs were estimated using bootstrapping with 200 resamples with replacement (VanderWeele, 2016). An additional mediation analysis was performed to identify the association of laser treatment with race disparities in wRVUs after adjusting for age, sex, insurance, and diagnoses of hemangioma and congenital malformations. All statistical tests were 2 sided, and the significance of each test was assessed at alpha = 0.05 without accounting for multiple testing. Statistical analyses were performed using SAS, version 9.4 (SAS Institute, Cary, NC).

Ethics Statement

The Duke University Health System Institutional Review Board approved this study and waived participant consent because the research carried no more than minimal risk, could not be conducted without the requested waiver, and did not adversely affect the rights and welfare of participants.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available owing to privacy or ethical restrictions.

ORCIDs

Udeyvir S. Cheema: http://orcid.org/0000-0003-3842-3482

Beiyu Liu: http://orcid.org/0000-0002-6706-9729

Ethan Borre: http://orcid.org/0000-0002-8666-5382

Cynthia L. Green: http://orcid.org/0000-0002-0186-5191

Lauren A. V. Orenstein: http://orcid.org/0000-0003-0534-2854

Suephy C. Chen: http://orcid.org/0000-0002-0678-7380

Conflict of Interest

LAVO is a board member of the HS Foundation, participated on advisory boards for Novartis and UCB, was a consultant and investigator for Chemocentryx, and received a research grant from Pfizer. The remaining authors state no conflict of interest.

accepted manuscript published online XXX; corrected proof published online XXX