OBJECTIVE: To assess whether underrepresented minority (URM) pediatricians serve minority and/or publicly insured or uninsured children to a greater degree than non-URM pediatricians, and to evaluate trends in these relationships.
PARTICIPANTS AND METHODS: Respondents to American Academy of Pediatrics surveys in 1993, 2000, and 2007, provided their race/ethnicity, practice locations, and estimated race/ethnicity and insurance sources of their patient panels. We classified pediatricians by race/ethnicity as either URM (black, Hispanic, Hawaiian/Pacific Islander, or American Indian/Alaskan Native) or non-URM (white, non-Hispanic/Latino, Asian, or other). We compared the average percentages of minority individuals (excluding Asians) in the patient panels of URM and non-URM pediatricians, and compared the average percentages of publicly insured (includes TRICARE) or uninsured patients.
RESULTS: Response rates were >50% for all surveys. There were 1003 (10.2% URM) postresidency respondents in 1993, 707 (11.8% URM) in 2000, and 900 (11.1% URM) in 2007. In all 3 surveys, the average percentage of minority children in URM pediatricians' patient panels was higher than for non-URM pediatricians by approximately 20 percentage points (all P < .001). The differences in the percentages of patients who were either publicly insured or uninsured widened over the study period, from 46% in URM pediatric practices versus 38.8% in non-URM practices in 1993 to 59.7% vs 40.7%, respectively, in 2007 (P < .001). In multivariable analyses, URM pediatricians' patient panels had a higher mean percentage of minorities (15.2 percentage points greater; 95% confidence interval [CI]: 12.1–18.4) and a higher mean percentage of publicly insured and uninsured patients (10.9 percentage points greater; 95% CI: 6.7–15.4) compared with non-URM pediatricians' patient panels.
CONCLUSIONS: Over the last 14 years, URM pediatricians were more likely than their non-URM peers to care for minority children and publicly insured or uninsured patients. This relationship has not appeared to attenuate over time, and these data support ongoing and enhanced efforts to recruit qualified URMs into pediatric careers to meet the growing needs of the expanding minority child population.
WHAT'S KNOWN ON THIS SUBJECT:
Historical data suggest that URM pediatricians disproportionately provide care for minority or poor children, but that the healthcare marketplace and child population continue to change.
WHAT THIS STUDY ADDS:
This study demonstrates that URM pediatricians continue to disproportionately provide care to minority and poor children. Over the past 14 years, URM pediatricians have experienced a large increase in the percentage of their publicly insured or uninsured patients.
Several racial or ethnic populations are considered to be underrepresented in medicine, including African Americans, Hispanics/Latinos, Native Americans, and Native Hawaiian/Pacific Islanders. The imbalance between the number of these underrepresented minority (URM) providers and the number of patients of these races or ethnicities in the general population have real health care implications. A diverse workforce will likely provide more culturally appropriate care and increased access to care for patients of minority race/ethnicity.1 However, the potential benefit of the improved access to care for minority or underserved children provided by higher numbers of URM physicians has received perhaps the most attention by research studies. In a systematic review, Saha and Shipman identified 14 studies demonstrating that URM physicians were more likely than non-URM physicians to provide care for minority patients or poor patients (usually defined by insurance type or status).2 In addition, evidence has suggested that care is improved, at least from the patient standpoint, when patients and physicians are of the same race/ethnicity.3,–,6 Only 1 study has evaluated these relationships among pediatricians. In 1996, Brotherton et al surveyed more than 1000 randomly sampled US pediatricians. The study oversampled minority pediatricians and demonstrated that URM pediatricians reported higher percentages of minority patients in their patient panels and higher percentages of publicly insured or uninsured patients than did non-URM pediatricians.7
The data supporting the tenet that URM physicians are more likely to provide care to underserved populations were largely published in the 1990s, and used data from as far back as the 1980s. Since the seminal studies in this field were published, several factors may have altered the relationship between URM-provider status and the likelihood of caring for more minority or underserved children. First, data from 1987 and 1994 began to demonstrate changes in practice specialty choices for URM medical students, including changes away from a greater inclination toward primary care specialties.8,9 In addition, data from the Association of American Medical Colleges (AAMC) demonstrated that although the United States experienced increasing numbers of minority science-major college graduates in the 1990s and early 2000s, the number of those graduates who chose medical careers remained stagnant, suggesting that nonmedical career opportunities seemed to be more attractive to more recent minority college graduates who majored in science.10,11 In addition, there is ongoing evidence that the interest in medicine has experienced greater declines among minority premedical college students than the declines among majority students.12
Taken together, all of these factors suggest that the interest in medical careers among minority science undergraduates and medical career options and practice patterns for minority medical school graduates have changed over the past 2 decades, which raises the question of how much the practice patterns of current URM pediatricians might have changed. Our study evaluated the changes in practice demographics of URM pediatricians over a 14-year period from 1993 to 2007. We had 2 research questions: first, do URM pediatricians have patient panels that are more diverse, as defined by percentages of minority patients or percentages of patients with public insurance (including TRICARE) or no insurance? Second, how have the patient panels of URM pediatricians changed over time?
