Special Article
Prevention of Lyme Disease after Tick Bites — A Cost-Effectiveness Analysis
N Engl J Med 1992; 327:534-541August 20, 1992
- Abstract
- Abstract
Background.
In areas of endemic disease, the probability of Lyme disease after a tick bite ranges from about 0.012 to 0.05. Early treatment with oral antibiotics prevents most complications of Lyme disease, but antibiotics are generally not prescribed until rash or other symptoms develop.
Methods.
We used decision analysis to evaluate the outcomes, costs, and cost effectiveness of three alternative strategies to treat patients bitten by ixodes ticks in areas of endemic Lyme disease: empirically treat all patients with two weeks of doxycycline, treat only patients in whom erythema migrans develops, and treat only patients with erythema migrans or a positive serologic test for Lyme disease one month after exposure.
Results.
Empirical treatment is the least expensive strategy and results in the fewest cases of Lyme disease and the fewest complications when the probability of Borrelia burgdorferi infection after a tick bite is 0.036 or higher. For probabilities of infection below 0.036, empirical therapy prevents most major complications, sequelae, and adverse events, but it incurs additional minor complications, especially as the probability of infection falls below 0.01.
Conclusions.
Empirical treatment of patients with tick bites is indicated when the probability of B. burgdorferi infection after a bite is 0.036 or higher, and this treatment may be preferred when the probability of infection ranges from 0.01 to 0.035. When the probability of infection after a tick bite is less than 0.01, empirical therapy is not warranted. (N Engl J Med 1992;327:534–41.)
Article Activity
- Article
LYME disease is a multisystem disorder caused by the spirochete Borrelia burgdorferi. Spread to humans by infected ticks, it is the most common vector-borne illness in the United States.1 The primary vectors are Ixodes dammini in the northeast and midwestern United States2 , 3 and I. pacificus on the West Coast.4 The probability of contracting Lyme disease after a tick bite in an area of endemic disease has been reported to range from 0.012 to 0.05 (pooled estimate from four studies, 0.027; 95 percent confidence interval, 0.013 to 0.049).5 6 7 8
Early Lyme disease (localized infection) is often marked by the development of influenza-like symptoms and erythema migrans, a characteristic target-shaped rash that typically occurs 2 to 32 days after inoculation.9 , 10 Early Lyme disease is usually cured by oral antibiotics.11 However, approximately 20 to 40 percent of people infected with B. burgdorferi either do not have or do not detect erythema migrans.12 13 14 In patients with untreated localized infections, hematogenous dissemination of spirochetes to secondary sites may occur. Cardiac, neurologic, and rheumatologic manifestations are the most common sequelae of disseminated Lyme disease. The sequelae of late Lyme disease may be refractory to oral antibiotic treatment and may require intravenous therapy.14 15 16 17 Untreated patients infected with B. burgdorferi may remain asymptomatic without sequelae.13 14 15 16 17 18 19
The purpose of this study was to address a question frequently faced by physicians practicing in areas where Lyme disease is common: What is the optimal treatment of persons bitten by ixodes ticks in such areas? The current standard of care, which is based on little data, is to observe the patient and treat only if clinical manifestations of Lyme disease (early or late) appear.6 , 20 We used decision analysis to compare outcomes, costs, and the cost effectiveness of alternative management strategies to treat patients bitten by ticks.
Methods
Structure of the Decision Model: The Basic Tree
Figure 1Figure 1
Model of the Decision Tree, Outlining the Treat All, Follow, and Test Strategies. shows the decision model used in this analysis. Three strategies were compared for the prevention of late sequelae of Lyme disease in patients bitten by a tick in an area of endemic Lyme disease: (1) Treat All — a strategy to treat all patients empirically with two weeks of oral antibiotics; (2) Follow — a strategy to treat only patients who have erythema migrans; and (3) Test — a strategy to treat only patients who have erythema migrans or who have a positive serologic test at follow-up one month after tick exposure. Because treatment of early Lyme disease has a very low cost and risk and greatly reduces the likelihood of late sequelae,11 a strategy of no treatment was not considered practical.The terminal branches of the basic tree that end with "Treated Lyme," "Untreated Lyme," and "Unnecessary Treatment" were extended to include the probability of late major adverse sequelae of B. burgdorferi infection (which are then treated with appropriate antibiotic regimens), the probability of antibiotic failure, and the probability of adverse reactions to treatment with oral antibiotics. The analysis assumed that the probabilities of neurologic sequelae, rheumatologic sequelae, cardiac sequelae, and reactions to antibiotics were mutually exclusive (a conservative assumption that biases the results against finding a benefit for empirical treatment).
