Assessment of Physician Knowledge, Attitudes, and Practice for Lyme Disease in a Low-Incidence State

Abstract

Lyme disease (LD), caused by the bacterium Borrelia burgdorferi, is transmitted to humans in California through the bite of infected blacklegged ticks (Ixodes pacificus). Overall, the incidence of LD in California is low: approximately 0.2 confirmed cases per 100,000 population. However, California’s unique ecological diversity results in wide variation in local risk, including regions with local foci at elevated risk of human disease. The diagnosis of LD can be challenging in California because the prior probability of infection for individual patients is generally low. Combined with nonspecific symptoms and complicated laboratory testing, California physicians need a high level of awareness of LD in California to recognize and diagnose LD efficiently. This research addresses an under-studied area of physicians’ knowledge and practice of the testing and treatment of LD in a low-incidence state. We assessed knowledge and practices related to LD diagnosis using an electronic survey distributed to physicians practicing in California through mixed sampling methods. Overall, responding physicians in California had a general awareness of Lyme disease and were knowledgeable regarding diagnosis and treatment. However, we found that physicians in California could benefit from further education to improve test-ordering practices, test interpretation, and awareness of California’s disease ecology with elevated levels of focal endemicity, to improve recognition, diagnosis, and treatment of LD in California patients.

Lyme disease, Borrelia burgdorferi, physician knowledge, incidence, physician attitude

Lyme disease (LD), caused by the bacterium Borrelia burgdorferi, is transmitted to humans in California by western blacklegged ticks (Ixodes pacificus [Cooley & Kohls]) (Burgdorfer et al. 1982, Clover and Lane 1995, Barbour and Benach 2019). Lyme disease has been a reportable disease in California since 1989 and became a nationally notifiable disease in the United States in 1991 (Werra 1991, Center for Disease Control and Prevention 2019). On 1 January 2022 the Council of State and Territorial Epidemiologists (CSTE) modified the LD surveillance case definition to improve specificity of reported cases in low-incidence states (CDC 2021a). Though LD is the most common tick-borne disease in the United States with over 30,000 cases reported annually (Center for Disease Control and Prevention 2019, Kugeler et al. 2021, Schwartz et al. 2021), the incidence of LD in California is low, with approximately 100 confirmed cases reported annually (0.2 cases per 100,000 population) (California Department of Public Health 2019). California’s unique ecological diversity contributes to focal high-endemic regions, where human incidence ranges from 1.1 to 6.2 cases per 100,000 (Lane et al. 1992, Ley et al. 1994, Eisen et al. 2006a, Eisen et al. 2006c). B. burgdorferi sensu stricto has been recovered from western blacklegged ticks from at least 44 of 58 CA counties (CDPH 2022).

The diagnosis of LD requires understanding of the clinical manifestations, laboratory results, and patient exposure (Hirsch et al. 2018, Van Hout 2018, Center for Disease Control and Prevention 2019). Recommended laboratory serologic testing for LD is the conventional two-tiered testing (STTT) sensitive first-tier enzyme immunoassay (EIA) followed by a more specific western blot test, either IgM within first 30 d post onset or IgG thereafter) or the modified two-tiered testing (MTTT) which involves the sequential use of two EIAs (Lipsett et al. 2019, Lantos et al. 2021, Branda and Steere 2021). Effective diagnostic approaches depend on a healthcare provider’s knowledge and awareness of LD (Henry et al. 2012, Schoen 2020).

The challenge for physicians is determining appropriate testing when a patient has symptoms compatible with LD to minimize potential for false positive or false negative test results (Aguero-Rosenfeld et al. 1996, Steere et al. 2008). These testing complexities, coupled with low overall exposure risk in a low-incidence state, can make diagnosis of LD more complicated. This research addresses an under-examined area of physician knowledge, testing, and treatment of LD in California, a low-incidence state. Physician assessment in areas of the United States where the incidence of LD is high (19.7 – 106.6 per 100,000) suggests physicians generally follow published guidelines for diagnosis and treatment of symptomatic LD (Lantos et al. 2021), but deviate from guidelines for serologic testing and management of asymptomatic tick bites (Fix et al. 1998, Ramsey et al. 2004, Henry et al. 2012). It is unknown if the findings apply to low-incidence areas. We surveyed physicians in California to understand their current knowledge, testing, and treating practices for LD. The results of this study could help identify opportunities for public-health education about diagnosis and treatment of LD in low-incidence states.