PARTICIPANTS AND METHODS
Study Design and Target Population
This study is a cross-sectional assessment from 3 distinct points in time: 1993, 2000, and 2007. At a summative level, these data provided a longitudinal assessment of the service patterns of pediatricians and their practices over the 14-year study period. We examined data from the American Academy of Pediatrics' (AAP) Periodic Surveys (PS) of Fellows from 1993 (Periodic Survey #21), 2000 (Periodic Survey #43), and 2007 (Periodic Survey #67). Each periodic survey was pilot tested and approved by the AAP institutional review board. The surveys were mailed to a unique sample of 1616 members in 1993, 1602 in 2000, and 1605 in 2007. The participants were randomly selected AAP fellows in the United States. Approximately 80% of the board-certified pediatricians in the United States are members of the AAP. For each survey, up to 6 follow-up mailings were sent to nonresponders.
Respondents identified their racial/ethnic group, gender, birth year, and practice location. We classified pediatricians by race/ethnicity into URM (black/African American, Hispanic/Latino, Hawaiian/Pacific Islander, or American Indian/Alaskan Native) and non-URM (white, non-Hispanic/Latino, Asian, or other) groups. These surveys did not collect details on provider or patient racial/ethnic subgroups. Pediatricians who were still clinically active estimated the percentage of their patient panels who were covered under different insurance types and the percentage of their patient panels who fell into various racial/ethnic groups. Different insurance types included private insurance (both fee-for-service and managed care), public insurance (both fee-for-service and managed care), TRICARE (military), and no insurance (uninsured). Patient race and ethnicity categories included white, non-Hispanic/Latino; Hispanic/Latino; black/African American; Asian; Native Hawaiian/Other Pacific Islander; American Indian/Alaskan Native; and other. Black/African American, Hispanic/Latino, Hawaiian/Pacific Islander, and American Indian/Alaskan Native patients were considered as minority patients for the purposes of this study. We also completed analyses including Asian children in the minority category. Although this changed the absolute values (percentages of minority patients) in our comparisons, it did not change the qualitative differences in patient panels between the URM and non-URM providers. From a financial standpoint, we grouped together for analyses those patients with public insurance (including TRICARE) or those who were uninsured. Publicly funded and uninsured patients were grouped together because both groups have reduced health care access and rates of use compared with privately insured patients, although children with publicly funded insurance have better access to medical care compared with uninsured children.13,14 Finally, respondents identified their practice location as urban (inner city), urban (non-inner city), suburban, or rural.
Analyses for this study were limited to the respondents who had completed residency training. We calculated frequencies for both the respondent and patient-panel variables. We evaluated the racial or ethnic concordance between providers and patients, with all 3 survey years combined owing to the relatively small groups of Hispanic, black, or Asian physicians that would have resulted had we grouped by year. For all other bivariate comparisons, each year was analyzed separately by using t-tests. We also tested for linear trends across survey years using 1-way analyses of variance with linear contrasts. We conducted bivariate analyses to compare the average percentages of minority patients in URM pediatrician and non-URM pediatrician patient panels and also compared the average percentages of those patients who were either publicly insured or uninsured in the respective patient panels. For multivariate analyses, provider URM status was the criterion variable, controlling for survey year, age (dichotomized at 40 years of age), practice location, and gender. We completed linear regression models for the percentage of minority patients in a panel and the percentage of patients who had either public insurance or TRICARE, or were uninsured, including the criterion and control variables used in the models. We included first-order interactions in the regression models, but these additions did not substantially change the outcomes. Therefore, this article presents the results without the interaction terms.
