Chapter 4. Substitutes for Pediatric Rheumatologists?  General Pediatricians and Internist Rheumatologists Involvement in Pediatric Rheumatology Care as Evidence of a Shortage

The approximately 200 practicing pediatric rheumatologists in the United States practice in a limited number of geographic areas.  As a consequence, other physicians---internist rheumatologists and primary care providers---may substitute for pediatric rheumatologists in some regions.  Several studies suggest, for example, that internist rheumatologists play a prominent role in the care of children with these diseases. ^{1, 4, 21}  A 2000 report revealed that one-third of children with a known rheumatic disease received their care from an internist rheumatologist. ²⁷  A study of Medicaid children eligible for Supplemental Security Income (SSI) found that only 18 percent of children with juvenile rheumatoid arthritis (JRA) saw a pediatric subspecialist and 30 percent saw an internist subspecialist. ¹

Receipt of Care by North Carolina Medicaid Children with Juvenile Rheumatoid Arthritis

North Carolina Medicaid claims data provide an opportunity to characterize physician utilization among North Carolina Medicaid enrollees with JRA.  North Carolina ranks 33^rd in the Nation in the ratio of children to pediatric rheumatologists.  It is also a predominantly rural State with pediatric rheumatologists (n=4) located only in Chapel Hill, Durham, and Greenville.  Pediatric rheumatology care is available also in bordering areas such as Knoxville, TN, Richmond, VA, and Norfolk, VA.

Children who had two or more physician visits with one of the following ICD-9 CM diagnosis codes in a given year were considered to have JRA:  714.3, 714.30, 714.31, 714.32, and 714.33.  Analyses were limited to children who were continuously enrolled for one or both of the following time periods:  July 1, 1998 through June 30, 1999 or July 1, 1999 through June 30, 2000. 

Physician specialty codes were used to classify JRA-related physician visits ^[vii] by provider specialty.  Unfortunately, North Carolina Medicaid claims data from hospital-based physician practices usually do not identify the specific physician for whom the claim was generated.  For example, claims from the Private Diagnostic Clinic at Duke University Medical Center and those from University of North Carolina (UNC) Physicians and Associates come from the overall physician practice without an individual provider identifier; as a result, one cannot definitively identify the treating physician.  JRA claims from these types of physician practices were classified as visits to pediatric rheumatologists when, according to the American College of Rheumatology, the center had a pediatric rheumatologist available and the primary diagnosis was JRA.

Internist rheumatologists were identified in two ways:  using the specialty code specific to rheumatology and comparing the list of multispecialty and “other” providers with the American College of Rheumatology file to identify misclassified rheumatologists.  Internist rheumatologists may be underrepresented because care received at physician practices based in medical centers, like North Carolina Baptist Hospital or Carolinas Medical Center, do not list the specific physician seen. Thus, the “multispecialty” and “other” categories may also include care provided by internist rheumatologists.

To avoid including children with a one-time “rule-out” diagnosis, the analyses included only children with two or more physician claims with a diagnosis of JRA.  The analysis included only continuously enrolled children to ensure that we had all physician claims for each study child.  In 1999 and 2000, respectively, 67 and 68 continuously enrolled Medicaid children had two or more physician visits with a diagnosis of JRA (Table 13).  Of these, slightly more than 50 percent had at least one visit to a center with a pediatric rheumatologist in each year. ^[viii]   In both years only 40 percent of all JRA-related physician visits were to a pediatric rheumatologist or a medical center with a pediatric rheumatologist.  General pediatricians and internist rheumatologists provided 15 to 20 percent of all JRA-related visits to this population. 

Table 13:  Number of Medicaid-Enrolled Children with JRA in North Carolina and Distribution of JRA-related Physicians Visits

North Carolina Medicaid data were also used to characterize the JRA-related physician management of children.  Using only visits with a JRA diagnosis, children were classified into groups based on the involvement in their care of a primary care physician (PCP), pediatric rheumatologist and/or an internist rheumatologist.  Children seen by a family practice physician or a pediatrician, but not seen by any rheumatologists, were classified as “PCP only;” those with JRA-related claims from pediatric rheumatologists only or internist rheumatologists only were classified accordingly.  Those with JRA-related claims from both a PCP and a pediatric rheumatologist were classified as having their care co-managed by these providers and those seen by both a PCP and an internist rheumatologist were similarly classified.

