Activity Transcript

Kevin Messacar, MD, PhD: Hi, I'm Dr Kevin Messacar. I'm an associate professor in the Department of Pediatrics, Section of Infectious Diseases at the University of Colorado, in Children's Hospital Colorado, in Aurora. Welcome to this Medscape program titled, "Clinical Pearls and Diagnostic Stewardship: From Knowledge to Application." It's wonderful to be joined today by 2 leaders in the emerging field of diagnostic stewardship. Dr Ferric Fang is a professor of laboratory medicine and pathology at the University of Washington School of Medicine in Seattle. And Dr Dan Morgan, who is a professor of epidemiology and public health at the University of Maryland School of Medicine in Baltimore. Welcome, Ferric and Dan.

Daniel J. Morgan MD, MS: Thank you.

Ferric C. Fang, MD: Thank you.

Dr Messacar: Today we'll be discussing diagnostic and antimicrobial stewardship and more specifically how rapid multiplex polymerase chain reaction (PCR) panels can be optimally implemented for clinical care to improve the selection of appropriate antibiotics and reduce unnecessary use. We'll start by going through some big picture concepts, then drill down to look at a few case studies to put rapid testing into a clinical context. To start, I think it's important to define a couple of terms you'll be hearing in this program and others in this series, diagnostic stewardship and antimicrobial stewardship. Dan, could you start us off by explaining what the difference is between these terms and what the goals are for each?

Dr Morgan: Certainly. So the goal of diagnostic stewardship is to select the right test for the right patient to generate accurate clinically relevant results. But then I think most importantly that that result comes at the right time, prompting the right clinical action to improve patient care. Whereas antimicrobial stewardship is a bit different. It's trying to ensure the right interpretation that leads to the right antimicrobial therapy at the right time to improve clinical outcomes and decrease unnecessary antimicrobial use. And it often occurs a little bit downstream from the diagnosis portion that diagnostic stewardship focuses on. And diagnostic stewardship I think is often done for different reasons. A goal of diagnostic stewardship really is to focus on the patient care outcomes. So to improve patient care, to avoid patient harms, things like side effects from antibiotics or antimicrobial resistance to optimize antimicrobial use. And then down near the bottom I would put improve institutional costs and metrics.

It's obviously relevant to any of us who work in a hospital, but that's not the driving force behind what we're doing. We're not trying to just improve metrics, but really trying to make things better for patients. And why is there a need for diagnostic stewardship? Well, we've seen the use of diagnostics has really blown up during my lifetime in medicine where there was very limited diagnostics probably 40 years ago. And now there is lots of high-tech and complicated diagnostics. And this survey of Emerging Infections Network (EIN) infectious disease physicians highlighted some of those problems that there's test characteristics that are important for people to know, like test accuracy. And that's really sort of how good is it detecting when someone has a disease, how good is it detecting when someone doesn't have disease. Having adequate turnaround time, so these results are able to be used in a real-life clinical setting.

And in this survey, it identified that the majority of respondents felt some testing is becoming too complex for non-infectious disease physicians. And almost 80% believe that there should be stewardship for complicated and/or expensive tests. And so these are things like multiplex molecular respiratory panels, broad-range PCR testing, and antigen-based tests for fungal infections.

Dr Messacar: Thanks Dan. That was some really helpful background to get us started. Ferric, now on you. What are some of the issues that clinicians need to think about when selecting a diagnostic test? Then can you explain how the diagnostic stewardship of these syndromic PCR panels can be used to improve antimicrobial stewardship?

Dr Fang: Thanks, Kevin. There are a few key questions you have to ask yourself when you're selecting a diagnostic test. First, is the test appropriate for the clinical setting? Second, will care of the patient be affected by the result? In this regard, it should be the goal to limit testing in patients who have a low pretest probability of disease and to use tests in those settings where they have been shown to be either cost-effective and/or improved patient care. Third, will the results be available in time to impact patient care? And fourth, will clinicians understand the result?