The response rates of the 3 surveys were 70.9% (1993), 52.0% (2000), and 55.7% (2007). There were 1003 postresidency respondents in 1993, of which 10.2% were URM pediatricians. There were 707 respondents (11.8% URM) in 2000, and 900 respondents (11.1% URM) in 2007. Descriptive information about the respondents is summarized in Table 1. Women comprised 36.7% of the respondents in 1993, 45.3% in 2000, and 52.3% in 2007. The percentages of white respondents varied from 74.4% to 79.6%, depending on the year, and the percentages of Hispanic/Latino respondents varied from 4.5% to 5.1%. There was a significant increase in black respondents from 2.1% in 1993 to 4.9% in 2007. There were very few Native Hawaiian/Other Pacific Islander or American Indian/Alaskan Native providers among the respondents. There was a strong concordance between the pediatricians' race/ethnicity and patients' race/ethnicity. When the data were grouped across the 3 surveys, URM pediatricians were more likely than non-URM pediatricians to care for minority patient groups (Fig 1). The mean percentage of minority patients was significantly higher for Hispanic (57.9%) and black (57.6%) pediatricians than for white (33.4%) and Asian (40.6%) pediatricians (P < .001). There was no difference in the percentage of minority patients between the Hispanic and black pediatrician groups (P = .999). Although the overall percentage of minority patients was the same for the Hispanic and black groups, the relative compositions of the minority patient groups varied because more patients had the same race/ethnicity as their pediatricians.
In each of the 3 surveys, the average percentages of minority children in URM pediatrician-patient panels was higher than those for non-URM pediatricians by approximately 20 percentage points; this relationship did not attenuate over the time period studied. In 1993, URM pediatricians estimated that 57% of their patient panels were minority children versus 36.6% in the non-URM pediatrician panels. In 2000, URM pediatrician-patient panels were 56.6% minority patients versus 29.8% for non-URM pediatricians. Finally, in 2007, the percentages were 56.7% versus 37.2%, respectively (P < .001 for all 3 years) (Table 2). In addition, there were no changes over time in percentages of minority children in either URM or non-URM pediatrician patient panels (test for trend, P > .60 for both).
When examining the insurance status of the patients in a pediatrician's panel (Table 3), URM pediatricians had higher percentages of publicly insured or uninsured patients. This difference widened over the study period, which was accounted for by a significant increase in the number of these patients in URM pediatrician practices (test for trend, P < .05) and little change in non-URM pediatrician practices (test for trend, not significant). In 1993, there was a nonsignificant difference in the percentages of publicly insured or uninsured patients (46% in URM practices versus 38.8% in non-URM practices). By 2007, however, 59.7% of patients of URM pediatricians were publicly insured or uninsured compared with 40.7% in non-URM practices (P < .001) (Table 3).
In multivariable analyses accounting for survey year, provider age, practice location, and gender, the provider race/ethnicity had a strong association with the percentage of minority children in the patient panel. For example, patient panels of URM pediatricians had a higher percentage of minorities (15.2 percentage points greater; 95% CI for the difference: 12.1–18.4) compared with those of non-URM pediatricians. Providers in both categories of urban practice locations had higher percentages of minority patients. Male pediatricians cared for lower percentages of minority children (Table 4).
In comparing insurance status representation among practices, patient panels of URM pediatricians had higher percentages of publicly insured or uninsured patients (10.9 percentage points greater; 95% CI for the difference: 6.7–15.4) compared with patient panels of non-URM pediatricians. Respondents from both urban location types and rural respondents were significantly more likely to have higher percentages of publicly insured and uninsured patients (Table 5).
These data demonstrate that URM pediatricians continue to be more likely than nonminority pediatricians to provide care to children of racial/ethnic minorities, as well as to children (of any race/ethnicity) who have public insurance (includes TRICARE) or are uninsured. This association has not attenuated over the past 14 years. In fact, the URM respondents in later years reported even larger percentages of their patients who were from nonprivately insured populations than did the non-URM pediatricians. Our findings closely mirror the findings from the only other study of URM pediatrician practice patterns, which was completed by Brotherton et al.7 In that study, the authors found that URM pediatricians had patient panels with higher percentages of minority children (by 24 percentage points) and higher percentages of publicly insured or uninsured children (by 13 percentage points) compared with non-URM pediatricians. Our data from 1993 are roughly contemporary with that of Brotherton et al, but we extended the analyses through 2 subsequent time points (2000 and 2007), which has shown that the predisposition of URM pediatricians to disproportionately serve the needs of minority and publicly insured or uninsured youth relative to non-URM pediatricians has been persistent over time. Thus, URM pediatricians continue to play a critical role in meeting the health care needs of poor children.