Data and Assumptions
The estimates of probability used in the model (Table 1Table 1
Estimated Probabilities for the Variables Included in the Decision Model to Determine Optimal Strategies for Treating Patients with Tick Bites.) were obtained by surveying the literature and consulting experts in infectious disease, cardiology, neurology, and rheumatology. For each variable a best estimate was determined, and a range of plausible values was selected to represent the degree of uncertainty surrounding the variable.Epidemiologic Variables
The model incorporated the probability of infection after a tick bite, the performance characteristics (sensitivity and specificity) of testing for antibody to the Lyme disease spirochete in cases of localized disease when a single test is performed one month after infection, and estimates of patients' probable degree of compliance with recommended care. Sequelae of Lyme disease were defined as conditions that occurred as a consequence of B. burgdorferi infection. Erythema migrans was classified as a minor sequela; cardiac, neurologic, and rheumatologic complications of late Lyme disease were considered major sequelae.
Treatment Variables
The treatment regimens used in the model were based on current recommendations.14 15 16 17 Oral antibiotics were assumed to be the treatment for patients with erythema migrans, first-degree atrioventricular block, isolated facial palsy without cerebrospinal fluid abnormalities, and arthritis of new onset. Doxycycline was used in this model, because it may be superior to penicillin and erythromycin in patients with early Lyme disease.11 , 17 Doxycycline achieves better tissue levels than other tetracyclines, is taken only twice a day, and is associated with minimal gastrointestinal distress.17 , 33 Doxycycline (100 mg orally twice daily for 14 days) was used to treat localized infection and first-degree atrioventricular block.14 , 15 , 16 , 17 A 30-day regimen of the drug was used in patients with arthritis or isolated facial palsy.14 , 15 , 17
Intravenous antibiotics were assumed to be the treatment for patients with high-grade atrioventricular block, neurologic abnormalities other than isolated facial palsy, and arthritis resistant to an initial course of oral antibiotics.14 , 15 , 17 Ceftriaxone (2 g intravenously daily for 14 days) was used, since intravenous ceftriaxone may be superior to intravenous penicillin in the treatment of patients with late neurologic and rheumatologic manifestations of Lyme disease.14 , 15 , 17 , 36 Patients who did not respond to an initial course of intravenous antibiotics received a second 14-day course of treatment.
The degree of a patient's compliance with prescribed medication has been reported to be directly related to the patient's perception of the seriousness of the illness and the associated level of anxiety.35 Using these data and the judgment of experts, we estimated the degree of compliance with prescribed antibiotic therapy in various phases of Lyme disease for each management strategy.
Reactions to Antibiotics
Reactions to antibiotics were defined as adverse occurrences resulting from antibiotic therapy. Minor reactions to oral doxycycline included the gastrointestinal effects of nausea, vomiting, diarrhea, and esophagitis; mild hypersensitivity reactions of rash, urticaria, and photosensitivity; and oral and vaginal candidiasis.37 , 38 Major reactions to oral doxycycline included enterocolitis, anaphylaxis (including angioedema), Stevens—Johnson syndrome, severe urticarial reactions, and a lupus-like syndrome.37 , 38 Minor reactions to intravenous ceftriaxone included the gastrointestinal symptoms of abdominal pain, nausea, vomiting, and diarrhea; hypersensitivity reactions of rash, pruritus, and fever and chills; candidiasis; and local reactions at the injection site.39 , 40 Major reactions to ceftriaxone included enterocolitis, serum sickness—like reactions, and renal failure.39 , 40
Evaluation of Use of Services and Costs
The use of medical services by patients with various sequelae of Lyme disease was estimated on the basis of audits of charts of patients with the disease. Twenty inpatient charts were reviewed, 10 from a community teaching hospital and 10 from a university hospital. Twenty outpatient charts, provided by two community-based physicians and two university-based physicians, were also reviewed.
In addition to the chart review, eight physicians expert in the management of Lyme disease and its sequelae (cardiologists, neurologists, and rheumatologists) were interviewed. The respondents included faculty members of medical schools and physicians practicing in the community who had admitting privileges at teaching hospitals. Clinical scenarios in which otherwise healthy patients had complications of Lyme disease relevant to the specialist (e.g., Bell's palsy in the case of neurologists, atrioventricular block for cardiologists, and arthritis for rheumatologists) or an antibiotic-related complication were presented to the physicians. They were asked in detail about how they would treat the patient (i.e., with hospitalization, diagnostic tests, or antibiotics).
The model incorporated various assumptions about treatment, as follows. Patients with first-degree heart block with PR intervals of less than 0.30 second received outpatient treatment; those with higher-grade heart block received inpatient treatment. Patients with arthritis were treated entirely as outpatients. Patients with chronic arthritis refractory to both oral and intravenous antibiotic therapy were treated with outpatient arthroscopic synovectomy. Patients with meningoencephalitis or meningitis (with or without cranial or peripheral neuropathy) were evaluated as inpatients; patients with an isolated cranial neuropathy or an isolated radiculoneuritis were treated as outpatients. Patients with major antibiotic reactions were admitted to the hospital. Patients with minor antibiotic reactions were treated as outpatients. Patients were not hospitalized for the sole purpose of receiving part or all of their intravenous antibiotic therapy, but were assumed to use a home infusion service.
