Availability of pediatric rheumatology training in United States pediatric residencies

Authors


Abstract

Objective

To characterize the availability of pediatric rheumatology training in general pediatric residencies.

Methods

We surveyed 195 pediatric residency program directors in the US using a combined Web-based and paper-based survey format. The survey asked directors about the availability of an on-site pediatric rheumatologist in their institution, the availability of formal pediatric rheumatology rotations, and the types of physicians involved in teaching curriculum components related to pediatric rheumatology. Survey responses were analyzed using descriptive and bivariate statistics.

Results

Of the 195 program directors surveyed, 127 (65%) responded. More than 40% of responding programs did not have a pediatric rheumatologist on site. Programs with on-site pediatric rheumatologists were significantly more likely than those without on-site pediatric rheumatologists to have an on-site pediatric rheumatology rotation available (94% versus 9%; P < 0.001). Although pediatric rheumatologists' involvement in 4 curriculum areas relevant to pediatric rheumatology is nearly universal in programs with on-site pediatric rheumatologists, nearly two-thirds of programs without on-site pediatric rheumatologists rely on internist rheumatologists, general pediatricians, or other physicians to cover these areas.

Conclusion

Programs without pediatric rheumatologists on site are less likely to have pediatric rheumatology rotations and are more likely to rely on internist rheumatologists and nonrheumatologists to address rheumatology-related curriculum components. Lack of exposure to pediatric rheumatology during residency may impede general pediatricians' ability to identify and treat children with rheumatic diseases, undermine resident interest in this field, and perpetuate low levels of supply.

INTRODUCTION

Past studies have demonstrated that the majority of pediatric subspecialists practice in academic medical centers (1), where many still function as the proverbial “3-legged stool” by providing patient care; educating physicians in training; and performing research to understand, diagnose, and treat pediatric conditions. An inadequate supply of these physicians not only limits their availability for patient care, but also negatively affects medical education. A pediatric subspecialty shortage may limit medical student and resident exposure to diseases treated by these physicians and may perpetuate discomfort and an unwillingness to care for children with complex medical conditions among general pediatricians. This lack of exposure may also perpetuate low levels of interest in a select number of fields (2).

As one of the smallest pediatric medical subspecialties, pediatric rheumatology is an excellent example of a supply-constrained field. The American Board of Pediatrics (ABP) first offered a certifying examination in pediatric rheumatology in 1992; as of December 2004, the ABP has certified 215 pediatric rheumatologists (3). Past studies have shown that the overwhelming majority of pediatric rheumatologists practice in university-based rheumatology settings (1, 4) and that nearly all pediatric rheumatologists practice in metropolitan areas (5). Based on 2004 data from the American College of Rheumatology (ACR), 13 states have no pediatric rheumatologists practicing within their borders (4). Currently, 80 medical schools in the US have 1 or more pediatric rheumatologists on faculty, leaving one-third of the nation's 125 medical schools without physicians in this subspecialty (6). This deficit raises concerns that the supply of pediatric rheumatologists may constrain the availability of pediatric rheumatology expertise in our nation's medical schools and residency programs.

A 2001 national survey of physicians' involvement in the care of children with rheumatic diseases and factors contributing to current referral patterns within pediatric rheumatology found that only 18% of pediatricians and 12% of family practitioners believed they were adequately trained to diagnose and treat juvenile rheumatoid arthritis (JRA) (7). Not surprisingly, 42% of pediatricians and 32% of family practitioners refer all JRA diagnoses and management to subspecialists. Bivariate analyses indicated that primary care physicians who reported having inadequate training in diagnosing JRA were twice as likely to refer patients as those who described their training as adequate (7).

Given the current geographic distribution of pediatric rheumatologists, it is likely that many pediatric residents have little or no exposure to pediatric rheumatologists during their years of training. To date, no studies have been conducted on pediatric rheumatology training within pediatric residency programs in the US. Using a survey of pediatric residency program directors, we characterized the availability of pediatric rheumatology training in general pediatric residencies, explored the relationship between the presence of pediatric rheumatologists at the training institution and the characteristics of pediatric rheumatology training in general pediatric residencies, and described pediatric residency directors' assessment of the adequacy of pediatric rheumatology supply both locally and statewide.

