Discussion

Our findings showed that the implementation of FA/ME panel orders did not significantly affect antibiotic prescribing patterns in patients with suspected community-acquired meningitis. Despite the rapid results of the FA/ME panel compared to CSF culture, there was no statistical difference between the duration of empiric antibiotic therapy or DOT when controlling for confounders. ^{Reference Hanson5}

Previous literature showed mixed findings on the impact of the FA/ME panel on antimicrobial usage. ^{Reference Hueth, Thompson-Leduc and Totev6,Reference Goodlet, Tan, Knutson and Nailor7} Most of the previous literature includes pediatric populations and non-antibiotic antimicrobials such as acyclovir, which may not reflect the utility of the FA/ME panel on antibiotic usage in adult populations. ^{Reference Hueth, Thompson-Leduc and Totev6,Reference Goodlet, Tan, Knutson and Nailor7} Meningitis presents with different symptoms, causative pathogens, and treatment algorithms in pediatric and adult populations, as well as different severity and treatment with bacterial and viral pathogens, limiting comparison between our study and previous research. ^{Reference Tunkel, Hartman and Kaplan8} In the systematic review by Goodlet et al. and meta-analysis by Hueth et al., a key limitation was that no studies with adult populations separated the usage of antibiotics and acyclovir, which leads to a strong confounder in how physicians approach viral compared to bacterial meningitis. ^{Reference Hueth, Thompson-Leduc and Totev6,Reference Goodlet, Tan, Knutson and Nailor7}

We found two other studies that tracked specifically antibiotic usage differentiated from total antimicrobial usage in adult populations. Of these studies, one showed a decrease in antibiotic usage, although another showed no change. ^{Reference Choi, Westblade and Gottesdiener9,Reference Kitagawa, Kitano and Uchihara10} A key difference between the study by Choi et al., which showed a reduction in antibiotic usage, is that clinicians received education regarding the interpretation of the FA/ME panel, which was not present in our study or the study by Kitagawa et al. ^{Reference Choi, Westblade and Gottesdiener9,Reference Kitagawa, Kitano and Uchihara10} This education included institutional emails, conferences, and direct communication of positive results from the Clinical Microbiology Laboratory to the clinician by phone. ^{Reference Choi, Westblade and Gottesdiener9}

Our study adds to the literature, given its focus on empiric antibiotic prescriptions in adults. Our study is also unique in that there was no pre-post design as in most previous studies, possibly limiting the non-contemporaneous control bias in quasi-experimental data. ^{Reference Hueth, Thompson-Leduc and Totev6,Reference Goodlet, Tan, Knutson and Nailor7,Reference Choi, Westblade and Gottesdiener9,Reference Kitagawa, Kitano and Uchihara10} The lack of impact of the panel at our hospital highlights the need for education and prospective antibiotic stewardship efforts when implementing new diagnostic tests.

Limitations of our study include possible selection bias due to the optional nature of the FA/ME panel orders. To address this bias, we controlled for patient baseline characteristics, presenting symptoms, comorbidities, and location that could affect the decision to order the FA/ME panel. We did not manually review all clinician notes, which could have led to confounding bias from patients receiving antibiotics for non-central nervous system (CNS) infections; however, we aimed to limit this bias by excluding all patients with any other positive non-CNS culture results. Finally, the intervention only included one academic health system, limiting generalizability.