Identifying providers who are committed to caring for low-income or minority children continues to be an increasingly important challenge. Population projections suggest that minority children are the fastest-growing child group in the United States.15 Furthermore, the growing number of working parents in the United States who are no longer eligible for employer-based insurance is leading to expanded rosters of publicly insured or uninsured children.16 Evidence demonstrates that physicians who care for high proportions of minority children face challenges in ensuring the delivery of quality care to those children, partly because of the financial pressures caused by poor reimbursement.17,18 In addition, the effects of current social and legal trends have worked against the effort to increase the entry of minority students into medical careers after undergraduate school and primary care fields after medical school.11,19,20 Our study supports the argument that recruiting URM providers into pediatrics is 1 method to increase access to care for growing populations of minority and publicly insured children. In addition, there is evidence that patients prefer race concordance with physicians, and this concordance has been shown to be associated with the quality of care.5,21 Our results demonstrate the striking concordance that exists between the race/ethnicity of pediatricians and their patient populations. Our findings suggest that efforts to recruit physicians into pediatric careers should focus on all minority providers, given their strong association with caring for minority children. The improved access and more culturally sensitive care offered by minority pediatricians can help achieve one of the overarching goals of Healthy People 2010, which is to reduce racial disparities in medical care.22
Our study amply demonstrates the well-documented increase in the number of publicly insured children. By 2007, publicly insured children comprised an estimated 38% overall in the patient panels of URM providers. This expansion is attributed in part to the expansion of the State Children's Health Insurance Program, along with individual state efforts to increase access.23,24 Despite this expansion of funding, the reimbursement shortcomings that accompany publicly funded insurance may continue to be a deterrent to providers.18 It is worth noting that URM pediatricians seem to be bearing a disproportionate share of the expansion of public insurance, with publicly insured and uninsured patients comprising nearly 60% of the URM patient panels in the 2007 survey, whereas the percentages were essentially unchanged for non-URM providers over the study period. We were unable to assess whether the practices qualified for federal cost-based reimbursement programs, but one must wonder how many of these practices are at a financial tipping point with such high percentages of publicly insured or uninsured patients.
Perhaps the most significant limitation of our data is the fact that the practice panel composition data were self-reported estimates. The survey was not designed to collect actual composition of the patient panel, and respondents were not required to justify their estimates. However, these percentages seem appropriate when compared with other national data. For example, year 2000 data from the National Ambulatory Medical Care Survey (NAMCS) demonstrated that 17.6% of the pediatric visits were by children of black race, which was almost identical to our results that showed 17.5% to 20% of the patients were black, depending on the year.25 NAMCS identifies Hispanic patients by ethnicity, and Hispanic patients can be of any race in NAMCS. Therefore, a direct comparison of the percentages of white and Hispanic participants from our data to NAMCS was not possible. On a population basis, Census Bureau data from 2004, compiled by the Child Trends DataBank, has demonstrated that 16% of the population of US children were black (vs 17.5%–20%, depending on the year, for our data) and 19% were Hispanic (compared with 13.8%–18.7%, depending on the year, for our data).26 Therefore, we believe that the estimates by the respondents to these surveys are in reasonable ranges that seem to be consistent with other national data. These data are also limited by the fact that they provide only a series of cross-sectional views and did not use the same respondents in each survey for a true time-series analysis. We were not able to control for whether respondents had participated in the National Health Service Corps, a factor that is intuitively associated with caring for more children with public insurance or no insurance. However, existing evidence has consistently shown that provider ethnicity has a stronger relationship with providing care for minority children than does National Health Service Corps participation.2 We were unable to account for the socioeconomic background of the providers and their educational debt, both factors that might be associated with practice location and focus on underserved populations.17 The response rates to the 3 surveys used for this study were 50% to 70%, rates that are admirable when compared with other national physician surveys.27 The survey respondents were all AAP fellows and may not have represented the complete national sample of pediatricians. To be sure, these data represent some of the largest data collected from pediatricians relevant to our research questions. Finally, these data represent aggregate characteristics of URM and non-URM pediatricians. At the individual level, there is wide variation within URM and within non-URM categories in the proportion of care provided to minority populations. Just as there may be URM pediatricians who do not disproportionately serve minority, publicly insured, or uninsured children, there may also be non-URM providers who do care disproportionately for those same groups.