Materials and Methods

Physicians practicing across 16 counties (representing higher and lower endemic counties for LD in California) were selected using a cross sectional approach. An electronic questionnaire was developed to capture LD knowledge, testing, and treating practices of physicians. Survey distribution was accomplished through multiple modalities. A medical marketing distribution company (MMS Distribution) was contracted to distribute the survey to physicians throughout California. The study was open to all physicians licensed in California with specialties or subspecialties in internal medicine, family practice, pediatrics, infectious disease, rheumatology, and neurology, namely those sub-specialties most likely to have a diagnostic encounter with a LD patient. The MMS healthcare provider database includes California-licensed physicians collected by the American Medical Association through the General Medical Education (GME) census. The physician survey by MMS inc. was launched in March 2020, followed by three reminders, in April, July, and August 2021. To increase participant response rates, additional surveys were directly emailed to physicians at universities and practice groups with large catchment areas of northern California including the University of California, Davis Medical Center, the University of California, San Francisco Medical Center, Stanford Medical School, Palo Alto Medical Foundation and Dominican Hospital. This study was approved by the Institutional Review Board (IRB) of the University of California Davis, protocol # 1388609-2.

Questionnaire

The survey design was adopted from similar studies from high-incidence states in the northeastern United States (Magri et al. 2002, Conant et al. 2018), with modifications to address testing and treatment challenges of practicing in a low-incidence state like California. Study data were collected and managed using REDCap electronic data capture tools hosted at the University of California, Davis (Harris et al. 2009, Harris et al. 2019). The modified survey consisted of questions on characteristics of each physician’s practice (years of practice, medical specialty, practice setting, and county), the physician’s knowledge about LD such as understanding of the tick vector and its role in the medical diagnosis of LD, serologic testing of LD, patient scenario questions addressing laboratory diagnostics (testing and interpretation of test results) and treatment of LD such as antibiotic treatment choices for LD following a tick-bite in a low incidence state. The survey was designed to take 5–10 min. The survey was reviewed and piloted by six practicing California physicians before deployment. The pilot physicians were not included in the final analysis.

Sample Size

Our study was designed before the global pandemic caused by SARS-CoV-2, but implementation occurred during the pandemic. In total, 3,488 surveys were distributed. A lower-than-expected survey response was obtained, motivating the follow-up with targeted sampling in which physicians were actively contacted within large medical providers in northern California as described above.

Statistical Analysis Plan

We obtained frequencies on physician practice characteristic variables such as setting, specialty, and years of practice, as well as responses to LD practice questions and patient scenario questions. Ten questions were scored as correct or incorrect to test physician knowledge of testing and treatment for LD. A total score for each physician involved calculating a percentage of responses that matched predefined correct answers for each question. All predefined correct response was based on the IDSA guidelines for the testing, and treatment of Lyme disease (Lantos et al. 2021). Exact 95% confidence intervals were constructed for the percentage of correct responses. All analyses were performed using SAS software version 9.4 (SAS Institute Inc., Cary, NC, USA).

Results

Physician Characteristics

Sixty-four physicians from 16 California counties responded to the survey. Of these, two physicians provided incomplete responses; thus, a total of 62 responses (23 responses from MMS and 39 responses from direct outreach) were included in the final analysis. Thirty-four (54.8%) practiced in internal medicine, family practice, and pediatrics and 28 (45.2%) practiced in infectious disease, rheumatology, and neurology. All participating physicians had Doctor of Medicine (MD) degrees. Twenty-seven (43.6%) worked in the outpatient setting, 10 (16.1%) worked in the hospital setting, and 22 (35.5%) worked in both a hospital and outpatient setting. On average, physicians had 20.6 years of practice, ranging from 1 to 50 years (Table 1).

Table 1.