Costs
The model was confined to the assessment of direct medical costs resulting from short-term and intermediate-term sequelae of Lyme disease (a conservative assumption that biases the analysis against finding a benefit for empirical treatment). Data for 1989 on mean payments for medical services (laboratory tests, physicians' services, hospital care, and hospital medications) were obtained from New Jersey Blue Cross—Blue Shield and were used to estimate the costs associated with the use of medical services. Average costs of antibiotics were calculated on the basis of a survey of charges at 10 local pharmacies. The cost associated with each potential outcome was determined by aggregating the various elements of the payments from Blue Cross—Blue Shield and pharmacies (Table 2Table 2
Estimated Costs of Management of Sequelae of Lyme Disease.). Since only short-term and intermediate-term complications of untreated or ineffectively treated Lyme disease were examined (those that occurred within one year of infection) and long-term complications were excluded, costs and benefits were not discounted.Health Outcomes
Health outcomes and costs were calculated for a hypothetical cohort of 100,000 persons bitten by ticks and treated with each of the three strategies. For each strategy, the major reactions to antibiotics and the major sequelae of Lyme disease were totaled in order to determine the number of major complications. Minor reactions to antibiotics and minor Lyme sequelae were totaled to determine the number of minor complications.
Outcome Effectiveness and Cost Effectiveness
Incremental cost-effectiveness ratios were used to compare management strategies when one alternative was more effective in preventing disease but was more expensive than the other strategy. The incremental cost effectiveness of one strategy relative to a second strategy was expressed as the extra dollar cost incurred, divided by the number of additional major complications prevented. Similarly, incremental outcome effectiveness was expressed as the number of extra minor complications incurred, divided by the number of extra major complications prevented.
Sensitivity Analysis
The estimates of epidemiologic characteristics, the incidence of Lyme sequelae, compliance with treatment, treatment effectiveness, reactions to antibiotics, and cost varied over a plausible range of values as determined from the literature and estimated by experts, and the effect of these changes on outcomes, costs, and cost effectiveness was calculated.
Results
Base-Case Analysis: Health Outcomes and Costs
The health outcomes and dollar costs for a hypothetical cohort of 100,000 persons bitten by ticks and treated by each of the three alternative strategies are shown in Table 3Table 3
Health Outcomes and Costs When Best Estimates Were Used in a Hypothetical Cohort of 100,000 Persons with Ixodes Tick Bites Treated with One of Three Treatment Strategies.. The Treat All strategy would incur the fewest major sequelae of Lyme disease and would be the least expensive strategy, but it would incur more minor complications than the other strategies because of the increased incidence of reactions to antibiotics. Under base-line conditions (probability of contracting Lyme disease after a tick bite, 0.03), Treat All would incur 432 major complications (major sequelae of Lyme disease plus major antibiotic reactions) and 3475 minor complications (minor sequelae of Lyme disease plus minor antibiotic reactions) per 100,000 patients. Follow would incur 1065 major and 2234 minor complications. Test would incur 838 major and 2329 minor complications. The average dollar costs per patient according to this best estimate would be as follows: Treat All, $24.58; Follow, $36.60; and Test, $60.91.Worst-Case Sensitivity Analysis
When the probability that B. burgdorferi infection will develop in a patient after a suspected ixodes tick bite was estimated at 0.03, the relative health outcomes and costs of the three strategies did not change when each variable was studied over the full range of its reasonable values. Therefore, we created a worst-case scenario in which the desirability of the Follow and Test strategies was maximally enhanced relative to the Treat All strategy.
The worst-case model included the following assumptions: the highest estimate of the probability of erythema migrans in B. burgdorferi infections (0.80); the lowest estimates of late sequelae of Lyme disease in untreated Lyme infections (cardiac, 0.04; neurologic, 0.15; rheumatologic, 0.50); the highest estimates of antibody-test sensitivity (0.80) and specificity (0.99) at one month; the highest estimates of patients' follow-up for antibody testing one month after the tick bite (0.75); the lowest estimate of compliance with empirical oral antibiotics (0.55); the lowest estimate of efficacy of oral antibiotics used to treat localized infection (0.85); the highest estimates of minor and major reactions associated with the use of oral antibiotics (0.06 and 0.0005, respectively); and the lowest costs of treating patients who have major sequelae of Lyme disease (25 percent less than the base-line estimates).
According to this worst-case scenario, the Treat All strategy would prevent the most serious complications and would still be the least expensive alternative. Follow would be the strategy with the fewest minor complications. The Test strategy would be more expensive than the other two management strategies examined and would be the optimal one in terms of major health outcomes only when the probability of B. burgdorferi infection after a tick bite was less than 0.0001. Even then, it would be almost twice as expensive as the other strategies. Therefore, our incremental analyses of effectiveness compare the Treat All and Follow strategies.
