Current state and practice variation in the use of Meningitis/Encephalitis (ME) FilmArray panel in children

This study provides an overview of the ME panel’s current use among pediatric providers in the U.S and highlights the variation in its use and interpretation. 68% of respondents in our survey use the ME panel, but only 25% are aware of ordering guidelines, and 26% have institutional guidelines directing its interpretation. A majority of participants still reported using single organism PCR in addition to the ME panel.

When a new diagnostic test is being introduced, it is imperative that there be clear guidelines on test indications/interpretation, evaluation of the cost-benefit ratio, and an ongoing assessment of the risks and benefits of the test. With the easy availability of the CSF ME panel, overutilization has been noted, with the test being performed in patients with little or no suspicion of CNS infection [13‐15]. Furthermore, use of a single organism PCR in addition to the ME panel may add to healthcare costs without providing additional diagnostic information. Although, there may be times where this is necessary, such as HSV PCR testing when this diagnosis is highly suspected, but the ME panel is negative. Test overuse causes unnecessary costs and may also lead to incorrect diagnosis and inappropriate treatment. Though less than half of the providers in our study report being aware of test performance characteristics, the majority reported confidence in interpreting the results. This may further point to overconfidence bias, a known cognitive bias associated with diagnostic inaccuracies and suboptimal management [16]. Increasing providers’ knowledge and providing tools like practice guidelines, diagnostic stewardship, and clinical algorithms are paramount in successfully implementing new testing method. Pairing antimicrobial stewardship with novel diagnostic tools is well-studied in the literature, including blood culture and MEP interpretation. In both instances, introduction of test use guidelines, audit and feedback, and interpretation guidance was associated with decreased antimicrobial use, ancillary testing, and increased antibiotic de-escalation [17, 18]. In the ME panel’s case, these tools ensure maximum clinical benefit (avoid unnecessary antimicrobials, reduce the length of stay, and avoid unnecessary diagnostic tests) without test overuse, inappropriate treatment, and overburdening the health care economy.

In our survey, providers were more likely to start antibiotics in the presence of a positive bacterial ME panel test; however, they hesitated to stop antibiotics with a negative bacterial ME test or with the detection of a viral pathogen. Similarly, less than half of the respondents reported confidence in narrowing the antibiotics based on ME panel results. Potential reasons could include a lack of knowledge and understanding of multiplex PCR by providers and/or concern for false negatives. The ME panel has a sensitivity of 90% and specificity of 98%; however, a small number of false positives and false negatives are reported with the test [19]. These findings suggest that providers’ hesitancy to discontinue antibiotics may in part be due to a lack of knowledge about ME Panel test performance. However, other factors may certainly influence these decisions, including provider experience, pre-test probability, or suspicion for organisms not detected on the ME panel given the clinical scenario. The hesitation in discontinuing antibiotics is in concordance with other studies where physicians start antibiotics more often with a negative viral test but only occasionally stop antibiotics in the presence of viral pathogen [20]. A prior study found that implementation of interpretation guidance for enterovirus testing results was associated with decreased antibiotic use in febrile infants found to have enterovirus meningitis [21]. Similarly, our study results indicate that there is likely an opportunity to provide interpretation guidance and information about ME Panel test performance characteristics to help inform clinical decision making. Inclusion of such tools with diagnostic and antimicrobial stewardship may help achieve desired clinical outcomes. As such, there is a clear role for clinical microbiology laboratory and antimicrobial stewardship teams to work together to develop test use and interpretation guidance to avoid unnecessary antibiotic use. These clinical guidelines or pathways as part of an antimicrobial stewardship team may help provide support to clinicians, especially with patient populations where management can be less straightforward, as is frequently encountered with infants < 2 months old and patients with positive viral targets of unclear significance, such as HHV-6. Studies evaluating such stewardship programs (including use of real-time decision support) indicate potentially positive effects on length of empiric antibiotic and ancillary test utilization [22, 23].

In our study, PHM providers were more aware of the test characteristics compared to PEM providers. This could be attributed to the fact that continuing, narrowing, and stopping antimicrobials falls under the hospitalist realm of practice; thus, they are required to be more knowledgeable. The context in which these two subspecialists encounter a suspected ME patient is also different. Hospitalists have the benefit of observing patients over the clinical course, whereas a PEM provider has the advantage of assessing patients firsthand. Though our study did not find a difference in clinical management between these two subspecialties, further analysis inclusive of subspecialties like ID and PICU is needed to fully elucidate variability. The inclusion of all providers from various specialties involved in the continuum of care is essential while developing clinical tools.

Our study has several limitations. We estimated the number of emails subscribed to the listserves; however, the exact denominator within each listserve changes frequently. Therefore, this might not be an accurate representation of all providers. Our study also has a lower response rate of 4.8% than the average listserve response rate of 8–20% [24‐26]. We could not account for potential differences in non-responders to know if our sample represents physicians at large. Furthermore, multiple providers may have responded from a single institution, which may over- or under-represent certain institutional practice. However, we believe this finding highlights intra-institutional practice variation and an opportunity for practice standardization. The high percentage of respondents who reported using the ME panel in our study also raises concern for bias, as people who use the ME panel were more likely to respond to the survey. We only used two specific clinical scenarios on the survey questions with limited information and we did not allow for a free-text response. Thus, our results may not fully reflect what providers would do in real-life clinical scenarios where more clinical and supporting laboratory evidence would be available. The survey was not validated beyond pilot testing with PHM providers, limiting its applicability to PEM and PICU providers. Our clinical vignettes did not specify which bacterial and viral targets were identified, which may have contributed to response variability. Lastly, most of the respondents in our survey were PHM and PEM physicians with limited response from PICU physicians, making the results less generalizable to intensivists. In general, PHM physicians were overrepresented in our sample with limited data from advanced practice providers, which limits generalizability of our study. Furthermore, very few ID physicians responded, and thus our results may not represent practice standards of this subspecialty. However, ID physicians frequently play an important role in clinical decision making in these patients. Despite these limitations, our study provides a good starting point for future research on guideline development for the ME panel and identifies areas for provider education. Comparative analysis of institutions and subspecialty based on the presence or absence of rigorously developed evidence-based guidelines focusing on clinical outcomes and missed diagnosis may provide an answer to this in the future.