Physician characteristics

Physician characteristics	Physician respondents (N = 62)
Specialty
Internal Med/Family Practice/Pediatrics	34 (54.8%)
Infectious Disease Rheumatology/Neurology	28 (45.2%)
Setting
Outpatient	27 (43.6%)
Hospital	10 (16.1%)
Hospital/Outpatient	22 (35.5%)
Other	3 (4.8%)
Years of Practice
Mean (SD)	20.6 years (12.7)
Range	1 year–50 years

Lyme Disease Practice

In California counties where LD is endemic, 18 (29.0%) physicians practicing in those counties reported that LD was not endemic, and 2 (3.2%) reported they did not know. Forty-nine (79.0%) reported they did not see an increase of LD patients in their practice in the past 12 months. Forty-two (67.7%) physicians reported patients have asked for LD treatment although the physician did not think their symptoms were caused by LD (Fig. 1).

Fig. 1.

Lyme disease belief and practice questions. The percentage of responses to questions regarding a physician’s belief and practice is shown. Question 1 (belief): Within your geographic area of practice, would you consider Lyme disease (LD) endemic? Question 2 (practice): Have the number of LD cases increased among patients in your practice? And Question 3 (practice): Have patients asked to be treated for LD though LD was the unlikely cause of their symptoms?

Physician Survey (Scored Questions)

The range of scores was between 20 and 100%. Four (6.5%) of the 62 physicians surveyed answered all 10 questions (100%) correctly. Overall, the mean score of physician’s knowledge was 71.9% (95% CI: 67.3–76.6%) and the median was 75.0% (Fig. 2).

Fig. 2.

Percentage correct of scored Lyme disease knowledge questions. The distribution of percentage correct across the scored questions on LD in California is portrayed with each individual point representing the score of a single physician. The top and bottom of the box represents the 75th and 25th percentiles of the distribution, respectively. The line through the box represents the median score. The top whisker extends to the highest score that falls within the distance of 1.5 times the interquartile range (IQR) = (Q3 – Q1) of the top of the box and the bottom whisker extends to the lowest score that falls within the distance of 1.5 times the IQR from the bottom of the box. Any points outside the whiskers are considered potential outliers.

Tick Vector and LD Diagnosis

Forty-seven (75.0%, 95% CI: 63.3–85.8%) responded knowing the species of a tick recovered from a patient would be helpful in diagnosing LD (Lantos et al. 2021). However, 64.5% (95% CI: 51.3–76.3%) responded that if a tick recovered from a patient with symptoms tested positive for B. burgdorferi, the positive tick result would inform their medical decision-making about LD (Table 2).

Table 2.

Patient scenario knowledge questions and responses

Interpretation of test results	Responses	Physicians respondents (N = 62)
(1) A healthy patient with a history of daily hiking in the month of April and in an area where ticks are found, presents in your office with a rash resembling an erythema migrans that began 3 d earlier. You order a serologic test for Lyme disease which yields a negative result. Would you consider this negative test result definitive to rule out Lyme disease as the cause of this patient’s rash?	Yes No Maybe	3 (4.8%) 57 (91.9%) 2 (3.2%)
(2) A 45-year-old patient from Southern California presents with fatigue and difficulty concentrating for the past two years. The patient does not remember a tick bite or rash but occasionally gardens in the backyard. The patient has not traveled out of Southern California for the past two years. A Lyme disease test was ordered at the time of the visit and the results were: Equivocal EIA, positive IgM Western blot (2/3 bands), negative IgG Western blot (1/10 bands). What is your interpretation of these results?	Unlikely to have LD Likely to have LD Other	47 (75.8%) 12 (19.3%) 3 (4.8%)
(3) How would you interpret this test result from a patient you tested for Lyme disease: Negative EIA, positive IgM western blot, and negative IgG western blot?	Unlikely to have LD Likely to have LD Other	38 (61.3%) 15 (24.2%) 9 (14.5%)
Treatment of Lyme disease
(1) A 35-year-old patient was diagnosed (based upon positive serology and compatible clinical symptoms) and treated for Lyme disease. Are additional serologic tests for Lyme disease warranted after treatment?	Yes No Not Sure	2 (3.2%) 53 (85.5%) 7 (11.3%)
(2) A patient presents to your clinic concerned with a tick bite received about 30 d ago. The patient has not traveled outside of CA, the patient has no symptoms, no laboratory testing performed to date, and normal examination findings. Which of the following describes your next action?	Treat for LD Treat tick bite prophylactically. Do not treat LD Other	0 (0.0%) 2 (3.2%) 59 (95.2%) 1 (1.6%)
(3) A patient presents with recurrent, asymmetric arthritis that began 3 months prior, involving large, weight-bearing joints. The patient has no history of an erythema migrans rash and has had multiple negative Western (IgM/IgG) blot test results for Lyme disease over the past 3 months. The patient does not recall a tick bite, but the patient spends a lot of time outdoors. Which of the following describes your next action?	Treat for LD Do not treat LD Other	2 (3.2%) 57 (91.9%) 3 (4.8%)
Vector and diagnosis
(1) If you submitted a tick recovered from a patient for identification, would knowing the tick species inform your medical decision-making about Lyme disease?	Yes No Not sure	47 (75.8%) 10 (16.1%) 5 (8.1%)
(2) If you submitted a tick recovered from a patient to be tested for Borrelia burgdorferi, would the tick testing result inform your medical decision-making about Lyme disease?	Yes No Not sure	40 (64.5%) 11 (17.7%) 11 (17.7%)
Serologic Diagnostic Testing
(1) What Lyme disease diagnostic tests do you commonly order for a suspected Lyme disease patient?a	WB IgG EIA/IFA/ELISA PCR (Blood tissue) Culture CD57 WB IgM PCR (Synovial fluid) Plasmid Other	42 (67.7%) 44 (71.0%) 9 (14.5%) 0 (0.0%) 1 (1.6%) 36 (58.1%) 0 (0.0%) 1 (1.6%) 10 (16.1%)
(2) A 50-year-old patient from Northwest California presents with a swollen, erythematous knee for the past week. The patient does not remember a tick bite or rash but is active outdoors and went on a hiking trip to the coastal foothills two months ago. You suspect Lyme disease. Which of the following testing approaches would yield the most diagnostic information?	No testing needed Order EIA only Order WB only Two-tier testing PCR joint fluid Other	4 (6.5%) 0 (0.0%) 2 (3.2%) 41 (66.1%) 11 (17.7%) 4 (6.5%)