This figure illustrates the different steps between ordering a test and getting a result. With a traditional culture-based test, the biggest time delay is allowing the organisms time to grow, which can take 24 hours or more before a report can be provided to the clinician. Molecular tests are more rapid, and if they're provided at or near the point of care, the results can be available to the clinician within even a few hours of sample collection. Diagnostic and antimicrobial stewardship are distinct, but they're intimately related. The laboratory can perform tests quickly, but it's also important for clinicians to receive and act upon those results in a timely manner. A result that's reported in the electronic health record is only going to be impactful if someone can act on it. So real-time antimicrobial stewardship decision support can close this gap by providing individualized antimicrobial recommendations at the time diagnostic test results become available. This provides the support for decision making that can really optimize care.

Dr Messacar: Thanks so much, Ferric. Now let's move on to our cases to put what we've learned into a clinical context. Dan, you're up first to present our first case.

Dr Morgan: Okay, so I'll be presenting a case of a patient with respiratory disease. And this is a simple case, but I think helps it illustrate some of the key things that we should be thinking about with diagnostic stewardship. So this is an adult patient with a history of heart failure who presents to the emergency department. They report 3 days' worsening shortness of breath and feeling sweaty. They report, they also intermittently forgot to take home furosemide and angiotensin-converting enzyme (ACE) inhibitor in the past week. And in the emergency room they are known to have fever but otherwise normal vital signs, and they have a chest x-ray that's abnormal and is read as, "consider infection." And the question for the emergency room physician is, should he get antibiotics? So I think this is a very simple bread-and-butter case. And the differential probably came to people quite quickly while listening to this, that this patient may have heart failure, they could have pneumonia, or they could have some aspect of an upper respiratory tract infection.

And we know that in settings like this, there's varied clinical practice giving the uncertainty if this is just heart failure, just pneumonia, or something else happening. So we know that antibiotics often depend on pneumonia being present, but there's some uncertainty with a chest x-ray like this in a patient that doesn't fit one clear description. So this is a place where I think quantifying probability can help. We often talk about something being unlikely, or uncertain, or likely, but it can be hard to quantify that a bit more. And so I think trying to think about how likely it is more objectively and then also using respiratory viral panel testing can help guide clinical care. So let's talk about the multiplex PCR panels. So a multiplex test is defined as just if there's greater than one PCR test run simultaneously on the same cartridge. There are some of these panel tests that are small, and they target high-yield pathogens. I think everyone's becoming quite familiar with these during COVID. But one that just has COVID, flu, and RSV, targeting sort of the most severe respiratory viruses that often have more treatment impact.

And there are broader panels that have up to 22 targets, including less relevant pathogens. These tests generally have a high sensitivity, and turnaround time is rapid compared to culture so long as they're being done in a laboratory that's on site. There is however, this issue that you're running multiple simultaneous tests. So that could increase the likelihood of a clinical false positive. So a positive test that may not be related to the actual disease the patient is experiencing. And medical centers implementing multiplex PCR panels tend to observe increased testing, including testing not recommended by guidelines. So how do you strike a balance? The next slide describes some of the organisms that are present on some of these panels. And I think it's not worth going into details about this. Really beyond seeing that there are lots of different viruses as well as some bacteria. The bacteria initially were concern that people thought this would often lead to overuse of antibiotics. But I think in real life we are seeing the bacteria are very infrequently positive, but we see a lot of common viral upper respiratory tract pathogens occur.

Okay, so back to this case. This patient who is in the emergency room is being treated for heart failure and they're now breathing better. The emergency room attending considers this possible viral pneumonia based upon working through the probability of disease. And this patient had a low probability of bacterial pneumonia based upon relatively normal vital signs and lack of other findings, and an alternative diagnosis that was more likely. They did a respiratory viral panel test and came back with human metapneumovirus. And they used this information to decide that this patient was probably viral pneumonia and not bacterial, and not to treat him with antibiotics. And you provided the patient with instructions to return if his condition worsens. Versus the alternative option that I think often occurs in this setting is someone would be treated with antibiotics and sent home with a diagnosis of possible bacterial pneumonia, which would put them at increased risk of Clostridioides difficile diarrhea or rash related to the antibiotics.

