NEW YORK (360Dx) – Researchers at a Mayo Clinic campus in Phoenix have found that the knowing the test results from BioFire's rapid multiplex diagnostic panel for detecting bloodstream infections does not actually alter certain aspects of clinician prescribing behavior, contradicting previous published reports.
Prior studies have established that antimicrobial stewardship interventions are required in combination with the BioFire FilmArray BCID test in order to move the needle on factors such as the time to first appropriate antimicrobial prescribing — or escalation — and the discontinuation of inappropriate treatment after a pathogen is identified, otherwise known as de-escalation.
Authors of the new research, however, suggested that even with standard stewardship interventions, multiplex molecular panel testing did not affect antimicrobial prescribing, although results of other kinds of testing — namely, gram staining and antimicrobial susceptibility testing — were able to induce physicians to alter course appropriately.
As described in Open Forum Infectious Diseases this month, the work separated the results of gram-positive and gram-negative bacteria. The lab performed gram staining immediately on a positive blood culture, with the FilmArray BCID results provided about two hours later. The stewardship effort during the study involved results of the gram staining and BCID being sent to an antimicrobial stewardship team, who, in turn, checked up on inappropriate therapies between the hours of 8AM and 5PM. They noted, for example, whether docs were continuing the antimicrobial drug vancomycin in a patient confirmed to have a gram-negative infection, and guided clinicians to stop ineffective therapies.
Consistent with other research, the BCID panel reduced the time to identify specific infectious organisms compared to standard techniques, like culture.
But how those ID results are then applied by doctors is a key question. Although the OFID study found trends toward improved prescribing given BCID results for gram-positive organisms, overall there were no statistically significant changes in time to escalation or de-escalation of antimicrobials due to the use of the BCID test, according to Thomas Grys, lead author on the study and the director of microbiology at the Mayo Clinic in Arizona.
This partially contradicts other work, including studies done previously by the Mayo campus at Rochester, Minnesota. Grys and his colleagues in Phoenix, however, had a different patient population that included more elderly people, more cancer and transplant patients, and that did not include pediatric patients, he said.
The study was part of a broader set of research at Mayo's Phoenix campus asking whether rapid diagnostic testing, including things like respiratory panels and procalcitonin testing, are impacting prescribing in the hospital. "I can tell you that without stewardship, the answer is generally no, not in the way we would want — and even with stewardship, sometimes it is difficult," Grys said.
In the current study, Grys said the group discovered that gram stains helped escalate therapy appropriately, and also, for gram-negatives, it wasn't until physicians were provided with susceptibility data that they made decisions to de-escalate therapy.
Grys said he believes the reluctance to alter course with BCID testing results may be in part a function of the limited susceptibility testing on the panel. For example, it only detects one beta-lactamase resistance-conferring mutation in the KPC gene, a type of resistance that is rare in Grys' region of the US, he said.
"The concern on the part of the physicians is that, with gram-negatives, they could have multidrug-resistant organisms that we were unable to detect with this method, and so they felt uncomfortable de-escalating until they had phenotypic susceptibilities back," Grys said.
Reports of rapid diagnostic testing failing to impact prescribing seems to be on the increase. Grys noted that, despite best efforts and training, people often make irrational decisions based on anecdotes and personal experience rather than data. Grys said his hospital had a pneumonia case with a low procalcitonin result, suggesting the infection was not bacterial, and the lab had also confirmed a virus was present using a multiplex respiratory panel. Yet, in an interview with the stewardship team, the physician said it was hard to de-escalate antimicrobials because the patient was not doing well, Grys said.
"Especially in the ICU, if the patient is very ill, they might be on four antimicrobial drugs. … Once the pathology results begin to emerge, prying those antibiotics out of the physician's hands is difficult, because they say, 'Well the patient is improving, we shouldn't change anything.'" Unfortunately, and particularly in elderly populations subject to polypharmacy, the antimicrobial drugs themselves can contribute to other risks, Grys noted, both for the patients and for the community.
Beth Lingenfelter, senior director of outcome research at BioFire, noted that the Mayo Phoenix study generally supports the firm's position that optimal deployment of the molecular panel requires on-site support of antimicrobial stewardship teams.
However, the results were "a bit surprising," Lingenfelter noted in an email, as they were not in line with "several high-quality publications that have shown that implementation of the BioFire FilmArray BCID Panel has a significant impact on patient management, and especially on antibiotic escalation and de-escalation."
Specifically, a Journal of Clinical Microbiology study looked at effects of deploying the BCID panel, both with and without intervention from an antimicrobial stewardship program. That study confirmed that the rapid multiplex panel reduced the time to identify an organism as compared to standard methods. And, among the patient population in that study – namely, hospitalized adult patients at a 700-bed academic medical center in Charleston, South Carolina – the time to effective therapy was also reduced, as was the time to first antibiotic de-escalation. As noted in an accompanying commentary to the JCM study, while the findings suggest that the BCID test played a more important role than stewardship alone in reducing the time to effective antibiotic therapy, the importance of the stewardship program was revealed in the antibiotic de-escalation results.