Incremental Analysis of Effectiveness and the Probability of Infection
Figure 2Figure 2
Incremental Cost Effectiveness of the Treat All Strategy Relative to the Follow Strategy, According to the Probability of Infection. shows the incremental cost effectiveness of the Treat All strategy relative to Follow as the probability varies for infection with B. burgdorferi after a tick bite. At probabilities of infection higher than 0.015, Treat All would both prevent more major complications than Follow and cost less. At probabilities of infection less than 0.015, the Treat All strategy would prevent more major complications than Follow but would cost more.Figure 3Figure 3
Incremental Outcome Effectiveness of the Treat All Strategy Relative to the Follow Strategy, According to the Probability of Infection. shows the incremental outcome effectiveness of the Treat All strategy relative to the Follow strategy as the probability of infection varies. At probabilities of infection higher than 0.055, Treat All would prevent more major complications than Follow and would incur fewer minor complications. At probabilities of infection ranging from 0.036 to 0.055, Treat All would incur more minor complications than Follow, but it would prevent more major complications and would be associated with fewer complications in all. For probabilities of infection of less than 0.036, Treat All would result in more than one additional minor complication for each major complication prevented. When the probability of infection falls below 0.01, the ratio of excess minor complications incurred per major complication prevented would rise sharply with Treat All as compared with Follow.Discussion
There are few data on the appropriate treatment of patients with tick bites in areas of endemic Lyme disease. Well-designed clinical trials of adequate size have not been reported. Thus, this study used decision analysis to create a model of the best available data, to address the question of how to treat patients bitten by ticks potentially infected with B. burgdorferi.
Treat All was the strategy of choice when the probability of infection was 0.036 or more, since it incurs the fewest major complications, has the fewest complications overall, and is cheapest. For probabilities of infection between 0.036 and 0.01 the physician and patient must determine the strategy they prefer on the basis of their beliefs about the value of preventing major complications in relation to the dollar costs and the number of minor complications incurred. However, at these probabilities, Treat All appears to be the strategy of choice, since it still incurs the fewest major complications, while incurring only relatively few additional minor complications. As the probability of infection after a tick bite falls below 0.01, Follow becomes preferred, since the number of excess minor complications incurred by empirical antibiotic therapy greatly exceeds the number of major disease sequelae averted.
All clinical models, data, and decisions are limited by the need to use imperfect data and make necessary simplifying assumptions. However, the results generated by this model appear robust. The base-line probabilities used, although imperfect, represent the best estimates available from the literature and from experts. Moreover, for a given probability of infection, the relative costs, benefits, and complications of the three strategies examined do not change when a wide range of assumptions is considered with regard to cost, the incidence of sequelae of Lyme disease, the risk of reactions to antibiotics, compliance with antibiotic therapy, and the efficacy of such therapy.
Although reimbursements from Blue Cross—Blue Shield and outpatient pharmacy charges may not fully represent the costs of treating patients, the preferred strategies will not change even if the absolute dollar values change, as long as the cost estimates are not biased with respect to the strategy evaluated. Finally, whereas this economic analysis considered only the relatively limited direct medical costs of managing the immediate sequelae of Lyme disease, consideration of nonmedical direct costs (e.g., travel costs incurred in the course of receiving medical care), additional out-of-pocket costs of care, direct medical costs incurred by patients with long-term complications, indirect costs (e.g., work loss), and intangible costs (e.g., pain and suffering) would increase the benefits of the Treat All strategy even more in relation to the other strategies.
This analysis applies only to patients eligible for doxycycline therapy. Patients in whom tetracyclines are contraindicated (children less than eight years old and pregnant or lactating women) were excluded. Discoloration and severe dysplasia of the teeth and inhibition of bone growth are associated with the administration of tetracyclines during the second and third trimesters of pregnancy, when most women are aware of their pregnancies. However, women of childbearing age who have missed one or more periods should have a serum pregnancy test before starting doxycycline therapy. With this precautionary measure, adverse effects on the fetus from tetracycline therapy are very unlikely.
Only a small portion of patients bitten by ticks are candidates for empirical treatment, because the large majority of bites are unrecognized. Hence, empirical therapy is not likely to reduce the number of cases of Lyme disease substantially. For similar reasons, the potential for increased resistance to antibiotics as a result of empirical therapy is unlikely to be a substantial problem. Tetracyclines are commonly prescribed. The increase in use attributable to empirical therapy of patients with identifiable ixodes tick bites in areas of endemic disease is very small. The efficacy of penicillin in killing Treponema pallidum, the spirochete with which we have the most experience, remains undiminished despite 40 years of extensive use.
In the one published randomized clinical trial of prophylactic therapy in patients bitten by ixodes ticks, 1 of 29 patients in the placebo group had erythema migrans, and 1 of 27 patients who received empirical antibiotic therapy had a penicillin-associated rash.6 The authors concluded that the risk of acquiring Lyme disease from a recognized tick bite approximates the risk of an adverse reaction to antibiotic therapy.
The model used in our study suggests that this conclusion is correct when only minor complications are considered. At a probability of infection of 0.03, the incidence of minor complications (early Lyme sequelae plus minor reactions to antibiotics) is about equal for the Treat All and Follow strategies. The potential value of empirical therapy, however, is in preventing the cardiac, neurologic, and rheumatologic sequelae that may arise in persons infected with B. burgdorferi in whom erythema migrans does not develop. When the probability of infection after a tick bite is 0.03, the model indicates that 640 extra cases of major Lyme sequelae are averted, but that only 7 additional cases of major drug reactions are incurred with the Treat All strategy.