The frequency and percentage for each response to each question are reported. Bolded responses are the correct responses.

LD = Lyme disease.

^aMore than one correct response (correct response is EIA and IgG or EIA and IgM – not mutually exclusive responses).

Descriptive characteristics of the practice and experience of the physicians responding to the survey.

Diagnostic Testing and Interpretation of Lyme Disease Test Results

Approximately 60% of respondents correctly chose laboratory serologic tests that satisfied the recommendation of the two-tiered testing approach for the diagnosis of LD. Fifty-seven (91.9%, 95% CI: 82.2–97.3%) responded correctly in interpreting diagnostic test result in the presence of an EM. Forty-seven (75.8%, 95% CI: 63.3–85.8%) correctly responded a patient from a non-endemic area with ongoing symptoms greater than two years was unlikely to have LD. Thirty-eight (61.3%, 95% CI: 48.1–73.4%) correctly responded that the patient was unlikely to have LD in the presence of a positive IgM western blot only (Table 2).

Treatment of Lyme Disease

Fifty-three (85.5%, 95% CI: 74.2–93.1%) physicians correctly responded that further testing was not appropriate in a diagnosed and treated patient. Fifty-nine (95.2%, 95% CI: 86.5–99.0%) and 57 (91.9%, 95% CI: 82.2–97.3%) correctly responded that physicians should not treat a tick bite prophylactically (question 2) and physicians should not treat long term arthritis with a negative test for LD (question 3) (Table 2).

Discussion and Conclusion

The purpose of this study was to characterize physician knowledge and practice of LD in California, as a low-incidence United States. We used surveys developed for high-incidence states as a foundation for our questionnaire. The survey results demonstrated that physicians in California could benefit from targeted education to improve test-ordering practices and test result interpretation. Additionally, increased education on LD geographic and seasonal patterns, as well as on characteristics of the diverse disease ecology across California, could be beneficial to physicians for patient evaluation. For example, some physicians did not think LD was endemic in their county of practice, despite documented presence of infected ticks. Overall, California has a low incidence of LD but has focal areas of higher endemicity which may pose a higher risk of human infection for residents living or recreating in those areas (Eisen et al. 2006a, b, c). Insufficient knowledge on where infected ticks are found in California can be problematic in the overall care of patients, affecting testing, diagnosis, and treatment of LD.

The diagnosis of LD in patients with compatible clinical symptoms and recent travel to high-incidence states is relatively straightforward (Forrester et al. 2015). However, patients with a locally acquired tick bite from a low incidence area may pose a diagnostic challenge complicated by a decreased positive predictive value in the serologic testing in a low pretest probability setting (Tugwell et al. 1997).