And I just emphasize, this is an example of focusing on the clinician part of diagnostic stewardship, not in the more traditional implementation of tests that occurs often in centers as a part of quality improvement in pathology services. And I'll talk a little bit more about respiratory panel testing. Because I do think I just talked about the positive. So in the right setting where you have a low-risk patient who you don't think has a bacterial pneumonia, but the information can help nudge you away from antibiotics, they can be useful. But we know that respiratory viral panels used without thoughtfulness about how the results will be interpreted is not helpful.

So there's a great randomized controlled trial of children. They applied respiratory viral panel testing to children in a large pediatric emergency department (ED). And these were all patients with acute respiratory illness, so influenza-like illness. And half of them received the rapid testing, half of them did not. And they noticed no change in antibiotic prescribing with the respiratory viral panels and they were more likely to receive antivirals and be hospitalized with the result, but no change in their outcome. So indiscriminate use of respiratory viral panel testing does not improve care. So we shouldn't just be doing this to all patients, but only in patients where we think it will change the outcome. I think a few key takeaways for people on respiratory viral panels, they tend to be better if used at initial presentation. They have a lower yield if used in hospitalized patients who have been present for more than 2 or 3 days. They're best in patients who've had previous testing. So repeat testing doesn't really add much besides cost and potentially confusing information. And they probably work better in immunocompromised patients. That's often the patients where finding some of these viruses may change management in some ways.

But really the key question I think for everyone to ask, just like any diagnostic is, do you know that you're going to do something based upon the results? If you stop and think for a minute, and think, well, this will change my management of a patient, well, then a diagnostic is often a good idea. But I think often we order diagnostics kind of upfront without thinking about what we'll do with them and they probably have no impact on care, or it could even potentially be harmful if they provide unuseful information. So, the biggest point I would try to leave clinicians with is, think about what you're going to do with the results before you order the test, and then tests are usually helpful.

Dr Messacar: Great point. Thanks, Dan, for a great case and review of diagnostic stewardship for respiratory testing. Ferric, you're up next for a gastrointestinal (GI) case.

Dr Fang: So I'd like to switch gears and talk about the use of the GI panels. I'd like to first discuss the case of a 27-year-old Seattle resident who presented with a week of watery diarrhea and severe abdominal cramping. The patient had a history of heroin use and was currently experiencing homelessness. On examination, diffuse lower abdominal tenderness was present without rebound, and a peripheral white blood cell count was elevated, and a still multiplex PCR panel was obtained. And this detected a target indicative of Shigella or enteroinvasive Escherichia coli. To understand our approach to this patient, it's important to know that Seattle has been experiencing multiple, ongoing, parallel outbreaks of shigellosis. And these have been going on for several years, and they primarily involve 2 populations: men who have sex with men, and persons experiencing homelessness. We perform genomic analysis of the Shigella isolates, and we found that different strains were responsible for the infections occurring within these populations, with occasional evidence of spillover between the populations.

Importantly, the antimicrobial resistance patterns of these Shigella isolates differ. The strains obtained for men who have sex with men have tended to be resistant to trimethoprim-sulfa[methoxazole], but susceptible to ciprofloxacin, while the strains obtained from homeless persons have shown the opposite pattern. Our emergency room physicians developed a protocol whereby the patients who had acute diarrhea were evaluated with a rapid stool PCR test. And depending on which at-risk group the patient belonged to, those with Shigella were treated with ciprofloxacin. And similarly, patients could receive trimethoprim-sulfa if they were in the appropriate at-risk group. Despite a high rate of multidrug resistance, this approach allowed our hospital to prescribe appropriate empirical antimicrobial therapy about 70% of the time. And this was an important intervention both to help the patients and to slow the spread of resistant Shigella downstream to other persons.

Our institution has developed specific indications for the use of these panels to ensure that they're not overused. The criteria are based on published Infectious Diseases Society of America (IDSA) and American College of Gastroenterology (ACG) guidelines, and they're designed to limit testing to patients in whom antimicrobial therapy might be appropriate. So just as Dan was saying, we want to use the test when the test results are likely to influence treatment decisions. These criteria include moderate to severe illness; diarrhea, 1 week or longer; dysentery, meaning fever, bloody or mucoid diarrhea, severe cramping, and potentially sepsis. And then an immunocompromised host or an outbreak that involve food handlers, healthcare workers, daycare attendees, or employees, and residents of institutions.

