Time to Diagnosis and Treatment of Lyme Disease by Patient Race

This cross-sectional study compares stages at which Lyme disease was diagnosed by race in a specialty clinic in the US.

case series of patients meeting a research case definition for post-treatment Lyme disease, with probable or confirmed prior Lyme disease verified by medical record review, and current functionally-impairing symptoms (n=286, 20.5%).These patients were primarily recruited from those seen at the clinic who met eligibility criteria and agreed to participate, and were excluded for the same list of conditions detailed above.Clinical information regarding initial Lyme disease presentation for these patients was abstracted from medical records at the time of initial Lyme disease diagnosis and treatment.Finally, a third source was a clinic chart review of patients seen solely for clinical care of Lyme disease, in which their initial presentation was also abstracted in the same way from their medical records (n=728, 52.2%).
In the current study, the merged sample across these sources was reviewed to ensure that any patients seen both for clinical care and research participation were not duplicated.The final sample therefore included 728 (52.2%) patients seen for clinical care only, and 667 (47.8%) seen as part of a research study.

Data Abstraction
© 2023 Starke SJ et al.JAMA Network Open.
For the 1,014 patients whose initial Lyme disease presentation information was abstracted from their medical record (the second and third sources), a standardized data abstraction instrument with explicitly defined variables was created by two of the authors (JNA and AWR), who also oversaw the abstraction process.The first series of abstractions (n=32, 3.2%) were performed by AWR as part of testing the validity of the instrument and training subsequent abstractors, one of whom performed n=10 (1%) of abstractions and the other performed the remaining n=972 (95.9%) of abstractions across several years.
All abstractors were unaware of the current study's hypothesis at the time of abstraction, and were encouraged to raise questions and concerns as they arose.The resulting data were edited for internal consistency and missing information prior to any analyses being performed.Our data integrity efforts worked to minimize any instances of missing data.These data were collected and managed using REDCap (Research Electronic Data Capture) tools hosted at Johns Hopkins University.

Variables of Interest
The primary variables of interest in the current study included a) race, which was first described by the categories displayed in Figure 1 then treated as a binary variable (Black vs. white) in subsequent analyses that excluded other races, and b) Lyme disease initial presentation, as described in the manuscript as three categories.Age was abstracted as a continuous variable and descriptively summarized in its original one-year units.For an accurate and easier interpretation of the estimated odds ratio in the final adjusted model, age was rescaled to 10-year units by dividing the original continuous age variable by 10.Gender was assessed through self-report, with one patient identifying as neither a man or a woman.Time to appropriate treatment was determined as the interval between the date of the patient's first self-reported new-onset sign or symptom of their Lyme disease and the date of initiation of the first appropriate antibiotic treatment for Lyme disease.Appropriate antibiotic treatment for Lyme disease was considered to be any of the following identified in the medical record as prescribed at the time of initial diagnosis: Doxycycline 100mg BID for at least 10 days, Tetracycline 500mg TID for at least 14 days, Amoxicillin 500mg TID for at least 14 days, Augmentin 875mg BID for at least 14 days, Ceftin 500mg BID for at least 14 days, and Ceftriaxone 2g Q24 for at least 14 days.