Researchers at the Mayo Clinic in Rochester conducted a prospective randomized study, published in Clinical Infectious Diseases, which compared standard testing to BCID with results reported to the electronic medical record, and to BCID with results incorporated into real-time antimicrobial stewardship support.
The Rochester research showed the BCID results reduced the amount of time a patient was on broad-spectrum antibiotics, regardless of how the results were delivered to clinicians. This research also looked at how gram staining impacted escalation, and found that adding BCID to that routine technique shortened the amount of time until a patient was put on appropriate therapy. In terms of de-escalation, the study found that only BCID combined with real-time stewardship support was able to induce clinicians to take a patient off unnecessary antibiotics.
The authors of the Rochester study reasoned that "notification from a stewardship team member might prompt a busy clinician to act on a test result more so than a telephone call from a laboratory technologist," they wrote, and, that in complex clinical scenarios involving critically ill or immunocompromised patients, or patients with polymicrobial infections on multiple broad-spectrum antibiotics, "providers may prefer discussion with infectious disease specialists prior to modifying antimicrobial management."
Grys pointed out that the population his hospital sees in Phoenix is quite different, however, including more immunocompromised individuals. "That is a higher-risk patient, so often our patients are empirically put on broader-spectrum [antimicrobials], and although we then have more opportunities to de-escalate, we're not going to until we know we have the right drugs remaining on board," Grys said.
Lingenfelter also highlighted a recent study of the FilmArray gastrointestinal panel that suggested physicians can become more efficient with the use of new diagnostic information as they gain experience. "Careful attention to test implementation, effective reporting of results, physician education, and physician comfort with the new information, are important to optimizing the ability of a diagnostic test to improve patient management," she said.
The Mayo Phoenix group now plans to continue using the panel, and to increase its stewardship efforts using the published study as a baseline, Grys said. Additional stewardship could involve extending the hours that the stewardship team operates. The Mayo system overall has now standardized its electronic health system technology across all sites to the Epic system, he said, and this affords new opportunities to incorporate electronic advisories and comments that pop up during results reporting to guide antimicrobial selection.
The best implementation technique, however, remains to be determined. For example, Grys said using the electronic health record to tell a physician that the patient is being switched to a particular drug unless he or she changes the order may make it is easier for the physician to say OK than if the lab simply reports out the raw results, because in the latter case, if the patient happens to be doing better, the clinician may delay deciding.
"If we're not going to make the decision rapidly, there's no use in providing the result rapidly," Grys said.
The gram-negative bacteria remain a bit of a hurdle in Phoenix. On the BCID panel, the susceptibility genes carried by gram-positive bacteria influence "relatively binary" decisions about antimicrobials for bacteria like Staphylococcus aureus or Enterococcus, Grys said. But, gram-negative bacteria have a broader number of species that can carry antimicrobial susceptibility genes, and there are also a larger number of those genes. "Unless you're looking for all possible genes or doing the phenotypic testing … you're not going to know it's there, and because these organisms can be hard to treat if you go too long without being on effective therapy, getting behind the eight ball would make it difficult to recover," Grys said.
Newer panels with more markers of gram-negative resistance — for example panels from T2 Biosystems, and Luminex's Verigene — might help with this issue, he said. Grys plans to evaluate the GenMark Diagnostics instrument and that firm's respiratory and BCID testing in the coming months. The GenMark BCID currently comprises separate multiplex panels for gram-positive, and fungal pathogens which were both cleared in the past week by the US Food and Drug Administration, as well as a gram-negative panel.
The Mayo Phoenix lab has also considered bringing on rapid phenotypic testing from Accelerate Diagnostics, but Grys noted that the cartridges require refrigerator space that the group doesn't currently have. However, the Phoenix Mayo campus is undergoing a $650 million expansion and will soon have space to bring in new testing, as well as total lab automation. For the latter, Grys said they are looking at the "two major players," namely Biomerieux's Copan and BD's Kiestra. "Our other sites are also interested, so we'll probably have to do a deep dive to figure out what works best for all of our sites and then pursue a contract, so we get good value out of that decision," he added.
Grys further noted that he is working with a group at Arizona State University Biodesign Institute to develop rapid susceptibility testing, and that the group recently won almost $6 million in funding from the National Institutes of Health. "I think the rapid phenotypic area is going to be an important area in the next few years," he said.
Grys' lab was the first in the Mayo system to bring in the BioFire FilmArray, back in 2013, he said. "There was a time five or 10 years ago where people thought syndromic panels were going to really change everything, but I think we're realizing in that past couple years that they are another tool, and like any other tool, it is important how you use it." The testing can either make a difference, or just cost more and not actually improve patient care, he said.
"In our hands, for the gram-positives, I think there's a lot of favorable trend there and if we had more numbers we'd probably have [significant results], so to me it was encouraging. And for gram-negatives, it was just a sign that in the way we've been doing it, we're not impacting decisions, and if we want to impact decisions we'll have to change something," Grys said.