We used 100% inpatient, outpatient, skilled nursing facility, and carrier standard analytical as well as Medicare enrollment files for calendar years 2006–2008 to assess babesiosis among elderly Medicare beneficiaries ages >65 years of age. Standard analytical files are generated to capture medical services rendered and patient diagnoses, and enrollment files help to ascertain coverage eligibility. To be eligible for the study, beneficiaries had to be continuously enrolled in Medicare fee-for-service Parts A and B for >365 days before and including the latest month of continuous enrollment in the calendar year. We identified likely new babesiosis cases on the basis of the first recording of the International Classification of Diseases, 9th Revision, Clinical Modification diagnosis code 088.82 during the calendar year, with no recorded babesiosis infection in the preceding 365 days.

We assessed annual babesiosis rates by estimating the number of cases recorded per 100,000 beneficiaries per calendar year, overall and by sex, age, race, and state of residence. Seasonal occurrence was analyzed by using the number of cases in each month and number of beneficiaries continuously enrolled in Medicare fee-for-service within 365 days of each month. Cases were assigned age on the basis of diagnosis date, and persons without babesiosis had age assessed at the beginning of the latest enrollment month in the year. Beneficiaries with babesiosis were excluded from denominators of subsequent calendar years or diagnosis months. We performed χ² tests comparing babesiosis rates by using Epi Info version 3.5.1 (wwwn.cdc.gov/epiinfo). We conducted this institutional review board-approved study in coordination with the Centers for Medicare & Medicaid Services and within the SafeRx Project.

Among 27,278,865, 26,381,435, and 25,908,122 elderly Medicare beneficiaries in 2006, 2007, and 2008, respectively, there were 985 (3.6/100,000), 851 (3.2/100,000), and 1,223 (4.7/100,000) babesiosis cases based on International Classification of Diseases, 9th Revision, Clinical Modification diagnosis code (Table 1). Annual babesiosis rates (per 100,000 beneficiaries) among the white elderly were 4.0, 3.6, and 5.2, and among nonwhite elderly the rates were 0.6, 0.9, and 1.4 for 2006–2008, respectively. Rate comparisons for whites versus nonwhites showed significant differences (p<0.0001) for each year. Babesiosis rates were significantly (p<0.001) higher for men versus women and younger elderly women (ages 65–84 years) versus older elderly women (>85 years old) (Table 1). The Figure displays highest babesiosis rates in July and August of each year, with 74.4% of cases diagnosed in May through October.

Babesiosis varied by state, with the top 10 states and District of Columbia accounting for 834 (84.7%), 709 (83.3%), 1,024 (83.7%) of all cases in 2006–2008, respectively (Table 2). Connecticut, Rhode Island, New York, and Massachusetts had the highest babesiosis rates among the elderly (Table 2). Babesiosis cases were also recorded in other states, including California, Florida, Texas, Pennsylvania, Minnesota, and Wisconsin, with those 6 states accounting for 9.8%, 10.8%, and 12.3% of all cases in 2006, 2007, and 2008, respectively (data not shown).

We report a national population-based study of babesiosis among the US elderly, which used large administrative databases. The study found variations in the number of babesiosis cases by year, state, race, sex, age, and diagnosis month. Overall, our 3-year study suggests that there were more cases of babesiosis in 2008 compared with previous years. Northeastern and Mid-Atlantic States accounted for most newly diagnosed cases among the US elderly, with state-specific rates up to 10× higher than national annual rates. Our results show highest babesiosis rates in known babesiosis-endemic states of Connecticut, Rhode Island, New York, and Massachusetts and suggest possible expansion of human babesiosis to Maryland, Virginia, and other states. Human encroachment into tick and deer habitat, growth of deer population, climatic effects, and travel to disease-endemic areas may be responsible for variations in number of babesiosis cases and spread of the infection to non–disease-endemic states (1–3,9,11).

Our findings show babesiosis trends similar to surveillance results in Rhode Island and New York State over the same period, with fewer cases reported in 2007, and more in 2008 (12,13). Data from the Connecticut Department of Public Health demonstrate broad annual variation in numbers of reported babesiosis cases in 2000–2008, with an increasing trend over time (14). Our finding of higher babesiosis rates among men versus women and among younger elderly women versus older elderly women are generally consistent with state surveillance data (12,15). Similarly to the literature (1,2,6,12,15), our study shows that most babesiosis cases are diagnosed during May through October. These findings are likely related to life cycle and activity of the tick vector and to activity of human and other mammalian hosts (1,2,9).

Study limitations are related to the use of administrative databases and include difficulty in identifying incident versus prevalent cases, possible misdiagnosis, and lack of clinical detail for diagnosis verification and transfusion-transmitted babesiosis case identification, as well as inability to differentiate Babesia species. Medical record review is needed to address above-mentioned limitations. Choosing a different continuous enrollment period could produce slightly different rates. Although Medicare data do not provide population-wide information on babesiosis among persons <65 years of age, younger persons are more likely to remain asymptomatic and less likely to get a diagnosis (1,2,7).

Our nationwide large medical database study is an additional tool to better understand regional, seasonal, and other babesiosis transmission patterns, by year and demographic characteristics, among the US elderly. Because the elderly are also known to use the majority of transfused blood, studies are needed to evaluate transfusion-transmitted babesiosis in this group. Overall, our study suggests that large administrative databases can be useful in assessing emerging infections in the United States.