Figure 1

Figure 1. Flowchart of enrollment and completion by participants in study of economic burden of reported Lyme disease in high-incidence areas, United States, 2014–2016. LD, Lyme disease.

During the enrollment period, we identified 2,991 Lyme disease patients who were classified as having confirmed cases or probable cases with symptoms reported (Figure 1). Of the 1,360 (45%) patients we were able to contact, 1,118 (82%) consented to patient cost surveys; we included 901 (81%) participants with complete survey data in the patient cost analysis. Last, 613 (68%) of these participants also had complete societal medical cost data, and we included them in the societal cost analysis.

The study population included 402 (55%) confirmed localized, 238 (21%) confirmed disseminated, and 261 (24%) probable cases (Table 1). Overall, 36% of participants were 46–65 years of age, 57% were men, and 94% were white. Most had income >$60,000 (71%) and private health insurance (70%). Demographic distributions were similar for the 613 participants who completed both patient cost surveys and societal medical cost collection (Appendix Table 3).

Participants reported a median of 2 provider visits and completed a median of 3 surveys (Table 2). Those with confirmed disseminated disease had the highest number of provider visits, reflecting the highest healthcare use, whereas those with probable disease had the highest number of surveys completed, reflecting the longest duration of costs incurred. Forty (4%) participants were still reporting symptoms and 25 (3%) were still incurring costs at survey 12.

Overall, the patient cost per participant ranged from $0.46 to $30,628. The median cost was $244 and the mean cost $1,252, reflecting a highly positively skewed distribution (Table 3). Participants with confirmed disseminated Lyme disease had the highest median cost ($358) and mean cost ($1,692), followed by those with probable disease (median $315 and mean $1,277), then participants with confirmed localized disease (median $170 and mean $1,070).

Figure 2

Figure 2. Mean and median cost per participant, by Lyme disease category and cost category of the total patient cost in high-incidence areas of the United States, 2014–2016. A) Confirmed localized disease;...

We calculated the median and mean cost per component of the patient cost by disease category (Figure 2; Appendix Table 4). For all disease categories, productivity losses had the highest mean cost of all cost components: $727 for those with confirmed disseminated disease, $627 for those with probable disease, and $540 for those with confirmed localized disease. However, the median cost of productivity losses for all disease categories was $0. Medical bills had the next highest cost: a median of $83 and a mean of $628 for those with confirmed disseminated disease, a median of $83 and a mean of $389 for those with probable disease, and a median of $42 and a mean of $314 for those with confirmed localized disease. All other cost components for all disease categories had median costs <$22 and mean costs <$80.

We collected 9,679 CPT codes to estimate societal medical costs. The most common codes were for office visits (17%) and routine venipuncture (6%) (Appendix Table 6). Overall, the societal medical cost per participant ranged from $50 to $121,869, for a median of $478 and mean of $1,333 (Table 4). Participants with confirmed disseminated disease had the highest median and mean societal medical cost ($696 and $2,537), followed by those with probable disease ($612 and $1,804), then confirmed localized disease ($374 and $668).

Overall, the societal cost of Lyme disease per participant ranged from $54 to $122,766; the median was $690 and the mean $2,032 (Table 5). Participants with confirmed disseminated disease had the highest median and mean societal cost ($1,081 and $3,251), followed by those with probable disease ($940 and $2,620), then confirmed localized disease ($493 and $1,307) (Appendix Table 7). Applying these per participant societal costs to estimates of the number of Lyme disease cases diagnosed each year (6), the aggregate cost to US society annually would be ≈$345 million using median costs and ≈$968 million using mean costs (2016 US dollars; Appendix Tables 9, 10).

In multivariable linear regression analysis, disease category, age, and state were associated with societal cost per participant (Table 6; Appendix Table 8). Costs for participants with confirmed disseminated disease were 120% higher than costs for participants with confirmed localized disease (p<0.001); participants with probable disease had costs that were 59% higher than for those with confirmed localized disease. Participants 18–45 and 46–65 years of age had costs that were 96% and 108% higher, respectively, than those <18 years of age (p<0.001); however, those >65 years of age did not have significantly different costs. Minnesota residents had 75% higher costs than did Connecticut residents, but Maryland and New York residents did not have significantly different costs from those for Connecticut residents.

In this study, persons with confirmed or probable Lyme disease had an average patient cost of ≈$1,200 (median cost ≈$240) and an average societal cost of ≈$2,000 (median cost ≈$700). In stratified analyses by disease category, those with confirmed disseminated or probable disease had double or more the societal cost per participant than those with confirmed localized disease, highlighting the importance of early and accurate diagnosis. Having disseminated or probable disease, being 18–65 years of age, and residing in Minnesota had the greatest effects on the societal cost of Lyme disease. Although median societal costs were typically <$1,000 for all disease categories, mean costs were substantially higher, indicating that most patients have low costs, but some experience very high costs related to this disease. Similarly, the low median number of provider visits and hours of lost productivity suggest that Lyme disease illness is manageable for most patients, but for a minority, it can be highly disruptive. Approximately 476,000 cases of Lyme disease are diagnosed each year in the United States, so the aggregate cost to society annually could be $345–968 million (2016 US dollars). This substantial economic burden underscores the need for effective prevention methods, such as a vaccine.