The relative desirability of the management strategies depends on the probability of incurring Lyme disease from a tick bite. The probability of infection from a tick bite is a function of the probabilities that the biting arthropod is an ixodes tick, that the tick is carrying B. burgdorferi, and that the spirochete is transmitted from tick to human. These factors must be assessed before it can be decided whether empirical therapy is warranted.
The nature of the tick is derived from the clinical history. If it cannot be established with a very high degree of certainty that the patient was bitten by an ixodes tick (if the tick is unavailable for identification by the physician and if the patient's description is inconsistent with the appearance of an ixodes tick), the probability of infection is too low to warrant empirical antibiotic therapy.
Once it has been determined that the patient has been bitten by an ixodes tick, the probability that the tick is infected with B. burgdorferi must first be determined on the basis of epizootiologic surveys. The Centers for Disease Control recently reported that I. dammini and I. pacificus are established in only 17 states (Maine, Massachusetts, Connecticut, Rhode Island, New York, New Jersey, Pennsylvania, Delaware, Maryland, Michigan, Illinois, Wisconsin, Minnesota, Iowa, Washington, Oregon, and California). In an additional seven states where these ticks have been found (Vermont, New Hampshire, Virginia, Indiana, Arizona, New Mexico, and Nevada), their presence may be a transient phenomenon caused by their transport (e.g., on birds) from adjacent states. In these states, one must consider the probability that an ixodes tick is infected with B. burgdorferi.
Although hundreds of cases of Lyme disease have been reported from the southeastern United States and from Missouri, where other species of ticks (I. scapularis and Amblyomma americanum) are found, empirical therapy is not recommended at this time in the 26 states where I. dammini and I. pacificus have not been found, since B. burgdorferi has never been definitely identified in ticks, animals, or humans.
Two small studies suggest that spirochete transfer occurs in approximately 10 percent of people bitten by infected ticks.5 , 6 Thus, empirical therapy is not warranted in areas where the prevalence of infected ticks is less than 0.10, because the product of the prevalence of tick infection and the probability of disease transmission from an infected tick is less than 0.01, the probability at which the model indicates that Follow is the preferred strategy. Two recent preliminary reports suggest that the prevalence of infection in ticks that bite may be significantly lower than the general rate of infection in ticks, as assessed on the basis of ticks gathered from vegetation.41 , 42 If these reports are substantiated, the prevalence of tick infection as determined from entomologic surveys will need to be adjusted downward to obtain an estimate of the rate of infection among biting I. dammini ticks. Alternatively, public health authorities could monitor and report rates of infection among biting ticks.
In the six western states where I. pacificus is found, the prevalence of B. burgdorferi infection ranges from 1 to 3 percent,4 , 43 and empirical therapy of tick bites is unwarranted. In areas of New England, the Middle Atlantic states, and the upper Midwest, approximately 0.1 to 1 percent of larvae, 25 percent of nymphs, and 50 percent of adult I. dammini ticks are infected with B. burgdorferi.14 , 44 45 46 47 48 49 Although tick infestations are highly localized, empirical therapy for tick bites should be considered in these areas. Recommendations for empirical therapy will need to be reexamined as better information on tick-infection rates (e.g., on a county-by-county basis) and the risk of infection after tick bites becomes available. Determinations of the efficiency of infection with B. burgdorferi through tick bites and the factors influencing the risk of infection should be major research priorities, given the importance of this information in patient care.
The duration of the tick's attachment may be a critical factor in spirochete transmission. Studies in animals suggest that there is little risk of infection within the first 24 hours, a risk of approximately 50 percent after 48 hours, and almost universal infection after 72 hours of attachment or if an infected tick feeds to engorgement.49 , 50 However, the generalizability of these findings to humans is unknown, and accurate estimation of the likelihood of transmission is difficult because of difficulty in assessing the duration of tick attachment and the degree of tick engorgement.
The detection of B. burgdorferi infection in an early, asymptomatic stage offers the potential to reduce unnecessary antibiotic treatment of uninfected patients and adverse reactions to antibiotic therapy, while minimizing disease sequelae in untreated patients. Unfortunately, currently available serologic tests for B. burgdorferi infection are not sufficiently sensitive. The strategy of treating only patients who have erythema migrans or who have a positive antibody test for Lyme disease one month after tick exposure is reasonable only if the sensitivity of the test exceeds 0.90, and then only for patients in whom the probability of B. burgdorferi infection is between 0.01 and 0.02.
This decision analysis provides guidance to clinicians caring for patients with recent ixodes tick bites in areas of endemic Lyme disease. Our approach offers the advantages of performing analyses applicable to a broad range of probabilities of infection, altering assumptions when better estimates become available or conditions change, and considering both economic and clinical consequences simultaneously. Thus, decision models are both comprehensive and flexible, allowing adjustment for variables that change with populations and time. Larger randomized clinical trials of empirical treatment of patients bitten by ixodes ticks are now in progress and should broaden our understanding of the benefits and drawbacks of prophylactic therapy, primarily by providing better estimates of the probability that Lyme disease will develop after exposure to infected ticks.7 , 8 , 51 Even then, models such as the one presented here will be required in order to evaluate management strategies in settings that differ from those of the clinical trials, and because some sequelae of Lyme disease are too infrequent to be studied adequately in a randomized trial.