Physician management of JRA care was fairly evenly distributed across types in 1999, with “PCP Only,” “PCP with a Pediatric Rheumatologist,” and “Internist Rheumatologist Only” each accounting for around 20 percent of visits (Table 14).  “Pediatric Rheumatologist Only” was the most common management type, accounting for over 30 percent of children with JRA.  In FY 2000 the percentage of children treated exclusively by a PCP declined nearly 6 percentage points and the percent being co-managed increased by nearly the same amount.  While these figures suggest that more than one-half of Medicaid enrolled children with JRA in North Carolina have been seen at a center with a pediatric rheumatologist on staff, one cannot ascertain that the visits to these centers involved a visit to a pediatric rheumatologist.  Nonetheless, it is clear that primary care providers and internist rheumatologists play an important role in the care of Medicaid-enrolled children with JRA.

Table 14:  Classification of JRA-related Physician Care among Medicaid-Enrolled Children with JRA, North Carolina

Internist Rheumatologists as Providers of Pediatric Rheumatology Care

A 2002 study of physician members of the ACR (n=4,673) divided these specialists into three groups: those who treat pediatric patients only, internist rheumatologists who treat adults only, and internist rheumatologists who treat both adults and children.  After restricting the sample to physicians who provide at least some patient care (n=4,304), 224 were classified as pediatric only providers, 3,030 as internist rheumatologist who treated adults only and 1,050 as internist rheumatologists who also treated children.  Of the 3,141 counties in the United States, 623 (20 percent) have an internist or pediatric rheumatologist involved in patient care on at least a part-time basis (Figure 4).

[D]

As Table 15 shows, the authors found that approximately 50 percent of the population under age 18 of the United States lived within 50 miles of a pediatric rheumatologist.  When the parameters were expanded to include internist rheumatologists who treat adults and children, the percentage of children living within 50 miles of a provider of pediatric rheumatology care (i.e., an internist who treats children or a pediatric rheumatologist) increased to 90 percent.  This indicates that internist rheumatologists are more geographically diffuse than pediatric rheumatologists and, therefore, their involvement in the treatment of children substantially reduces the distances that must be traveled to obtain care.

Table 15:  Percent of the Population Living within Selected Distances of Rheumatology Providers by Rheumatology Provider Type

Source:  Mayer ML, Mellins ED, Sandborg CI. Access to pediatric rheumatology care in the United States. Arthritis Rheum 2003; 49:759-65 

A survey of California rheumatologists found that children under the age of 18 represented a small percentage of internist rheumatologists’ patients. ²²  Respondents were asked how many patients they treat by age group (under 18 years, and 18 years and older); in addition a separate question asked if they “treat pediatric rheumatology patients.”  Among internist rheumatologists who reported treating pediatric patients, approximately 3 percent of their patients are under the age of 18.  While the number of pediatric patients seen by individual internist rheumatologists is small, collectively they treat a large number of patients under 18.  Using self-reported data on patient volume, this survey estimated that, as a group, internist rheumatologists who report treating children saw a total of 217 under 18 patients per week.  In addition, many internist rheumatologists who reported not treating pediatric patients did, in fact, report treating patients under the age of 18.  These providers, as a group, treat approximately 202 patients under the age of 18 in a week.  Collectively, internist rheumatologists were seeing almost as many patients as were pediatric rheumatologists who treated an estimated 550 patients under 18 per week. 

Factors Influencing Internist Rheumatologists Involvement in the Care of Children 

Distance to the nearest pediatric rheumatologist appears to be an important determinant of internist rheumatologists’ involvement in the care of children.  The aforementioned 2002 national study of ACR members found that, controlling for a variety of other factors such as practice type and non-clinical professional activities, ⁶ distance to pediatric rheumatology care was significantly related to internist rheumatologists’ involvement in the care of children. Moreover, a study found that internist-rheumatologists in Washington State found those who reported treating children lived a significantly greater average distance from a pediatric rheumatology referral center than those who referred children (101 miles vs. 21 miles) and distance was the most frequently reported reason (66 percent) for not referring a child to a pediatric rheumatologist. ⁴  A survey of California rheumatologists found that internist rheumatologists practicing between 10 and 50 miles from the closest pediatric rheumatologist were significantly more likely than those within 10 miles of a pediatric rheumatologist to treat children.  The odds of treating pediatric patients among those practicing 50 or more miles from the nearest pediatric rheumatologist were nearly 7 times higher than among those practicing within 10 miles of a pediatric rheumatologist. ²²    

The AF/ACR Survey conducted in 2004 found that internist rheumatologists involved in the care of children practiced a significantly greater distance from a pediatric rheumatologist, on average, than those who do not treat children (66.2 vs. 46.1 miles, p=0.017).  In multivariable analyses that controlled for provider characteristics, such as age and distribution of work hours, distance did not maintain its significance.  These results should be interpreted with caution, however, as it is estimated that the response rate among internist rheumatologists in this survey was less than 20 percent.