MATERIALS AND METHODS

In spring 2004, we conducted a survey of all 195 pediatric residency program directors in the US and Puerto Rico to assess the status of pediatric rheumatology training in general pediatric residency programs. One pediatric rheumatologist and one pediatric residency director pilot tested a newly developed survey, which was then revised based upon their comments. An 11-item survey resulted that included questions about program size and characteristics, pediatric rheumatology staffing, resident exposure to clinical pediatric rheumatology, recent attempts made by affiliated hospitals and/or medical centers to hire a pediatric rheumatologist, and perception of the adequacy of the pediatric rheumatology supply locally and statewide. Nine of the 11 questions asked respondents to choose from a set of closed-item responses. Open-ended questions allowed the respondent to indicate the number of years the program had been in existence and the number of trainees. The survey also included an optional open-ended section for comments. The Institutional Review Board at the University of North Carolina at Chapel Hill School of Public Health approved the survey.

Respondents were initially contacted via an e-mail message that explained the study and provided a URL link to the Web-based version of the survey. In addition to the initial e-mail, 3 followup messages were sent every 7–10 days to nonresponders; a paper version of the survey was sent to all nonresponders ∼1 month after the initial e-mail message. Responding directors were tracked using the Web site to allow reminder e-mail messages and the mail survey to be sent to nonresponders only. This tracking system also allowed the use of incentives for responding directors. In the data cleaning and analysis stage, all program director and residency program identifiers were removed so that programs remained anonymous in the analysis phase. The initial screen of the Web-based survey described the study in detail and asked respondents to formally consent to the study. The mail-based survey included a cover letter, a consent letter, and a stamped and addressed return envelope; a returned mail survey served as consent. Each respondent was sent an electronic gift code worth $10 to use at an online book store.

Programs were considered to have a pediatric rheumatologist available if program directors reported having any pediatric rheumatologists available on site at their medical center. Directors were asked if a pediatric rheumatology rotation was available as an on-site or away rotation. In addition to questions about the availability of a formal pediatric rheumatology rotation, we asked directors to indicate which faculty members were involved in 4 curriculum components relevant to rheumatology: joint examination, rheumatology laboratory evaluation, JRA diagnosis, and JRA treatment. For each curriculum component, directors could choose from 1 or more of the following: pediatric rheumatologists, internist rheumatologists, general pediatricians/continuity clinic, and lectures/guest speaker. We collapsed faculty involvement into 3 mutually exclusive categories: pediatric rheumatologist involved, internist rheumatologist involved without a pediatric rheumatologist, and general pediatrician or other nonrheumatologist. We compared these classifications between programs with and without staff pediatric rheumatologists for each of the 4 curriculum components.

Residency directors were also asked if their institution had tried to recruit a pediatric rheumatologist in the 5 years prior to the survey. Response options were 1) yes, we have successfully recruited 1 or more; 2) yes, but we have not been successful yet; 3) no, we have a sufficient number of rheumatologists; 4) no, we would like to but are unable to do so for financial or other reasons; and 5) I do not know. Finally, residency directors were asked to describe the supply of pediatric rheumatologists in their catchment area as inadequate for the current level of patient demand; adequate for patient demand but insufficient to allow time for research and teaching; or adequate to fulfill clinical, research, and teaching responsibilities. Respondents were also asked to describe statewide supply of pediatric rheumatologists using the same categories.

Bivariate analyses were used to compare programs with and without pediatric rheumatologists with regard to program characteristics, the availability of a formal pediatric rheumatology rotation, and the faculty involved in teaching rheumatology curriculum components. Pearson's chi-square analyses were used to assess associations between categorical variables. In comparisons where any cell sizes were <5, these associations were tested using Fisher's exact test. Associations between categorical and continuous variables were tested using 2-sided, 2-sample t-tests or Kruskal-Wallis tests. All analyses were performed using Stata software, release 8.0 (8).