Supported in part by grants from the Charles A. Dana Foundation and the John A. Hartford Foundation, grants (AI 26853 and AR 4007) from the National Institutes of Health, and an unrestricted grant to the Leonard Davis Institute from the Merck Foundation. Dr. Magid was a Charles A. Dana Foundation Scholar and Dr. Brian Schwartz an Andrew Mellon Foundation Fellow at the University of Pennsylvania.
Presented in part at the annual meeting of the Society for Medical Decision Making, Minneapolis, October 15–18, 1989, and the National Meeting of the American Federation for Clinical Research, Washington, D.C., May 4–7, 1990.
Source Information
From the Emergency Medical Services, Denver General Hospital, and the Colorado Emergency Medicine Research Center, University of Colorado Health Sciences Center, Denver (D.M.); the Department of Environmental Health Sciences, Division of Occupational Health, Johns Hopkins University School of Hygiene and Public Health, Baltimore (B.S.); the Section of Rheumatology, Department of Medicine, School of Medicine, Yale University, New Haven, Conn. (J.C.); and the Division of General Internal Medicine and the Clinical Epidemiology Unit, Department of Medicine, School of Medicine; the Health Care Systems Department, the Wharton School; and the Leonard Davis Institute of Health Economics; all of the University of Pennsylvania, Philadelphia (J.S.S.). Address reprint requests to Dr. Schwartz at the Leonard Davis Institute of Health Economics, 3641 Locust Walk, Philadelphia, PA 19104.
- References
References
1
Lyme disease — Connecticut . MMWR Morb Mortal Wkly Rep 1988;37:1–3.
2
Steere
Web of Science | Medline3
Levine
Web of Science | Medline4
Burgdorfer
Web of Science | Medline5
Falco
CrossRef | Web of Science6
Costello
CrossRef | Web of Science | Medline7
Agre 8
Shapiro ED, Gerber MA, Persing D, Berg AT, Feder H. Prevention of Lyme disease: a randomized clinical trial of antimicrobial prophylaxis for people bitten by a deer tick. In: Proceedings and abstracts of the Fifth International Conference on Lyme Borreliosis, Washington, D.C., May 30—June 2, 1992. Bethesda, Md.: Federation of American Societies for Experimental Biology, 1992:A47. abstract.
9
Malawista
Medline10
Asbrink
CrossRef | Web of Science | Medline11
Steere
Web of Science | Medline12
Steere
Web of Science | Medline13
Steere
CrossRef | Web of Science | Medline14
Steere
Full Text | Web of Science | Medline15
Treatment of Lyme disease . Med Lett Drugs Ther 1989;31:57–9.
Medline16
Schoen
Medline17
Rahn
Web of Science | Medline18
Hanrahan
CrossRef | Web of Science | Medline19
Lastavica
Full Text | Web of Science | Medline20
Centers for Disease Control 21
Craft
CrossRef | Web of Science | Medline22
Shrestha
CrossRef | Web of Science | Medline23
Russell
CrossRef | Web of Science | Medline24
Magnarelli
Web of Science | Medline25
Magnarelli
CrossRef | Web of Science | Medline26
Berardi
CrossRef | Web of Science | Medline27
Steere
Web of Science | Medline28
McAlister
Web of Science | Medline29
Reik
CrossRef | Web of Science | Medline30
Pachner
Web of Science | Medline31
Clark
Web of Science | Medline32
Steere
Web of Science | Medline33
Neu
CrossRef | Web of Science | Medline34
Liu 35
Cockburn
CrossRef | Web of Science | Medline36
Dattwyler
CrossRef | Web of Science | Medline37
Gelman CR, Rumack BM, eds. Doxycycline drug evaluation monograph. DRUGDEX Information System. Vol. 72. Denver: Micromedex, 1992. (Electronic data file, edition expired May 31. 1992.)
38
Scheife RT, Levy M, Greenblatt DJ. Antimicrobial agents: tetracyclines. In: Miller RR, Greenblatt DR, eds. Drug effect in hospitalized patients: experiences of the Boston Collaborative Drug Surveillance Program 1966–1975. New York: John Wiley, 1976:263–4.
39
Gelman CR. Rumack BM, eds. Ceftriaxone drug evaluation monograph. DRUGDEX Information System. Vol. 72. Denver: Micromedex, 1992. (Electronic data file, edition expired May 31, 1992.)
40
Moskovitz
CrossRef | Web of Science | Medline41
Mather TN, Hu R, Hyland KE, Jones AJ, Mitchell SE. Reduced rates of spirochetal prevalence in host-attached vs. host-seeking deer ticks (Ixodes dammini). In: Proceedings and abstracts of the Fifth International Conference on Lyme Borreliosis, Washington, D.C., May 30—June 2, 1992. Bethesda, Md.: Federation of American Societies for Experimental Biology, 1992:A49. abstract.