Among California internist rheumatologists who treat children, 79 percent indicated that personal expertise in pediatric rheumatology motivated them to treat children.  Patient preferences and distance to the nearest pediatric rheumatologist also were selected as motivating factors by 73.3 percent and 65.8 percent, respectively, of internist rheumatologists who treated children.  Among internist rheumatologists who do not treat children, the overwhelming majority (85.2 percent) refrain from seeing children because of inadequate personal expertise in pediatric rheumatology.  Most internist rheumatologists (70.3 percent) also cited the availability of nearby pediatric rheumatology care as a reason for their decision not to treat children with rheumatic diseases.  Among Washington State rheumatologists, the most common factor in an internist rheumatologist’s decision not to refer a child to a pediatric rheumatologist was distance (66 percent), with the second most common reason cited as inconvenience to the child’s family (60 percent). 

Data from the California survey also suggested that practice setting may play a role in internist rheumatologists’ involvement in the care of children; those practicing in multispecialty clinics were significantly more likely to be involved in the care of pediatric patients than those who practiced in other settings.  Interestingly, when queried about their reasons for treating pediatric patients, 46.2 percent of those in multispecialty practices cited insurance barriers to referral versus only 28.8 percent of those in other practice settings.  It is possible that physicians in these types of groups may be more involved in independent practice associations (IPA).  IPAs may be less likely to include pediatric subspecialists and providers may face disincentives to referring outside their network of providers, ^{28, 29}   which might explain the somewhat increased tendency, among those in multispecialty settings, to report insurance barriers as a motivating reason for treating children.

Comments from Internist Rheumatologists

While quantitative analyses are useful in summarizing survey data, qualitative data can provide rich and varied insights into the context of physician practice decisions.  A sample of comments from California rheumatologists surveyed suggests that some internist rheumatologists treat children only because of the lack of available pediatric providers and appear to be uncomfortable with their involvement in the care of these patients.  In contrast, other internist rheumatologists feel they are fully capable of meeting the needs of this population and are not called upon to do so often enough.   The following comments are from this survey.

Distance/Access

• “There is a major shortage of pediatric rheumatologists.  I treat some pediatric populations out of inaccessibility to pediatric rheumatologists – I would like to not treat any or get more training.”
• "I love kids, and feel that we desperately need more pediatric rheumatologists but with no pediatric experience in residency, only 12 clinics in fellowship, I am unprepared to see kids.  I could easily handle mild JRA cases with MTX/NSAIDs [methotrexate/non-steroidal anti-inflammatory agents], pain injections, but the biggest hold back is my comfort with diseases of children.  Legally and medically, I would be asking for trouble.”
• “Straight forward pediatric rheum [atology] or older pediatric patients I am comfortable taking care of.  If there is an element of doubt, I refer to Children Hosp LA.”
• “Because of the shortage of pediatric rheumatologists, I do not mind seeing old pediatric patients (> 13 yrs) if I have to.  But I definitely prefer not to treat anyone under 13 years of age because my training in internal medicine did not prepare me to treat pediatric patients.”
• “I rarely receive referrals from ‘peds’ [general pediatricians] for questionable reasons.  ‘Peds’ may feel we are incapable of treating children with rheumatic disorders even with training/experience.”
• “The major obstacle to the treatment of pediatric rheumatology cases in our area is the hesitance by the local pediatricians.  They all seem to shy away from the care of really sick children and prefer not to be involved in their care.  Thus, they almost always refer them out of the area, causing great inconvenience and, often, suboptimal care for the patients.  I would strongly suggest that pediatricians be made aware that most rheumatologists are quite experienced and able to care for pediatric rheumatology patients.”
• “Pediatricians were never taught to do joint exams.  Therefore, they don't recognize a swollen joint when they see it, so the child gets referred to an orthopedist, who does the "only" sensible thing:  they "cast" it.  Weeks later the cast comes off, oops, now we have a contracture.  So they refer to the university where the child disappears - where the child never gets referred back to me (unless HMO insists).  I like kids and could do your follow-ups locally, save the ‘ped rheum’ time (and the pts/family time for appointments).  But you never ask and patients are never even referred back to me.”