RESULTS

Of the 195 program directors surveyed, 127 (65%) responded: 84 responded to the Web-based survey and 43 responded to the mailed survey. The total number of pediatric residents per program ranged from 9 to 123, and the average was 41 (Table 1). In addition, 59.6% of programs also had internal medicine/pediatric programs, averaging 16 of these residents per program. Likewise, 15.6% of programs had combined pediatric residencies, such as pediatrics/physical medicine and rehabilitation, averaging 5 of these residents per program. Approximately one-quarter of programs reported that ≤50% of their graduates pursued primary care careers.

Table 1. Characteristics of responding pediatric residency programs by availability of a pediatric rheumatologist*
CharacteristicAll programs (n = 127)Pediatric rheumatologist
None (n = 55)On site (n = 72)
  • *

    Values are the percentage unless otherwise indicated. Bivariate analyses performed using Pearson's chi-square test for categorical variables and Student's t-test for continuous variables unless otherwise indicated. IM = internal medicine.

  • Difference between programs with pediatric rheumatologists and those without these physicians significant at P < 0.05.

  • Difference between programs with pediatric rheumatologists and those without these physicians significant at P < 0.001.

  • §

    Among those programs with any IM/pediatric residents. Difference between programs with pediatric rheumatologists and those without these physicians significant at P < 0.05.

  • Bivariate comparison performed using Fisher's exact test.

  • #

    Among those programs with any combined pediatric residents.

  • **

    Bivariate comparison performed using Fisher's exact test. Difference between programs with pediatric rheumatologists and those without these physicians significant at P < 0.001.

Pediatric rheumatologist on staff at affiliated institution   
 Yes56.7  
 No43.3  
Bivariate results   
 Years program in existence, mean ± SD32.2 ± 17.128.2 ± 15.435.1 ± 17.9
 General pediatric residents, mean ± SD41.1 ± 23.829.9 ± 16.349.6 ± 25.1
 Any IM/pediatric residents59.659.559.7
 IM/pediatric residents, mean ± SD§15.5 ± 10.112.2 ± 9.217.6 ± 10.2
 Any combined pediatric residents15.69.519.4
 Other combined pediatric residents, mean ± SD#4.9 ± 5.92.5 ± 1.35.7 ± 6.5
 Percentage of graduates in primary care   
  ≤25%0.00.00.0
  26–50%26.010.937.5
  51–75%58.370.948.6
  >75%15.818.213.9
 Program type   
  University based43.321.878.2
  Community based, university affiliated34.761.738.6
  Other22.157.142.9
 Programs with a pediatric rheumatology rotation**   
  On-site rotation available57.59.194.4
  Away rotation available22.045.44.2
  No rotation available20.545.41.4

We asked respondents to indicate the number of pediatric rheumatologists with patient care responsibilities at their institution. More than 40% of responding programs lacked an on-site pediatric rheumatologist. Seventy (56.7%) of the responding programs had ≥1 pediatric rheumatologists on staff, including 2 institutions that reported sharing a physician with another institution. The number of pediatric rheumatologists in these programs ranged from 1 to 6, with a mean of 1.8. Programs with pediatric rheumatologists on staff were significantly larger, in terms of pediatric residents, than those without these physicians and had significantly fewer residents pursuing general pediatrics careers. Programs with pediatric rheumatologists were significantly more likely to be university based.

Residency directors were also asked if their program offered a pediatric rheumatology rotation, either on site or as an away elective; overall, 57.5% of responding programs offered an on-site rotation and an additional 22.0% offered a rotation as an away elective. More than 90% of residency programs with a pediatric rheumatologist on staff reported offering an on-site pediatric rheumatology rotation. Among those programs without a pediatric rheumatologist on staff, 45% offered the rotation as an away elective and another 45% of programs did not offer any pediatric rheumatology rotation. Five programs (9%) without a pediatric rheumatologist available on site did offer an on-site rotation. One of these programs had an allergist precepting the rotation; however, it is not clear who precepted the rotation at the other 4 sites.

Programs that lacked pediatric rheumatology rotations were significantly more likely to have a greater percentage of graduates in primary care positions (Table 2). Directors from 64% of programs with on-site rheumatology training estimated that >50% of their graduates practiced in primary care; more than 90% of directors in programs without a pediatric rheumatology rotation available estimated that >50% of their graduates practiced in primary care.