We found physicians in this study deviated from national guidelines in diagnostic testing for LD when patients sought care for symptomatic disease and asymptomatic tick bites, whereas physicians from high incidence states were more likely to deviate from diagnostic testing guidelines in patients with asymptomatic tick bites only (Ramsey et al. 2004, Lantos et al. 2021). Serologic testing for LD in a low incidence state performs well when appropriate information is assessed, such as obtaining complete medical, travel, and exposure history of the patient. Over-testing and potential overdiagnosis are common in the United States and California (Ley et al. 1994, Webber et al. 2019), and maybe exacerbated by patients' request to be treated although LD is unlikely to be the cause of their symptoms as most physicians indicated (Ley et al. 1994, Webber et al. 2019).

The interpretation of LD antibody testing is a critical educational opportunity for Californian physicians. Some physicians in this study incorrectly interpreted false-positive IgM in patients with longstanding symptoms, which could suggest lack of knowledge. Studies showed that a positive IgM western blot in the presence of a negative screening test and IgG western blot accounted for more than 50% of all false positive results (Seriburi et al. 2012, Webber et al. 2019). In a low incidence state for LD, the pretest probability for LD testing is low, increasing the false positive rate. The false positive rate of an IgM western blot (>30 d of infection) may exceed true positive rates (Sivak et al. 1996, Fritz and Vugia 2001, Seriburi et al. 2012, Moore et al. 2016, Waddell et al. 2016, Webber et al. 2019).

Post tick-bite antibiotic prophylaxis (single 200-mg dose of doxycycline) in low incidence states is not recommended as risk of infection after a tick bite is low versus the risk of antibiotic side effects (Perea et al. 2015). A survey by the CDC found 48% of physicians practicing in a low-incidence state prescribed post tick bite antibiotics (Perea et al. 2015), while most physicians in our study responded they would not treat patients prophylactically to prevent LD.

This study has several limitations. This research was performed during a global pandemic (SARS-CoV-2) resulting in a low response rate. Physicians are challenging to survey, but amidst a global pandemic, was especially problematic (Noel and Huang 2019). Other approaches to increase response rates among physicians are incentive based and mixed-mode surveys (web-based and in person interviews) (VanGeest et al. 2007). Another limitation is whether physicians who responded to the survey systematically differed from physicians who did not respond to the survey. Another limitation is the data are self-reported and may not actually reflect a physician’s actual practice in a clinical setting.

As new LD tests become available (Lipsett et al. 2019) and with an update in the LD surveillance case definition (CDC 2021a), our study suggests that physicians could benefit from specific education to better understand LD risk in California and to improve recognition of symptoms with appropriate use and interpretation of serologic testing. Future knowledge attitude and practice studies are needed on a larger scale in other low-incidence states for LD to clarify LD medical approaches as two preventatives are on the horizon for LD (VanGeest et al. 2007, Comstedt et al. 2017).

Disclaimer: The findings and conclusions in this article are those of the author(s) and do not necessarily represent the views or opinions of the California Department of Public Health or the California Health and Human Services Agency.

Acknowledgments

We would like to thank Thomas Deetz, MD, Deborah Weschler, RN, MPH, Alison Hinckley, PhD and Kiersten Kugler, PhD for their help with survey questions and manuscript review. We would also like to acknowledge those individuals who helped with the distribution of the survey during a global pandemic: Patricia Conrad, DVM, PhD, Stuart Cohen, MD, Jake Scott, MD, Harry and Sue Fujita. We would like to acknowledge the physicians who provided their expertise in developing and piloting the questionnaire. We are grateful to all the physicians who responded to this survey during a global pandemic and a BIG thank you to all our front-line heroes. SIB, WAS, and CMB acknowledge funding support from the Pacific Southwest Regional Center of Excellence for Vector-Borne Diseases funded by the U.S. Centers for Disease Control and Prevention (Cooperative Agreement 1U01CK000516). This study was approved by the Institutional Review Board (IRB) of the University of California Davis, protocol # 1388609-2.

Author notes

Sharon I. Brummitt. Current address: School of Veterinary Medicine, One Shields Avenue, University of California Davis, Davis, CA 95616, USA