Practice Constraints

• “Because my office is not set up for children, I only deal with teens who like being treated like an adult.... I am more at ease with prescriptions.”
• “Too busy with adult patients to see pediatric patients.  Not really interested in seeing pediatric patients.”
• “In Fresno…there are two excellent pediatric rheum [atology] MDs but there is a shortage of adult rheum [atologist]s.”
• “I choose not to tx [treat] pediatric patients as I never did a pediatric rotation in my training.  Even if I had done one month of ped [iatric] rheum [atology] training, I probably would not treat this population due to professional liability concerns (I wouldn't be Board certified in pediatric rheumatology).  Currently, I have a five month waiting period to see adult patients - I have little incentive to see pediatric patients as well.”

Additional comments, not presented here, underscore the finding of the quantitative analysis that internist rheumatologists generally restrict their involvement in pediatric rheumatology to the care of adolescents and suggest that some of these providers also limit their involvement to children with mild forms of the more common rheumatic diseases.   

It is important to note that while many internist rheumatologists currently treat a small number of children and adolescents with rheumatic disease, their role in caring for the under-18 population may decrease in coming years.  The aging of the “baby-boomer” generation will increase adult demand for the services of internist rheumatologists and limit their availability to care for children.

Role of Primary Care Physicians in Treating Children with Rheumatic Diseases

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 11 percent of pediatricians and 38 percent of family practitioners had not seen any suspected or confirmed cases in the 5 years prior to the survey. ³⁰ Only 3 percent and 1 percent of pediatricians and family practitioners, respectively, saw more than 10 cases; only one percent of respondents diagnose and treat patients with JRA on their own. ³⁰  Forty-two percent of pediatricians and 32 percent of family practitioners refer all JRA diagnosis and management to subspecialists.  Most of the respondents indicated that they refer patients to a pediatric rheumatologist (92 percent of pediatricians and 76 percent of family physicians); a substantial percent of family physicians indicated, however, they refer to general rheumatologists (37 percent).  This may be because many are located in rural areas where access to a pediatric rheumatologist may be limited. ³⁰  

In analyzing survey responses from general pediatricians and family physicians, Freed et al. ³⁰ found that 42 percent of pediatricians and 19 percent of family practitioners felt comfortable diagnosing JRA, but only 18 percent of pediatricians and 12 percent of family practitioners felt they were adequately trained to diagnose and treat JRA.  Only 10 percent and 4 percent of pediatricians and family practitioners, respectively, described themselves as current on the latest JRA treatments.  Bivariate analyses indicated that PCPs who reported having inadequate training in diagnosing JRA were twice as likely to refer patients as those who described their training as adequate, implying that improvements in training may enhance primary care providers willingness to be involved in the care of children with rheumatic diseases. ³⁰  

Reported Pediatric Referral Sources

The 2003 survey of California internist rheumatologists asked providers to indicate which groups had referred children with rheumatic diseases to them.  Compared with internist rheumatologists who did not report treating children, internist rheumatologists who did report treating children were significantly more likely to be contacted by all referral sources (Figure 5). ²²   It is not clear if internist rheumatologists decide to treat children because they are asked to, or if these providers have a reputation for treating children and, as a consequence, are more frequently approached for treatment.  It is also notable that internist rheumatologists who do not treat children frequently reported being contacted about seeing a pediatric patient with a known or suspected rheumatic condition despite their unwillingness to treat children.

In their National survey, Freed and colleagues found that internist rheumatologists reported that their primary referral sources of JRA patients were family physicians, followed by pediatricians, then orthopedists.  Seventeen percent of internist rheumatologists reported that they never refer JRA patients to another specialist, and 11 percent reported referring all of their juvenile patients.  Ninety-six percent of all referrals from internist rheumatologists were to a pediatric rheumatologist. ³⁰

Figure 5:  Percentage of Internist Rheumatologists Reporting Requests to See Pediatric Patients by Requesting Source

[D]

Comfort in Treating Children

Freed and colleagues found that 88 percent of internist rheumatologists reported they are “adequately trained to diagnose JRA” and 72 percent felt they were “adequately trained to manage JRA.” ³⁰  Internist rheumatologists reported, nonetheless, that they often referred these patients to pediatric rheumatologists.  Factors that were considered most important in the referral decision were age of patient, parental request, and refractory clinical course.³¹  In a focus group composed of internist rheumatologists, several indicated that they would begin treating a patient with JRA and only refer them to a specialist if there was no improvement.  Most agreed that proximity to such a specialist also played a large role in the decision to refer.