Table 2. Characteristics of pediatric residency training programs: percentage of programs by availability of a pediatric rheumatology rotation
 Overall (n = 127)Rotation availability
On site (n = 73)Away elective (n = 28)None (n = 26)
  • *

    Difference between programs significant at P < 0.01.

  • Difference between programs significant at P < 0.001.

Percentage of graduates in primary care*    
 26–50%26.035.617.97.7
 51–75%58.352.153.680.8
 ≥75%15.812.328.611.5
Program type    
 University based43.380.05.514.6
 Community based, university affiliated34.740.930.029.6
 Other22.139.342.917.9
Percentage of general pediatric residents participating in a pediatric rheumatology rotation    
 ≤25%70.645.3100.0100.0
 26–50%15.126.0  
 51–75%7.112.3  
 ≥76%7.112.3  

Despite the availability of formal pediatric rheumatology rotations in 79% of pediatric residency programs, few pediatric residents elected to participate in these rotations. Compared with programs with away electives, programs with on-site rotations reported having a significantly higher percentage of residents completing a rheumatology rotation. Nonetheless, nearly half of directors from programs with on-site rotations estimated that <25% of their residents complete a pediatric rheumatology rotation during their residency. Among those programs with an away elective, all directors estimated that ≤25% of their residents complete this rotation. Interestingly, 7 programs that reported not having a pediatric rheumatology rotation reported that <10% of their residents had completed such a rotation (i.e., rather than reporting “none”). For 3 of these programs, a rotation with an adult rheumatologist was available or a pediatric rheumatologist had been on staff previously; for the remaining programs, the format of the rotation was unclear.

When one examines faculty involvement in the pediatric rheumatology curriculum components by availability of a staff pediatric rheumatologist, the importance of having a pediatric rheumatologist on staff becomes readily apparent. Programs without pediatric rheumatologists at their institutions were significantly more likely to rely on internist rheumatologists and/or general pediatricians to address these curriculum areas (Table 3). For each curriculum component, nearly 100% of the program directors with a pediatric rheumatologist on staff at their affiliated institution reported the pediatric rheumatologist's involvement in these training areas. In contrast, more than two-thirds of program directors without pediatric rheumatologists at their affiliated institutions reported that training in these areas is the domain of internist rheumatologists, general pediatricians, continuity clinics, and lectures and/or nonrheumatologists. It is important to note that approximately one-third of programs without pediatric rheumatologists on staff were able to involve pediatric rheumatologists in their resident training nonetheless.

Table 3. Faculty involvement in pediatric rheumatology curriculum components: percentage of programs by availability of an on-site pediatric rheumatologist*
 Joint examinationLaboratory workJRA diagnosisJRA treatment
No PRPRNo PRPRNo PRPRNo PRPR
  • *

    For each curriculum component and faculty classification, the difference between programs with and without staff pediatric rheumatologists (PR) is significant at P < 0.001 using Fisher's exact test. JRA = juvenile rheumatoid arthritis.

PR only or in combination with other physicians, continuity clinic, and/or lectures24.595.830.898.632.197.235.998.6
Internist rheumatologist only or in combination with other physicians (except PRs), continuity clinic, and/or lectures20.80.032.71.430.21.434.01.4
General pediatricians or continuity clinic with lectures and/or other nonrheumatologist54.74.236.50.037.71.430.20.0

Residency directors were asked, to the best of their knowledge, if efforts had been made to recruit ≥1 pediatric rheumatologists to their institutions in the previous 5 years. Nearly 25% reported that their institution had successfully recruited ≥1 pediatric rheumatologists and an additional 11.2% had been unsuccessful in their recruitment efforts. More than 33% of program directors reported an interest in recruiting a pediatric rheumatologist but an inability to recruit due to financial or other reasons. Only 13% of the program directors believed they did not need such a physician and 16% did not know about their institution's interest in hiring a pediatric rheumatologist. University-based programs were significantly more likely than other programs to have had a successful recruitment effort, whereas community-based, university-affiliated, and other programs were more likely to have been unable to recruit despite a need for a pediatric rheumatologist (Table 4). Nearly 90% of program directors without an on-site pediatric rheumatologist at the time of the survey reported an interest in hiring a pediatric rheumatologist but an inability to do so for financial or other reasons. The average number of pediatric rheumatologists was highest in programs with successful recruitments (mean 2.35), followed by those reporting no need for additional rheumatologists (mean 1.18).