42
Lacombe E, Rand PW, Smith RP, Vary CP. Disparity of Borrelia burgdorferi infection of adult Ixodes dammini on deer and vegetation. In: Proceedings and abstracts of the Fifth International Conference on Lyme Borreliosis, Washington, D.C., May 30—June 2, 1992. Bethesda, Md.: Federation of American Societies for Experimental Biology, 1992:A49. abstract.
43
Lane
CrossRef | Web of Science | Medline44
Steere
Full Text | Web of Science | Medline45
Magnarelli
Web of Science | Medline46
Magnarelli
Web of Science | Medline47
Bosler
Web of Science | Medline48
Bowen
Web of Science | Medline49
Burgdorfer
Web of Science | Medline50
Piesman
Web of Science | Medline51
Nadelman RB, Forseter G, Horowitz H, et al. Natural history of patients bitten by Ixodes dammini (Id) in Westchester County, New York, USA. In: Proceedings and abstracts of the Fifth International Conference on Lyme Borreliosis, Washington, D.C., May 30—June 2, 1992. Bethesda, Md.: Federation of American Societies for Experimental Biology, 1992:A80. abstract.
- Citing Articles (41)
Citing Articles
1
Scott Morlando, Stephen J. Schmidt, Kathleen LoGiudice. (2011) Reduction in Lyme Disease Risk as an Economic Benefit of Habitat Restoration. Restoration Ecologyno-no
CrossRef2
R. Kaiser, V. Fingerle. (2009) Neuroborreliose. Der Nervenarzt 80:10, 1239-1251
CrossRef3
Roger P. Clark, Linden T. Hu. (2008) Prevention of Lyme Disease and Other Tick-Borne Infections. Infectious Disease Clinics of North America 22:3, 381-396
CrossRef4
Gary P. Wormser, Raymond J. Dattwyler, Eugene D. Shapiro, John J. Halperin, Allen C. Steere, Mark S. Klempner, Peter J. Krause, Johan S. Bakken, Franc Strle, Gerold Stanek, Linda Bockenstedt, Durland Fish, J. Stephen Dumler, Robert B. Nadelman. (2006) The Clinical Assessment, Treatment, and Prevention of Lyme Disease, Human Granulocytic Anaplasmosis, and Babesiosis: Clinical Practice Guidelines by the Infectious Diseases Society of America. Clinical Infectious Diseases 43:9, 1089-1134
CrossRef5
C. Venkata S. Ram. (2006) Review of resistant hypertension. Current Hypertension Reports 8:5, 398-402
CrossRef6
Linden Hu. (2005) Lyme Arthritis. Infectious Disease Clinics of North America 19:4, 947-961
CrossRef7
Dirk M. Elston. (2005) Rickettsial skin disease: uncommon presentations. Clinics in Dermatology 23:6, 541-544
CrossRef8
G??ran G??nther, Mats Haglund. (2005) Tick-Borne Encephalopathies. CNS Drugs 19:12, 1009-1032
CrossRef9
Z. Hubalek, J. Halouzka, Z. Juricova. (2004) Borreliae in Ixodes ricinus ticks feeding on humans. Medical and Veterinary Entomology 18:3, 228-231
CrossRef10
Joseph Piesman, Marc C. Dolan. (2002) Protection Against Lyme Disease Spirochete Transmission Provided by Prompt Removal of Nymphal Ixodes scapularis (Acari: Ixodidae). Journal of Medical Entomology 39:3, 509-512
CrossRef11
Mary Elizabeth Wilson. (2002) Prevention of tick-borne diseases. Medical Clinics of North America 86:2, 219-238
CrossRef12
(2001) Single-Dose Doxycycline for the Prevention of Lyme Disease. New England Journal of Medicine 345:18, 1348-1350
Full Text13
Shapiro, Eugene D., . (2001) Doxycycline for Tick Bites — Not for Everyone. New England Journal of Medicine 345:2, 133-134
Full Text14
Nadelman, Robert B., Nowakowski, John, Fish, Durland, Falco, Richard C., Freeman, Katherine, McKenna, Donna, Welch, Peter, Marcus, Robert, Agüero-Rosenfeld, Maria E., Dennis, David T., Wormser, Gary P., . (2001) Prophylaxis with Single-Dose Doxycycline for the Prevention of Lyme Disease after an Ixodes scapularis Tick Bite. New England Journal of Medicine 345:2, 79-84
Full Text15
G A Poland. (2001) Prevention of Lyme disease: a review of the evidence.. Mayo Clinic Proceedings 76:7, 713-724
CrossRef16
E. D. Shapiro, M. A. Gerber. (2000) Lyme Disease. Clinical Infectious Diseases 31:2, 533-542
CrossRef17
G. P. Wormser, R. B. Nadelman, R. J. Dattwyler, D. T. Dennis, E. D. Shapiro, A. C. Steere, T. J. Rush, D. W. Rahn, P. K. Coyle, D. H. Persing, D. Fish, B. J. Luft. (2000) Practice Guidelines for the Treatment of Lyme Disease. Clinical Infectious Diseases 31:Supplement 1, S1-S14
