The present study of American Indians with RMSF-associated hospital and outpatient visits adds current and important information to our knowledge of the occurrence of RMSF among American Indian populations, which to date has been based exclusively on inpatient records or national surveillance data, both of which substantially underestimate RMSF disease incidence among American Indians.9,10
The present study demonstrates that there is a significant increasing trend of RMSF incidence during 2001–2008, which supports findings from other studies that have demonstrated that RMSF has increased among American Indians during this decade.7,9
A similar increasing pattern of RMSF was also observed for the overall U.S. population by using national surveillance; these trends have been attributed in part to a heightened physician awareness of RMSF after recent publications on the topic, and by a changing RMSF case definition and new diagnostic tests of varying sensitivity and specificity.7,9
In this study, the evaluation of IHS outpatient records in addition to inpatient records permitted insight into RMSF treatment trends among the American Indian population that past studies have not provided. During 2001–2008, approximately 90% of all patients with a clinical diagnosis of RMSF were treated as outpatients, which is a higher percentage than is suggested by current national surveillance systems for the general U.S. population. This finding suggests that IHS providers are considering suspected RMSF diagnoses more frequently than national surveillance data. In fact, national surveillance systems may be biased toward the capture of more severe outcomes for RMSF because persons with mild illness or who respond quickly to outpatient treatment may not return for laboratory testing necessary to meet the RMSF confirmed or probable case definition for capture under national reporting guidelines. The inclusion of outpatient data also enables improved assessment of RMSF incidence among American Indians.
The use of inpatient and outpatient records with an ICD-9-CM code of RMSF showed that the estimated average annual incidence for American Indian patients with RMSF in 2001–2007 (87.9 per 1,000,000) is approximately five times the incidence for American Indians reported to the passive national surveillance system (16.8 per 1,000,000) for the same time period.7
Although these patients may not all have had appropriate diagnostic tests conducted to meet the required RMSF case definition for reporting under national notifiable disease guidelines, and this report may include some suspected RMSF cases for which infection is eventually ruled out, it seems clear that the incidence of disease among American Indians is likely underestimated by national surveillance systems.
The findings from the present study also support past studies that have suggested that American Indians have a higher average annual incidence of RMSF compared with the overall U.S. population.7–9
Possible reasons for this include the fact that RMSF is common in rural locations and American Indians within the IHS system typically live in rural areas and may participate in different recreational, social, or community activities that more frequently expose them to ticks compared with other race groups.8,32
Additionally, American Indians may be at overall higher risk for infectious diseases than other populations because of increased exposure, susceptibility, or underlying co-morbidities that increase risk for infection after exposure.33,34
This analysis also highlights several important epidemiologic differences in cases of RMSF among IHS regions. The Southern Plains had a higher annual average incidence rate than the Southwest region. However, the average annual rate in the Southwest could be the result of the fact that RMSF did not emerge as an important public health problem in Arizona (which is located in the Southwest region) until 2003, two years into our study period.14
Cases of RMSF in the Southern Plains and other regions tended to occur in older age groups, with the highest age-specific incidence among persons 50–59 and 60–69 years of age. In contrast, RMSF cases from the Southwest region were found more often in younger persons (median age = 12 years), and persons 1–4 and 5–9 years of age showed the highest incidence rates. In the Southwest region RMSF patients were also more likely to be hospitalized compared with the patients in the Southern Plains and other regions. The Southern Plains region also was the only region to show a trend toward decreasing hospitalization rates during the study period. In the Southern Plains and other regions, most RMSF health care facility visits occurred during June–September. In the Southwest region, there was a later seasonal shift; most visits occurred during July–October.
The high numbers of American Indian patients reported with RMSF in the Southern Plains region correspond with previous national surveillance reports stating that Oklahoma has one of the highest numbers of RMSF cases in the United States.7
In addition, American Indians in Oklahoma have historically had a higher rate of RMSF than the overall Oklahoma population. This finding could be caused by differences in frequency of exposure to ticks; ticks infected with R
may be highly concentrated in certain geographic areas and thus contribute to high rates of transmission in these areas.8
National surveillance for RMSF also suggests high numbers and rates of RMSF cases in North Carolina and Tennessee.7
Although our analysis did not show a corresponding high number of RMSF cases from the East region, even these few cases resulted in a fairly high incidence. The small number of American Indians living in this large and geographically diverse region makes it difficult to assess this finding more closely.