Table 4. Factors associated with successful recruitment of a pediatric rheumatologist in the last 5 years*
 Recruitment effortNo recruitment effortsDo not know (n = 20)
Successful (n = 31)Unsuccessful (n = 14)No need (n = 17)Identified need (n = 43)
  • *

    Values are the percentage unless otherwise indicated. Bivariate comparisons performed using Fisher's exact test for program type and any pediatric rheumatologist and Kruskal-Wallis test for number of pediatric rheumatologists.

  • Significant at P < 0.05.

  • Significant at P < 0.001.

Program type     
 University based35.214.811.118.520.4
 Community based, university affiliated23.37.09.348.811.6
 Other7.110.725.042.814.3
Any pediatric rheumatologist     
 Yes42.38.521.17.021.1
 No1.914.83.770.49.3
Mean ± SD pediatric rheumatologists on site2.35 ± 1.280.44 ± 0.641.18 ± 0.640.14 ± 0.411.20 ± 1.23

Finally, residency directors were asked to assess the adequacy of pediatric rheumatologist supply in their catchment area and statewide. The majority believed that either the supply was inadequate (41.7%) or the supply was adequate to allow patient care but inadequate to allow time for research and teaching responsibilities (26.0%). Significantly more directors in institutions that lacked pediatric rheumatologists believed that supply was inadequate compared with institutions with pediatric rheumatologists (65.0% versus 23.6%; P = 0.001). The majority of directors similarly believed that the statewide supply of these physicians was inadequate (48.8%) or adequate for patient care only (14.2%). Directors of programs without pediatric rheumatologists were significantly more likely than directors of programs with pediatric rheumatologists to describe the statewide supply as inadequate (61.8% versus 38.9%; P < 0.01); however, only 12.5% of directors of programs with a staff pediatric rheumatologist described the statewide supply as adequate for patient care as well as other responsibilities.

DISCUSSION

Pediatric rheumatology is one of the smallest pediatric subspecialties in the US. Past studies have demonstrated that current distribution of these physicians leaves many geographic areas and medical schools lacking physicians with pediatric rheumatology expertise. Our study is the first to quantify the availability of pediatric rheumatologists in pediatric residency programs and to explore the relationship between their availability and the characteristics of pediatric rheumatology training.

Our results suggest that >40% of pediatric residency programs lacked an on-site pediatric rheumatologist. On average, pediatric rheumatologists were located at larger, university-based training programs. The availability of an on-site pediatric rheumatologist was positively associated with the availability of an on-site pediatric rheumatology rotation. Even in programs offering formal rotations, however, the percentage of pediatric residents electing to participate in these rotations was rather low. In half of the programs with formal, on-site pediatric rheumatology rotations, directors estimated that <25% of residents participate in a pediatric rheumatology elective. Therefore, the availability of a pediatric rheumatology rotation alone is not sufficient to encourage residents to spend dedicated time learning about the diagnosis and management of these diseases. The reasons for this are unclear but may relate to heavy clinical loads for the pediatric rheumatologists or a disinterest in the curricular elements that are usually the responsibility of pediatric rheumatologists.

The availability of an on-site pediatric rheumatologist is positively associated with the involvement of a pediatric rheumatologist in curriculum components related to rheumatology. In programs with on-site pediatric rheumatologists, these rheumatologists' involvement in teaching rheumatology-related curriculum components is nearly universal. Among those programs without on-site pediatric rheumatologists, however, nearly one-third rely on internist rheumatologists to provide instruction in rheumatology laboratory evaluation, JRA diagnosis, and JRA treatment. Likewise, one-third of programs without a pediatric rheumatologist on site rely on a general pediatrician or other nonrheumatologist to provide this training. In institutions without pediatric rheumatologists, the diversity of the reported specialists providing important curricular activities such as the musculoskeletal examination suggests that there is no systematic or established curriculum for this area. The extent to which nonpediatric rheumatologists are qualified to provide this training is not known.