At 3 time points over the last 14 years, URM pediatricians have been consistently more likely than their non-URM peers to provide care to minority or nonprivately insured children, and this relationship has not attenuated over the study period. The percentage of children in pediatricians' patient panels who receive public insurance has increased over the past 14 years, and comprises approximately 38% of the panels. URM pediatricians have experienced a disproportionate increase in patients who have public health insurance or who are uninsured, and these 2 groups combine to account for approximately 60% of the URM pediatricians' patient panels. These data support ongoing and enhanced efforts to recruit qualified URMs into pediatric careers to meet the growing needs of the expanding minority child population. Recent efforts to reduce or eliminate support for programs that target minority college students' interest in and preparation for medical careers may adversely affect the child physician workforce.
This study was supported by the American Academy of Pediatrics. We thank Sanford Sharp for assistance with data collection and analyses. This project was supported by grant number K08HS015679 from the Agency for Healthcare Research and Quality (Dr Basco) and grant number 57691 from the Robert Wood Johnson Foundation Generalist Physician Faculty Scholars Program (Dr Shipman). The content is solely the responsibility of the authors and does not necessarily represent the official views of the funding agencies.
- Accepted September 3, 2009.
- Address correspondence to William T. Basco Jr, MD, Medical University of South Carolina, Division of General Pediatrics,135 Rutledge Ave, PO Box 250561, Charleston, SC 29425. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
- URM =
- underrepresented minority •
- AAMC =
- Association of American Medical Colleges •
- AAP =
- American Academy of Pediatrics •
- NAMCS =
- National Ambulatory Medical Care Survey
- Friedman AL
- Saha S,
- Shipman SA
- Cooper LA,
- Powe NR
- Collins KS,
- Hughes DL,
- Doty MM,
- Ives BL,
- Edwards JN,
- Tenney K
- Garrison G
- 15.↵United States Census Bureau. U.S. Hispanic population surpasses 45 million—now 15 percent of total. Available at: www.census.gov/Press-Release/www/releases/archives/population/011910.html. Published May 1, 2008. Accessed June 29, 2009
- Fronstin P
- 17.↵Institute of Medicine.In the Nation's Compelling Interest: Ensuring Diversity in the Health-Care Workforce. Washington, DC: National Academies Press; 2004
- Reschovsky JD,
- O'Malley AS
- Terrell C,
- Beaudreau J
- Saha S,
- Taggart SH,
- Komaromy M,
- Bindman AB
- 22.↵US Department of Health and Human Services. Healthy People 2010: Understanding and Improving Health. 2nd Ed.Washington, DC: Government Printing Office; 2000. Available at: www.healthypeople.gov/Document/pdf/uih/2010uih.pdf. Accessed June 29, 2009
- Strunk BC,
- Reschovsky JD
- Slora EJ,
- Thoma KA,
- Wasserman RC,
- Pedlow SE,
- Bocian AB
- 26.↵Child Trends Databank. Racial and ethnic composition of the child population. Available at: www.childtrendsdatabank.org/indicators/60RaceandEthnicComposition.bpcfm. Updated spring 2006. Accessed June 29, 2009
Platelet-Rich Plasma Therapy: Does It Really Work?: One of the newest trend therapies for orthopedic injuries involves injecting people with their own plasma concentrated so it is mostly platelets to help heal inflamed tissues. The treatment referred to as “platelet rich plasma (PRP)” according to an article in The New York Times (Kolata G, January 12, 2010) is now being used to treat muscle sprains, tendon tears, arthritis, bone fractures and surgical wounds. Yet a well done study in the January 13 issue of the Journal of the American Medical Association (de Vos RJ, et al. JAMA, 2010; 303(2):144–149) shows that PRP therapy is no more effective than injecting saline into the injured extremity, making this therapy what Dr Bruce Reider, editor of the American Journal of Sports Medicine calls the “platelet-rich panacea” that may be anything but. Advocates of this therapy argue that there may be specific injuries where it may work better than others, but for the moment, insurers are generally not paying for these treatments, forcing those who want it to pay $1000 or more out of pocket.
Noted by JFL, MD
- Copyright © 2010 by the American Academy of Pediatrics