The high RMSF incidence rate and significant increasing trend in incidence in the Southwest region during the study period is likely caused by emergence of the brown dog tick (Rhipicephalus sanguineus
), which was shown to be a new vector for RMSF in eastern Arizona and for which cases began to be recognized as early as 2003.14
Despite the higher numbers of RMSF cases in the Southwest, RMSF incidence in this region is possibly diluted because the outbreak is focal (i.e., affecting only a few tribes), while the Southwest IHS service region covers a large geographic area and includes a large American Indian population.14,25
The seasonal variation in RMSF cases from this region may be caused by a different disease ecology in the Southwest.14,35,36
Specifically, Rh. sanguineus
, the tick responsible for transmission in this region, may have a different cycle of peak tick activity. Recent studies have shown that transmission of rickettsioses by Rh. sanguineus
is influenced by changes in climate; it has been shown that warmer weather increases the aggressiveness of Rh. sanguineus
, thus leading to more human attacks.37,38
These findings have raised the concern that changing climates could result in more emerging pathogens transmitted by the Rh. sanguineus
in the future.38
Furthermore, the unique predilection for pediatric cases in the Southwest region may reflect the close interaction of dogs and children in the peridomestic environment, and appears different from the epidemiology observed in the Southern Plains and other regions, and in the general U.S. population, in which older age groups have higher rates of RMSF.7
There are some important limitations to the use of hospital discharge and outpatient visit data based on ICD-9-CM-coded diagnoses. Many studies have illustrated the inaccuracy of ICD-9-CM codes for surveillance because of coding errors, physician errors, absence of laboratory verification, artificially abbreviated diagnosis code lists, and codes that do not correspond directly to clinical syndromes.39–41
However, studies have also been conducted to validate the use of ICD-9-CM diagnoses in the IHS health care system and have showed sensitivity and specificity for disease identification greater than 90%.42,43
In general, all current surveillance systems for RMSF are limited because complicated serologic testing is required to meet the confirmed RMSF case definition; interpretation of laboratory findings for cases in which there are antibodies present from previous exposures or cross-reactive antibodies may result in an incorrect designation of RMSF.7,44
Even with the national surveillance system, data for RMSF captured through the National Electronic Telecommunications System for Surveillance, which was designed to capture cases based on a laboratory and clinical diagnosis, cannot be reviewed for accuracy of case designation, and no details on the laboratory diagnosis are available. In 2001–2005, only 12.5% of American Indian cases met the confirmed RMSF case definition.9
Similarly, ICD-9-CM diagnoses for inpatient and outpatient visits are based on clinical diagnosis, and may not represent true RMSF cases. A previous medical chart review on RMSF hospital discharge records of American Indian hospitalizations in Oklahoma showed that nearly 45% of American Indian patients had a reported ICD-9-CM code for RMSF with no serum tested or only a single serum sample submitted for testing.8
Furthermore, other tick-borne diseases including R
infection and ehrlichiosis/anaplasmosis, have similar presenting symptoms, may occur in the same areas, and are clinically indistinguishable from RMSF. A study in North Carolina, a state that traditionally reports RMSF and ehrlichiosis/anaplasmosis, found that febrile patients with a history of tick bites were as likely to show serologic evidence of exposure to ehrlichiosis and anaplasmosis agents as to RMSF.45
In the present study, a small percentage of cases had Lyme disease (1.9%) and ehrlichiosis (0.2%) reported as accompanying diagnoses to the RMSF diagnoses, which suggests the lack of laboratory verification of the RMSF diagnosis. In addition, diagnostic criteria may vary by region and physician. Thus, in areas in which RMSF and ehrlichiosis/anaplasmosis are endemic, without laboratory diagnostics, other rickettsioses may be reported as RMSF. Efforts to identify the etiologic cause of disease with laboratory testing could help to properly classify these cases.
Other limitations of the present study include the fact that the incidence rates by region were calculated on the basis of residence rather than location of occurrence. Location of occurrence cannot be determined from the available data and location of residence was chosen rather than location of hospital or outpatient encounter for the purposes of incidence calculations and comparable user population denominators; this may bias regional results if persons were seen at a facility outside their region of residence and may account for RMSF cases occurring out of season. In addition, the American Indian/Alaska Native user population is an estimate of the number of American Indian/Alaska Natives who are eligible to use the IHS healthcare system and may not include all American Indian/Alaska Natives who are eligible for care at IHS/tribal facilities. Finally, the American Indians in this study may not be representative of all American Indians in the United States.
The findings from this study raise concerns about the high and increasing rates of RMSF among American Indians, and underreporting in the national RMSF passive surveillance system. Furthermore, the unique RMSF disease patterns in the Southwest region, including the higher number of pediatric cases and the increasing incidence trend in this region, highlight the need for further assessment of RMSF exposure risks caused by the emergent tick vector in this area. The suggestion of seasonal variation on the basis of region may also be an important detail to include in educational messages because RMSF cases were reported in later months, especially in the Southwest region. Physicians in the Southwest region need to be aware that RMSF cases may occur later in the season, and should plan to treat and test suspect patients accordingly. The finding that age-related risks vary by region is also important in terms of planning public health prevention messages. For example, educational messages targeted to school age children and parents regarding risks around their homes and the importance of treating homes and dogs may be important in the Southwest region, while general prevention messages for personal repellent use and behavioral modifications may be more appropriate in the Southern Plains region. It is important to continue to explore and monitor the epidemiology and trends of RMSF cases among American Indians to assess the efforts of public health interventions and to provide informative data for future public health policies and interventions.