Multiple commercial PCR panels for evaluating patients with gastroenteritis are now widely available, and they range from as few as 3 to as many as 24 different targets. Labs can select the ones that best fit their patient population, workflow, and budget. At our institutions at the University of Washington, we found that a multiplex PCR panel dramatically increased the sensitivity of stool testing. When we compared our culture and PCR in parallel, nearly 6 times as many pathogens were detected by PCR. PCR was much more rapid than traditional testing. Even though early on we were only reporting results during the day shift, clinicians still found that they were sufficiently rapid to allow them to withhold empiric antibiotics in many cases while awaiting results. In other words, this testing improved antimicrobial stewardship. Over time, our turnaround time has gotten shorter, and clinicians are now typically receiving the results while their patients are still in the emergency room or office. And this allows optimization of therapy, as we saw with our Shigella case.

A recent publication that my colleagues and I made in the Journal of Clinical Microbiology analyzed real-world data to look at the clinical impact of multiplex PCR testing in adults who presented with acute gastroenteritis. This study used the premier healthcare database, a database of hospital-associated visits in the United States that includes approximately 25% of the hospital encounters in the United States -- more than a billion encounters. More than 36,000 outpatient adults with gastroenteritis severe enough to warrant diagnostic testing were included in the study. The results confirmed that PCR is much quicker than culture. The mean turnaround time was 32 hours for traditional culture workup, and just 6 hours for the large multiplex panels. And this is across hundreds of hospitals. Although the laboratory costs were slightly higher for PCR, we also found that the costs were offset by lower numbers of secondary testing, required imaging, or follow-up visits. So that the overall healthcare costs for the index visit and the 30-day follow-up period were virtually identical for the diagnostic approaches.

So with PCR, you got a more rapid result, a more accurate result, and it came at virtually no cost for overall healthcare system. And as we saw in our hospital, the PCR panels detected many more pathogens and the ability to rapidly detect viral causes of gastroenteritis enabled clinicians to use fewer antibiotics in the large multiplex PCR panel group. So this is a really clear real-world demonstration of diagnostic stewardship.

Dr Messacar: Thanks, Ferric. That was a great case and review of the literature on diagnostic stewardship of GI panel testing. Now I get to present the last case, and we'll move to my realm of pediatrics to highlight a central nervous system (CNS) case. So this is a three-and-a-half-week-old, full-term, previously healthy infant who presented in the summer with fever to 101 degrees, no sick contacts or other symptoms. And his exam was really just notable for fussiness. Otherwise had a flat non-bulging fontanelle, no meningeal signs, and no other focal findings. As part of the recommended workup for a febrile neonate, got blood and urine cultures sent, which are pending, and a urinalysis that was normal. Had a peripheral complete blood count (CBC) with a white count of 4.6, and a normal C-reactive protein (CRP) at 0.9 mg/dL. The baby did get a lumbar puncture (LP) due to persistent fussiness, which showed a pleocytosis with 14 white blood cells, 82% neutrophils, 4% lymphocytes, 14% monocytes, a normal glucose in protein, a negative Gram stain, and cerebrospinal fluid (CSF) cultures were sent and are pending.

The baby, due to the pleocytosis, got doses of vancomycin, ceftriaxone, and acyclovir in the emergency room and is getting admitted to the pediatric ward. So, what are the next steps at this point? Now we are in an era where often there's pathogen-specific testing available with singleplex PCRs for enterovirus, parechovirus, herpes simplex virus 1 and 2, and syndromic panels with multiplex PCRs for meningitis/encephalitis causes. To highlight the FDA-approved meningitis/encephalitis panel, I wanted to go through some of the targets, which includes 6 bacteria, some of the most common causes of bacterial meningitis in infants, including Escherichia coli, and Streptococcus agalactiae or group B strep, and community-acquired bacterial meningitis. Streptococcus pneumoniae, Haemophilus influenzae, meningococcus, and then listeria, which we can see in infants and elderly and immune compromised. We have treatable viruses like HSV-1 and -2, and varicella zoster virus, as well as alternative etiology viruses like enterovirus and parechovirus, which may give us the cause of symptoms and allow us to stop antibiotics.