There is evidence that internist rheumatologists limit their involvement in the care of children to adolescents.  California internist rheumatologists were significantly less likely to treat children ages 0-5 and 6-11 than pediatric rheumatologists. ²²  Patients ages 16-17, on average, represent over 50 percent of internist rheumatologists’ pediatric patients.  The diseases treated did not differ significantly, however, between these providers; for both pediatric and internist rheumatologists the majority of patients had JRA or SLE.

Compared to pediatric rheumatologists significantly fewer California internist rheumatologists were comfortable treating each of 18 listed conditions. ²²  The majority of internist rheumatologists were comfortable treating the JRA subtypes, SLE, dermatomyositis, and spondylarthropathy; however, less than half of those responding were comfortable treating Kawasaki’s disease, Wegener’s granulomatosis, polyarteritis nodosa, reflex sympathetic dystrophy, psychogenic rheumatism, and fever of unknown origin.  Most Washington State internist rheumatologists similarly reported comfort with treating children with common diseases such as JRA; however, the percentage of surveyed physicians who reported comfort treating rarer diseases that typically affect younger children was much lower. ⁴

Analyses of the recent AF/ACR survey by the Arthritis Foundation and the ACR had similar findings.  Internist rheumatologists who treat children were less likely than those who do not treat children to report that there is an age below which they are uncomfortable diagnosing or treating pediatric patients.  Even among those who treat children the majority do report being uncomfortable diagnosing (78.3 percent) and treating (80.3 percent) children below a certain age.  The average age below which they are uncomfortable diagnosing is lower for those who treat children than those who do not (9.5 years vs. 14.9 years, respectively, p<0.001).  Likewise, the minimum average age at which a provider feels comfortable treating a child with a rheumatic disease is lower for those internist rheumatologists who treat children than among those who do not (9.8 years vs. 15.3 years, respectively, p<0.001).  As expected, internist rheumatologists who care for children are significantly more likely than those who do not treat children to report being comfortable treating pediatric rheumatic diseases.  As shown in Table 16, internist rheumatologists who treat children are most uncomfortable treating Kawasaki’s disease (67.1 percent), periodic fevers (54.3 percent), somatiform disorders (46.9 percent), and pediatric vasculitis (42.1 percent).  Those who do not treat children are significantly more likely than those who treat children to report being uncomfortable treating all reported illnesses.

Table 16.  Comparison of Internist Rheumatologists By Involvement in Care of Children AF/ACR Survey (N=523)

Practice Guidelines

Freed and colleagues also explored the need for continuing education for general practice physicians and internist rheumatologists who are likely to treat children with rheumatic diseases.  In a survey 71 percent of pediatricians, 73 percent of family physicians, and 73 percent of the internist rheumatologists indicated that they saw a need for a JRA practice guideline to be disseminated to physicians within their specialty; ³⁰ among focus groups, however, the results were less consistent.  Most pediatricians agreed that practice guidelines were not necessary because they usually refer potential JRA patients to specialists; they indicated that even if guidelines were available they would most likely not use them.  Opinions from the family physicians were mixed; some were skeptical of the benefit of guidelines because they see potential JRA cases so infrequently.  Those family physicians who were likely to play a significant role in the management of JRA care thought that guidelines would be especially relevant and helpful.  Of the internist rheumatologists who participated in a focus group, most agreed that a guideline would be helpful to keep them updated on recent developments in the treatment of JRA patients. 

Summary

Internist rheumatologists play a prominent role in the care of children with rheumatic diseases; evidence suggests that the lack of available pediatric rheumatologists influences the involvement of internist rheumatologists in the care of children.  Nonetheless, studies suggest that they may limit their involvement to the care of adolescents and those with mild cases.  The practice locations of internist rheumatologists are more geographically diffuse than those of pediatric rheumatologists and, as such, their involvement in the care of children with rheumatic disease certainly decreases the distances that children need to travel for care.  Efforts to enhance the ability of internist rheumatologists to provide quality care to children with rheumatic diseases may help ameliorate the current shortage. 

In contrast to internist rheumatologists, primary care providers appear to play a smaller role.  Primary care providers’ lack of involvement in the care of children with rheumatic diseases may reflect the lack of pediatric rheumatology training available in many pediatric residency programs, especially those programs with a high percentage of trainees who enter primary care practice.  Enhanced training of primary care providers may enable them to perform initial evaluation on children with suspected rheumatic diseases and minimize the number of unnecessary referrals, which increase the demand for pediatric rheumatology care.  Moreover, enhanced training of primary care providers may increase their willingness and ability to co-manage the care of children with rheumatic diseases and ease some of the patient care burden affecting pediatric rheumatologists.