Lyme Disease in Hispanics, United States, 2000–2013

Hispanics comprise a growing portion of the US population and might have distinct risk factors for tickborne diseases. During 2000–2013, a total of 5,473 Lyme disease cases were reported among Hispanics through national surveillance. Hispanics were more likely than non-Hispanics to have signs of disseminated infection and onset during fall months.


The Study
LD is a nationally notifiable condition, and cases are reported by state and local health departments to the Centers for Disease Control and Prevention (CDC) through the National Notifiable Diseases Surveillance System in accordance with previously established protocols (7). LD cases reported during 2000-2007 were confirmed cases only. In 2008, a revised case definition was implemented that altered the laboratory criteria and distinguished between confirmed and probable cases; cases reported during 2008-2013 included both categories (8).
We used 2010 US Census population data to calculate incidence rates (5,9). Weighting was applied to state-and county-specific numbers of cases to account for variations in completeness of ethnicity data. Descriptive statistics and comparisons were calculated by using SAS version 9.3 (SAS Institute, Cary, NC, USA). We compared median age of Hispanics and non-Hispanics with LD using the Kolmogorov-Smirnov 2-sample test. Risk ratios (RRs) were used to compare categorical data.
CDC human subjects review of the protocol determined it was not research involving human subjects. Thus, Institutional Review Board approval was not required.
During 2000-2013, a total of 374,338 LD cases were reported to CDC, of which 148,444 (39.7%) reports contained information about ethnicity and were included in this analysis. Among these, 5,473 (3.7%) persons selfidentified as being of Hispanic ethnicity. Most (54.8%) Hispanics with LD were male; median age was 32 years (interquartile range 15-46 years).
Annual incidence of reported LD among Hispanics was 0.8 cases/100,000 persons, compared with 4.0/100,000 among non-Hispanics. During 2000-2001, Hispanics comprised 2.8% of all persons with LD, whereas during 2009-2013, Hispanics comprised 3.7%-4.9% of persons with reported LD. In comparison, the proportion of Hispanics in the US population increased slightly during this period, from 13% in 2000 to 16% in 2010 (5).
Although a bimodal age distribution was evident among both Hispanics and non-Hispanics with LD, the peak in children was less pronounced among Hispanics ( Figure). Highest incidence among Hispanic children was in boys 10-14 years of age, whereas among non-Hispanic children, incidence was highest in boys 5-9 years of age. In adults, highest incidence among both Hispanics and non-Hispanics was in men 65-74 years of age.
Hispanics were significantly less likely than non-Hispanics to have disease onset during the summer months (RR 0.85, 95% CI 0.83-0.88) and more likely to have disease onset during the fall months ( reported clinical feature for both groups, it was less commonly reported among Hispanics than non-Hispanics (RR 0.83, 95% CI 0.80-0.86). Manifestations of disseminated disease, such as arthritis and facial palsy, were more commonly reported among Hispanics than non-Hispanics (Table 1). As expected, >90% of LD cases overall were reported from high-incidence states, although Hispanics with LD were slightly less likely to report residence in a high-incidence state (RR 0.90, 95% CI 0.82-0.98). All of the statistical associations were similar when analysis was restricted to confirmed cases only, with the exception of residence in a highincidence state, which became nonsignificant (RR 0.99, 95% CI 0.89-1.10).
After weighting, nearly half of all estimated cases of LD among Hispanics were from New York or New Jersey ( Table 2). Among counties with at least 75 estimated LD cases among Hispanics during the study period, highest incidence among Hispanics occurred in Columbia County, New York (170.4 cases/100,000 persons); Sussex County, New Jersey (111.4/100,000); and Hunterdon County, New Jersey (106.3/100,000).

Conclusions
Overall, the epidemiology of LD among Hispanics was similar to that among non-Hispanics: bimodal age distribution, slight predilection in males, and clustering in states to which LD is highly endemic were apparent (10). However, we identified several important differences. Most notably, Hispanics with LD were significantly more likely than non-Hispanics with LD to have signs of disseminated infection and symptom onset during fall months.
Although the overall incidence of LD in Hispanics was lower than that in non-Hispanics, additional research is needed to determine the reasons underlying these differences and the extent of any LD underdiagnosis in the Hispanic population. Inadequate healthcare access, language barriers, and lack of LD awareness could cause both underdiagnosis and delays in diagnosis in the Hispanic population. During 2009-2013, a total of 41.5% of Hispanics lacked health insurance, compared with 15.1% of non-Hispanic whites; 15.5% of Hispanics described delay in or nonreceipt of medical care because of cost (11). Furthermore, whether the predilection toward symptom onset in the fall  months for Hispanics results from delays in medical care or other factors, such as seasonal outdoor labor patterns, is unclear. Lastly, because a larger proportion of Hispanics than the overall US population live in urban areas (12), the risk for LD might be differentially diluted in Hispanics. Our findings were subject to several limitations. First, we had to exclude more than half of reported LD cases because of missing ethnicity data. Although we have no reason to believe that case reports with missing ethnicity data differed otherwise from those included in this study, we cannot exclude this possibility. Ethnicity reporting is also subject to error. Finally, surveillance data are limited by underreporting and reporting bias, which might differ by state and between Hispanic and non-Hispanic populations.
Reaching at-risk populations with culturally and linguistically appropriate prevention education is essential.
Although some educational materials about prevention of tickborne diseases have been translated to Spanish (13,14), additional translations and modifications to address cultural differences would be helpful. Furthermore, targeted educational campaigns could enhance use of these materials and improve the reach, retention, and overall impact of prevention education.
We identified specific risk groups and patterns of LD within the US Hispanic population. Direct and more in-depth assessments regarding prevention practices, knowledge, and LD epidemiology on local and national scales will further the understanding of LD risk in this population and guide future targeted prevention and education efforts. ‡Incidence calculated as number of annual estimated cases in county/100,000 Hispanic residents in county. Only counties with a substantial number of cases were included in this comparison. Seventy-five weighted cases was chosen as the cutoff based on distribution.