And lastly, some more confusing viruses to interpret like cytomegalovirus (CMV) and human herpesvirus 6 (HHV-6), which could be detected but not be the cause of disease either due to reactivation or chromosomal integration. And lastly, more pertinent to the adult immunocompromised patient population, we can detect Cryptococcus neoformans and Cryptococcus gattii. So how do we implement these new rapid syndromic panels? And I wanted to walk through a diagnostic stewardship implementation package that we used at our institution when we rolled out this panel. And this is a build-out of the diagram that Ferric displayed earlier as far as diagnostic stewardship concepts.

When we have a child with suspected CNS infection, do a clinical evaluation to decide they need a LP, the clinician can order a meningitis/encephalitis panel (MEP) if they use indication selection to tell us why they're ordering the test. That indication selection in the electronic medical record is populated with approved use criteria that has database indications for testing. This is driven by adult data that suggest CSF without a pleocytosis, so without cells there, cannot undergo this extensive testing, decreasing costs without impacting yield. Yet we have some pediatric caveats there. Our young infants less than 2 months of age often don't have a pleocytosis. Our immunocompromised patients may not have a pleocytosis. Our encephalitis patients get a ton of testing, which incurs a lot of costs that can be run on one panel. If you don't meet one of those criteria, we use conditional testing, which means the CSF goes down to the lab and they'll only run the test if it has greater than 5 white blood cells following that data from the adult literature. Then that test gets in the hands of the lab, is run real-time on demand, and that test result comes out with real-time decision support intervention using antimicrobial stewardship -- either through the ID team if it's a treatable organism, like bacterial meningitis or one of the herpesviruses to get them on the right dose of antibiotics from the start, or our stewardship team. If we detect alternative etiologies like enterovirus and parechovirus, that could allow us to stop antibiotics.

And ultimately, the goal is to get to targeted therapy and discontinue unnecessary therapies quicker. So we studied this rollout using diagnostic stewardship for implementation, which we started in 2017, and just published the results of this large clinical trial. We showed not only faster diagnosis, cutting the time to positive CSF results in half, but also clinical impact. Decreasing the intravenous (IV) antimicrobial duration, detecting more infectious causes of neurologic disease, and meeting the study endpoint of decreasing the time to optimal antimicrobials by about 10 hours. Similar to Ferric's data, though this syndromic panel was more expensive than the conventional testing that was conducted when you looked at overall total hospital costs, they were unchanged, so improved clinical outcomes without impacting total hospital costs. So let's bring it back to our case.

We had this young baby with suspected CNS infection who underwent an LP. He met the approved use criteria being less than 2 months of age and had a pleocytosis. The MEP was run in real time and enterovirus was detected in a CSF within 2 hours. That result went out to our antimicrobial stewardship team who contacted the provider with the literature that suggests when you detect enterovirus in a young febrile infant that's otherwise well-appearing, you can stop antibiotics and acyclovir safely. And in this case, we decrease antibiotic exposure, we decrease the baby's risk of kidney injury from the vancomycin and acyclovir, and we shorten the length of stay. So that wraps up our clinical cases, but I wanted to touch briefly on the practical implementation of these diagnostic stewardship interventions. Dan, could you explain who the key stakeholders are in diagnostic stewardship and how they can work together to create and implement these types of interventions?

Dr Morgan: Yeah, thanks Kevin. And that was a great case, and love to discuss it more later. So diagnostic stewardship is a relatively new idea, unlike antimicrobial stewardship where you may have somebody who has a job where they're the antimicrobial steward for the hospital. Typically, diagnostic stewardship is run by a group of people. It's often led by antimicrobial stewardship, or infection control, or clinical microbiology. But you need a lot of people involved to make certain that those end users of the test, that the people who kind of shape how tests are implemented in the electronic health record and in the laboratory as well as people who are affected by test results, like those who collect metrics or oversee hospital costs, all need to be involved to make certain that the changes that are being made to testing unfold smoothly and therefore are best able to improve patient care.

