Although rare individually, the rheumatic diseases, taken together, are among the most common chronic health conditions affecting children [1
]. Exact prevalence rates among children living in the United States are difficult to obtain, owing, in large part, to the de-centralized delivery of health care in this country. The IHS represents an exception to that decentralization and is, arguably, the closest representation to a nationalized health care delivery system currently functioning within the United States. Thus, records and data available through the IHS represent a unique opportunity to assess population-wide health needs not otherwise available to child health researchers in this country. This report provides a first-ever population-wide estimate of the prevalence of chronic arthritis in American Indian children living within the United States. While the prevalence rate for the Oklahoma City Area was within the same order of magnitude as the most recent reports from Europe [35
], the prevalence rate in the Billings Area was nearly 10 times this recent European estimate. These findings are consistent with earlier studies of rheumatoid disease in American Indian adults, where prevalence rates 10 times higher than the general population were reported [49
]. It should be pointed out, however, that prevalence estimates of JRA vary widely, ranging between 16 to 113 per 100,000 [30
The reasons for the discrepancy in prevalence estimates between the Oklahoma City and Billings Areas are not clear. One possibility is that the northern plains tribes are particularly susceptible to rheumatoid disease in ways that other groups (e.g., Eastern Woodlands or Southwestern tribes) are not. It is also possible that the difference in ethnic composition of the two populations accounts for this difference. While many Oklahoma tribes require at least a 25% blood quantum of tribal ancestry (e.g., the Kiowa tribe [56
]), other tribes require only proof of descent from an individual on the original Dawes rolls of 1893 [57
]. Thus, the Oklahoma City Area includes many individuals whose degree of American Indian ancestry is 1/4 or less and may include individuals with less than 1/64 American Indian ancestry. In contrast, there has been less intermingling between Caucasian and American Indian populations on the northern plains, and a larger percentage of the Billings Area population includes individuals with full-blooded American Indian ancestry.
Our study once again points out the rarity of the pauciarticular form of JRA in non-European populations. In studies of European and European-descended populations, pauciarticular JRA is the most common form of chronic childhood arthritis [3
The Oklahoma City Area database listed a single child with ICD-9 codes #714.32 (pauciarticular JRA) or 714.33 (monoarticular arthritis), the codes used to identify such children. These findings are consistent with reports from the Indian subcontinent,[7
] Kuwait [5
], Turkey [12
], Thailand [13
], Japan [14
], South Africa [15
], and with our experience with African American children in Detroit [16
The slight female-to-male preponderance for spondyloarthopathy is also worth noting. High prevalence rates for spondyloarthopathies have been noted in both Northwestern and Southwestern tribes [24
]. However, in these studies, a strong male preponderance was noted. Whether the findings from Oklahoma City and Billings represent a novel finding or inaccuracies in the ICD-9 coding await confirmatory studies, as we discuss below.
An important limitation to this study is the fact that we did not have the means to verify every individual case listed in Oklahoma City database and were unable to confirm any diagnosis in the Billings Area database. However, our limited test of the accuracy of the Oklahoma City data provided surprising confirmation of the accuracy of coding for known cases. While we could not confirm any of the Billings cases, our search of the database of a single IHS facility in the neighboring Aberdeen Area corroborated the prevalence statistics we derived from the Oklahoma City and Aberdeen databases. Indeed, our experience suggests that a search strategy like the one we used is likely to under-estimate rather than over-estimate the prevalence of rheumatic disease in the IHS user population.
We are aware that there are many factors that might overestimate disease prevalence using this type of database search. The first is the possibility that a given ICD-9 code might have been used to designate a "working" diagnosis that was never established by the patient's clinical course. The second opportunity for overestimation of prevalence would occur if children were systematically misdiagnosed. This could occur easily if physicians use serologic data as the sole criterion for diagnosis. For example, many physicians routinely screen children with musculoskeletal complaints using antinuclear antibody (ANA) tests. However, the prevalence of low-titer positive ANA tests is extraordinarily high in the pediatric population [58
]. Thus, if ANA-positive children with musculoskeletal pain [59
] are listed as having "JRA," then there would be a gross overestimation of the actual prevalence.
Similarly, there are factors that might have led to underestimation of JRA prevalence by relying solely on a three-year database search. Children or adolescents with well-controlled JRA may not have seen an IHS physician during the relevant time period, and thus would have been excluded. Similarly, physicians who rely on rheumatoid factor tests as a diagnostic criterion for JRA might fail to diagnose the disease in a child, since only a small number of children with JRA have detectable IgM rheumatoid factor [60
The ideal method for obtaining true disease prevalence rates would include rigorous, pro-active case finding in a known population at risk. This approach was taken by Manners and Diepeveen in a study of school children in Australia . Using such an approach, these authors reported a prevalence rate of 4 per 1,000 for JRA, significantly higher than any previous estimates. We are now preparing a similar project involving American Indian communities on the northern plains.