CrossRef18
J MELSKI. (2000) Lyme borreliosis. Seminars in Cutaneous Medicine and Surgery 19:1, 10-18
CrossRef19
J HIRSCHMANN. (2000) Antimicrobial prophylaxis in dermatology. Seminars in Cutaneous Medicine and Surgery 19:1, 2-9
CrossRef20
(1998) Perinatal Transmission of Human Granulocytic Ehrlichiosis. New England Journal of Medicine 339:26, 1941-1943
Full Text21
(1998) Immunization against Lyme Disease. New England Journal of Medicine 339:22, 1637-1639
Full Text22
Edith Maes, Pascal Lecomte, Nancy Ray. (1998) A cost-of-illness study of lyme disease in the United States. Clinical Therapeutics 20:5, 993-1008
CrossRef23
DH Walker. (1998) TICK-TRANSMITTED INFECTIOUS DISEASES IN THE UNITED STATES. Annual Review of Public Health 19:1, 237-269
CrossRef24
Leonard H. Sigal. (1998) MUSCULOSKELETAL MANIFESTATIONS OF LYME ARTHRITIS. Rheumatic Disease Clinics of North America 24:2, 323-351
CrossRef25
David T Dennis, Martin I Meltzer. (1997) Antibiotic prophylaxis after tick bites. The Lancet 350:9086, 1191-1192
CrossRef26
Dattwyler, Raymond J., Luft, Benjamin J., Kunkel, Mark J., Finkel, Michael F., Wormser, Gary P., Rush, Thomas J., Grunwaldt, Edgar, Agger, William A., Franklin, Michael, Oswald, DonaldCockey, LouiseMaladorno, Dionigi, . (1997) Ceftriaxone Compared with Doxycycline for the Treatment of Acute Disseminated Lyme Disease. New England Journal of Medicine 337:5, 289-295
Full Text27
Larry I. Lutwick. (1996) POSTEXPOSURE PROPHYLAXIS. Infectious Disease Clinics of North America 10:4, 899-915
CrossRef28
Martina H. Ziska, S. T. Donta, F. C. Demarest. (1996) Physician preferences in the diagnosis and treatment of Lyme disease in the United States. Infection 24:2, 182-186
CrossRef29
E. I. Korenberg, Henrietta G. Moskvitina, Natalia N. Vorobyeva, Lidiya Ya. Gorban'. (1996) Prevention of borreliosis in persons bitten by infected ticks. Infection 24:2, 187-189
CrossRef30
M. A. Cimmino, G. L. Moggiana, Maria Parisi, S. Accardo. (1996) Treatment of Lyme arthritis. Infection 24:1, 91-94
CrossRef31
D.W.J. Dippel, J.D.F. Habbema. (1995) Decision analysis in the clinical neurosciences: a systematic review of the literature. European Journal of Neurology 2:6, 523-539
CrossRef32
Marva M. Moxey-Mims, F. Bruder Stapleton, Leonard G. Feld. (1994) Applying decision analysis to management of adolescent idiopathic nephrotic syndrome. Pediatric Nephrology 8:6, 660-664
CrossRef33
(1994) Treatment of late (stage III) disease. The Lancet 343:8904, 1018-1020
CrossRef34
RobertD Nadelman, John Nowakowski, GaryP Wormser. (1993) Can Lyme borreliosis be prevented after tick bite?. The Lancet 342:8878, 1052
CrossRef35
Spach, David H.Liles, W. ConradCampbell, Grant L.Quick, Robert E.Anderson, Donald E. Jr.Fritsche, Thomas R.. (1993) Tick-Borne Diseases in the United States. New England Journal of Medicine 329:13, 936-947
Full Text36
F-R. Matuschka, A. Spielman. (1993) Risk of infection from and treatment of tick bite. The Lancet 342:8870, 529-530
CrossRef37
(1993) Antimicrobial Prophylaxis after Tick Bites. New England Journal of Medicine 328:19, 1418-1420
Full Text38
(1993) Prevention of Lyme Disease after Tick Bites. New England Journal of Medicine 328:2, 136-139
Full Text39
Peter Wahlberg, Hans Granlund, Dag Nyman, Jaana Panelius, Llkka Seppälä. (1993) Late Lyme Borreliosis: Epidemiology, Diagnosis and Clinical Features. Annals of Medicine 25:4, 349-352
CrossRef40
Shapiro, Eugene D., Gerber, Michael A., Holabird, Nancy B., Berg, Anne T., Feder, Henry M. Jr., Bell, Gillian L., Rys, Paul N., Persing, David H., . (1992) A Controlled Trial of Antimicrobial Prophylaxis for Lyme Disease after Deer-Tick Bites. New England Journal of Medicine 327:25, 1769-1773
Full Text41
Kassirer, Jerome P., . (1992) Is a Tick's Bark Worse Than Its Bite?. New England Journal of Medicine 327:8, 562-563
Full Text