Classification of a reported case as probable means a clinician has diagnosed Lyme disease in a patient and laboratory evidence of infection exists. However, any reported symptoms are typically nonspecific and do not meet clinical criteria for a confirmed case (24). Further, laboratory evidence of infection includes single-tier IgG immunoblot seropositivity, which might indicate past, rather than current, infection. As such, the increased costs for probable cases might result from higher healthcare use for disease unrelated to Lyme disease, which highlights the need for improvements in Lyme disease diagnosis and clinician education.

In a geographically limited study of Lyme disease patients residing on the eastern shore of Maryland during 1998–2001, Zhang et al. (9) reported mean total costs of $3,494 and median total costs of $500 (2016 US dollars) per patient attributable to this disease. However, their case definition differed from ours in its inclusion of patients with early, late, and suspected disease, as well as those with tick bite and other related complaints, as identified using diagnosis codes in medical records. Therefore, these figures might not be directly comparable to our mean and median societal costs ($2,032 and $690). Zhang et al. reported mean and median total costs of $2,275 and $689 (2016 US dollars) for participants with clinically defined early disease, which are higher than what we found for confirmed localized disease (mean $1,307 and median $493). However, in regression analyses, Zhang et al. found that disease category and age, but not sex, were significantly associated with societal medical costs, similar to our findings for societal cost. In another US study using nationwide commercial insurance claims data to compare cases with matched controls during 2006–2010, Adrion et al. (8) estimated an increase of $3,009 (2016 US dollars) in societal medical costs attributable to Lyme disease over a 12-month period. That cost is higher than our overall mean societal medical cost ($1,333), likely because of study population differences, but is similar to that found for our participants with confirmed disseminated disease ($2,537). In a recent study in the Netherlands, Van den Wijngaard et al. (17) used a societal perspective to estimate a total cost of $137 for patients with erythema migrans only and $6,398 (2016 US dollars) for those with disseminated Lyme borreliosis. These costs are lower than those for our societal results for confirmed localized disease ($1,307) and higher than those for our societal results for confirmed disseminated disease ($3,251). These cost differences might result from different healthcare financing systems in the United States versus Europe or from variations in clinical manifestations resulting from infection with different B. burgdorferi sensu lato strains (15–17).

Our study adds to the scarce literature on the economic burden of Lyme disease and provides a comprehensive estimate of its costs to both the patient and society. In addition, our prospective collection of all patient costs, including nonmedical costs and productivity losses, enables more accurate and more comprehensive cost estimates compared with previous studies in the United States. Further, these results provide estimates of the cost savings per case averted, which can be used in cost-benefit analyses of prevention interventions, such as a potential vaccine.

The first limitation of our study is that our estimates might be affected by recall error, either by participants or providers, although we attempted to mitigate such error by enrolling patients as close to disease onset as possible, by surveying them monthly to capture ongoing costs, and by requesting codes from providers for a date range instead of for individual visits. However, by requesting codes over a date range, some billing codes unrelated to Lyme disease (e.g., for other comorbidities) might have been included despite our excluding codes definitively unrelated to Lyme disease, potentially leading to overestimates. Information bias might have occurred in our measure of association between disease category and cost because those with milder disease might be more likely to forget some costs than those with more severe disease, with a potential bias away from null. In addition, although the use of quota sampling to recruit reported cases was necessary to enroll patients near disease onset, this nonprobability sampling method limits our ability to meet assumptions for calculating sampling error. Use of surveillance data to weight responses by disease category was intended to ensure representativeness by disease category. Nevertheless, in surveillance data, confirmed localized cases are likely underreported, resulting in confirmed disseminated cases representing an artificially large proportion of cases; therefore, our overall cost might be overestimated (35,36). Finally, this study did not include costs related to deaths from Lyme disease, because no enrolled participants died. Although very rare, death from Lyme carditis has been reported (2,3), and associated productivity losses would greatly increase cost estimates.

Our results reflect the costs of diagnosed cases meeting the Lyme disease surveillance case definition in high-incidence states (24); as such, these costs likely reflect that of actual infections. However, we were not able to evaluate whether our estimates accurately represent the cost of diagnosed but unreported Lyme disease, cases that reflect some proportion of overdiagnosis (6). Further, our results might not reflect costs in states with emerging or low incidence of Lyme disease. Therefore, our results for extrapolation of costs to an estimated ≈476,000 diagnosed cases nationally per year should be interpreted with caution. Last, our results do not include costs for suspected Lyme disease (e.g., consultation and prophylactic treatment for tick bite, negative diagnostic tests), undiagnosed disease, or nonacute disease (e.g., patients experiencing long-term symptoms). These costs would further increase the total economic burden attributable to Lyme disease. Future efforts should include cost-effectiveness analyses of current and future prevention methods, such as a vaccine, in addition to economic evaluations of unreported, suspected, and nonacute disease.

In conclusion, Lyme disease represents a substantial economic burden to individual patients and US society. The aggregate cost of diagnosed Lyme disease could be nearly $1 billion annually, not including suspected, undiagnosed, or nonacute cases. These findings emphasize the importance of early and accurate diagnosis to reduce both illness and its associated personal and societal costs.