It is notable that one-third of programs without a pediatric rheumatologist available still manage to involve these physicians in pediatric residency training, albeit in a limited manner. The ACR facilitates the involvement of pediatric rheumatologists in the training of pediatric residents at institutions without an on-site physician through the Pediatric Rheumatologist Visiting Professor Program (9). The endowment funds ∼4 visitation professorships per year at pediatric residencies accredited by the Accreditation Council for Graduate Medical Education that are affiliated with medical schools lacking pediatric rheumatology programs. Two responding program directors cited this program unprompted. One noted, “Pediatric rheumatologists are like gold. This year I received a visiting professorship from the ACR which hopefully will be helpful….” Another echoed these sentiments, “We desperately need a ped[iatric] rheumatologist and have now for the second year in a row secured a visiting prof[essor] … a finger in the dike both for our p[atien]ts' care and our housestaff education.” The American Academy of Pediatrics Section on Rheumatology also sponsors one annual visitation professorship each year (10). Expansion of these programs may facilitate pediatric residents' exposure to pediatric rheumatology and address the effects of the shortage on resident education in the short term.

Nearly 90% of program directors without an on-site pediatric rheumatologist reported that they would like to hire one but are unable to do so for financial or other reasons. Open-ended comments from these residency directors suggest that financial factors heavily influence their program's ability to hire a pediatric rheumatologist: “… our rheumatologist functions as a generalist and teacher in the clinic, newborn nursery, and on the pediatric floor … He spends the majority of his time in these endeavors, not in rheumatology … We would not be able to support a full time rheumatologist …”; “The major barriers to bringing pediatric rheumatology to our center are lack of available ped[iatric] rheumatologists to recruit, funding based on clinically generated dollars …, convincing [institution name deleted] re: the financial viability of the position, finding ancillary dollars (education, research, etc.) in a community-based academic residency, providing cross-coverage for on-call, etc.”; “We probably do not have sufficient patients within our tri-county referral area to justify a full-time on-site ped[iatric]s rheumatologist, nor do we have anything close to the budget …”; “To get an on-site specialist we would have to show that it is ‘cost-effective’ to hire them, and with our population this would not be the case.”

As with most physician surveys, our response rate is an important limitation. Although a response rate of 65% exceeds typical response rates for physician surveys (11), the possibility of response bias remains. To assess the possibility of response bias, we compared responding and nonresponding programs on 3 characteristics for which we had available information: whether the program was located in a metropolitan statistical area, distance to the nearest pediatric rheumatologist, and whether the program was university based (4). There was no significant difference between responders and nonresponders with regard to these characteristics. Nonetheless, nonresponding programs may differ from responding programs in other ways, such as program size.

Discussions of physician workforce issues rarely address pediatric subspecialties separately. As a group, pediatric subspecialties remain largely academic-based specialties. As such, those working in these fields treat patients, educate physicians in training, and perform research. Evaluation of the adequacy of pediatric subspecialty supply, therefore, must consider the effects of supply constraints on education and research.

Our results demonstrate that programs without pediatric rheumatologists are significantly less likely to have formal pediatric rheumatology rotations available and are more likely to rely on internist rheumatologists and nonrheumatologists to address rheumatology-related curriculum components. Most programs that lack an on-site rheumatologist have a need for such a provider but an inability to hire one. Lack of exposure to pediatric rheumatology during residency may impede resident interest in this field and perpetuate low levels of supply. One director noted, “I haven't been successful in trying to convince residents in the last 6 years to enter this field, I am hoping this will change now [that the institution has a pediatric rheumatologist].” Furthermore, lack of exposure to pediatric rheumatology potentially renders general pediatricians incapable of and uncomfortable with evaluating children with rheumatic diseases, as past studies have suggested a relationship between adequacy of pediatric rheumatology training and involvement in the care of children with rheumatic diseases (7). Efforts to increase the availability of pediatric rheumatology expertise in pediatric residency programs may serve not only to increase interest in this field but also to prepare general pediatricians to participate in the care of children with rheumatic diseases within their home communities.

Ancillary