So those groups, as I mentioned, microbiology is key, antimicrobial stewardship, infection control, infectious disease, clinician end users are key. So often hospitalists, or surgeons, or emergency room physicians, and really targeting the group that will be most impacted by the testing. And then also working with hospital administration who you may need them to sign off on potentially greater cost for these tests upfront, even if those costs are recovered in other places. As well as for support, for changing the way that tests are ordered, which sometimes it requires a bit of time for clinicians to accept that degree of input.

Dr Messacar: Great. Thanks so much Dan. It really shows diagnostic stewardship is really a team sport. So lastly, I'd like to pick your brains while I have you and ask you to share your expertise on where you think this field is headed. Going forward, what information do we still need to guide best use of these multiplex assays for diagnosing infectious diseases? Let's start with you Ferric.

Dr Fang: Well, I think all of these tests are different, and so I'd like to see more studies of real-world data like you performed with the meningitis panel and the one we performed for GI panels to help us understand the clinical impact of the coupling of rapid diagnostics and antimicrobial stewardship. Really looking at important real-world outcomes like antibiotic use and patient outcomes, including costs so that we can really optimize the use of these tests. I think each one is nuanced, and the advantages that they bring to specific kinds of clinical situations are different. So we can't really generalize from one kind of test to another. I think that's an area where we still need a lot more data.

Dr Messacar: I think that's a really important point, Ferric. And I think you hit on the point that one size really doesn't fit all, not every test and not every setting. And I think that's where a lot of the implementation science work is being done in this realm. What that implementation package I talked about with MEP probably isn't feasible at a lot of centers that may not have the resources to do that. Yet, there are lower labor-intensive impacts that you can have from reporting language or allowing providers to order a test, but using the lab itself to decide when the test is run. Dan, what have you seen as far as effective implementation strategies in different settings?

Dr Morgan: Yeah, that's a great question, I think one thing that I've observed is that there's just a great variation in terms of use of these tests. It seems like there's some hospitals where they're used quite frequently and other hospitals where they may be unavailable or blocked. And I think both of those examples illustrate ways that are not optimal for patient care. I think all of us agree that these are kind of remarkable tests as far as what they are able to do, and they can be very useful, especially maybe in a laboratory that doesn't have the traditional methods that may be more labor intense or require more expertise, especially say for a GI panel. So they're great tests, but we also know that they're overused in the low-risk situations and sometimes not used in high-risk situations where it could be very helpful to changing management.

So, I think we need, just like you described Kevin, studies to look at these. I mean, ideally, I think even randomized trials are helpful for very common indications, and that guidance can then be rolled out to other hospitals where they may not have the time or resources to figure that out. But the indications can be developed for how to use these tests most appropriately in patients, and even how that can be built into the system, say with the electronic health record for ordering or the microbiology lab for approving tests so that they give the optimal input and can improve patient care without causing any harms or unnecessary costs for a system.

Dr Messacar: So, as we learn today, diagnostic stewardship is an important movement in the infectious disease field, helping us to utilize both existing tests and newly emerging rapid molecular diagnostic technologies to their full potential. There are many parallels to the now established field of antimicrobial stewardship with which diagnostic stewardship interventions can be paired to ensure optimal use of tests leads to improvements in antimicrobial use in patient outcomes. Similar to antimicrobial stewardship, not being the antibiotic police, but working to make best use of our antibiotics when needed. It is important that diagnostic stewardship is not seen as the test police, but instead as systematic interventions to ensure we are using these diagnostic technologies in the most impactful and cost-effective manner. As we saw from the cases today, sometimes that may mean limiting testing in low-yield situations, but also can mean promoting use of a more rapid syndromic panel when there are data to support a positive clinical or financial impact.

The simple adage of right test, right patient, right time, with the right interpretation leading to the right antimicrobial at the right time remains this guiding principle of this movement. And I'm excited to see all the amazing implementation science work being done in this field, the new technological advances, and where the diagnostic stewardship field goes from here.

Ferric and Dan, thank you for joining me today. This has really been a great discussion that I think our audience will benefit from hearing. And thank you all for watching and participating in this Medscape activity. Please continue on to answer the questions that follow and complete the evaluation.

This transcript has not been copyedited.