Latest / The SHEA Podcast / VAP Diagnostic Stewardship

Transcript

VAP Diagnostic Stewardship · The SHEA Podcast

• 0:06
  SPEAKER_01: Welcome to the SHAYE podcast.
• 0:08
  SPEAKER_01: I'm Dr.
• 0:08
  SPEAKER_01: Kimberly Clays, a pharmacist and associate professor in the Department of Practice, Science, and Health Outcomes Research at the University of Maryland School of Pharmacy.
• 0:16
  SPEAKER_01: I'm also a member of the SHAE's Antimicrobial Stewardship Committee, which works to advance initiatives, guidance, and educational programs that promote antimicrobial stewardship across healthcare settings.
• 0:25
  SPEAKER_01: And today we're discussing diagnostic stewardship in the hospital, particularly as related to ventilator-associated pneumonia or VAP.
• 0:33
  SPEAKER_01: And VAP remains one of the most challenging infectious syndromes to diagnose.
• 0:36
  SPEAKER_01: Many clinical findings that we associate with VAP, fever, leukocytosis, changes in secretions, and radiographic abnormalities, are nonspecific.
• 0:44
  SPEAKER_01: And at the same time, respiratory cultures from balayed patients frequently identify organisms that may represent colonization rather than true infection.
• 0:51
  SPEAKER_01: This makes VAP particularly well suited for diagnostic stewardship.
• 0:55
  SPEAKER_01: And rather than focusing only on when antibiotics should we start or stop, diagnostic stewardship asks us to consider the entire diagnostic and treatment pathway to discuss this topic.
• 1:04
  SPEAKER_01: I'm pleased to be joined today by two leading experts who have helped develop and define this field.
• 1:10
  unknown: Dr.
• 1:10
  SPEAKER_01: Owen Albin is assistant professor of internal medicine at the University of Michigan School of Medicine.
• 1:14
  SPEAKER_01: His work focuses on improving the diagnosis and management of VAP through pragmatic diagnostic stewardship interventions, alveolar biomarkers that support personalized antimicrobial therapy, and patient-informed hierarchical composite endpoints for use in nosoclineal pneumonia trials.
• 1:29
  unknown: Dr.
• 1:29
  SPEAKER_01: Annestick Samuels is an associate professor of pediatrics in the division of pediatric infectious diseases at Johns Hopkins University School of Medicine and associate hospital epidemiologist at Johns Hopkins Hospital.
• 1:40
  SPEAKER_01: Her work, including the Bright Star Respiratory Collaborative, have focused extensively on understanding and proving and improving respiratory culture utilization in children with artificial airways.
• 1:49
  SPEAKER_01: Thank you both for joining us.
• 1:52
  SPEAKER_00: Pleasure to be here, Ken.
• 1:53
  SPEAKER_02: Hi.
• 1:54
  SPEAKER_02: This is uh one of my favorite topics.
• 1:56
  SPEAKER_02: So I really appreciate the invitation and uh chance to get to discuss this with you today.
• 2:02
  SPEAKER_01: Yes, thank you both for being here.
• 2:03
  SPEAKER_01: And I think just to start broadly, I want to uh talk about what makes VAP uniquely challenging from a diagnostic stewardship perspective and get both of your insights on that.
• 2:13
  SPEAKER_00: I will uh tee things off, I guess.
• 2:16
  SPEAKER_00: Um infectious diagnoses can be relatively straightforward because they they require more or less just the isolation of a germ from an otherwise sterile site in the body.
• 2:32
  SPEAKER_00: But not unlike the genito-urinary tract.
• 2:34
  SPEAKER_00: The respiratory tract is not entirely sterile.
• 2:38
  SPEAKER_00: And so uh the finding of a germ from the respiratory tract, uh, unlike something like bacteremia with a real plausible, convincing pathogen, is not sufficient to make a diagnosis.
• 2:51
  SPEAKER_00: And instead, uh, what's required is a syndromic diagnosis.
• 2:56
  SPEAKER_00: So we don't have a gold standard reference diagnostic test for ventilator-associated pneumonia beyond lung histopathology.
• 3:06
  SPEAKER_00: And that is obviously very challenging to get in patients who have not uh passed away and undergone an autopsy, oftentimes.
• 3:14
  SPEAKER_00: So, as a result, what we have to do is leverage probabilistic reasoning by amalgamating a number of different clinical, radiographic, and lab-based signs or symptoms that patients might have and cumulatively weight their respective probabilities in informing the likelihood of whether or not the patient has pneumonia.
• 3:40
  SPEAKER_00: That is extraordinarily challenging for every single doctor to do because, as human beings, we really struggle with probabilistic reasoning.
• 3:49
  SPEAKER_00: Um, and I think more than anything else, uh it's the lack of a really firm, easily accessible gold standard diagnostic test that prevents us from really uh accurately ascertaining whether or not pneumonia is present.
• 4:07
  SPEAKER_02: Thanks, Owen.
• 4:08
  SPEAKER_02: I think all of what you said is uh equally true in pediatrics, um, that you know, it's this challenge that clinicians are mindful of the risk that their patients could develop a low or l infection or pneumonia, um, especially patients with artificial air radius.
• 4:27
  SPEAKER_02: So if they are on a ventilator, for example, they have an increased risk associated with those devices.
• 4:33
  SPEAKER_02: And so they want to make sure they're not missing things and they want to make sure they're covering their bases and treating things that are treatable.
• 4:41
  SPEAKER_02: Um, but pneumonia has a lot of overlapping findings with other uh non-bacterial causes, so viruses is another big one, as well as non-infectious causes of respiratory stress, so pulmonary edema or other lung pathology that can even look very similar in a chest x-ray, give similar uh signs and symptoms, or they have a patient who's very complex and now they have a new fever and they don't know what's driving it.
• 5:13
  SPEAKER_02: There's this diagnostic uncertainty about it.
• 5:15
  SPEAKER_02: And so I do think that that diagnostic uncertainty is one of the drivers for these cultures.
• 5:20
  SPEAKER_02: And then we're stuck with a very non-specific test with germs that uh I have heard labeled as cathogens, but are in fact are often just normal bacteria in the respiratory tract.
• 5:32
  SPEAKER_02: And they're only a pathogen if they're causing an infection.
• 5:36
  SPEAKER_02: And so we have two challenges mixed together of a lack of clinical uh that diagnostic uncertainty and lack of a clear clinical diagnosis, and a test that, you know, has some opportunities uh for misinterpretation.
• 5:52
  SPEAKER_01: Um great responses overall.
• 5:54
  SPEAKER_01: And I think, Owen, when you mentioned urine cultures, I know that um before you've helped coin the term asymptomatic bacteria spunia.
• 6:01
  SPEAKER_01: And I have tried so hard to make that a thing.
• 6:04
  SPEAKER_00: Yeah, you know, we tried to make it a thing.
• 6:06
  SPEAKER_00: It it doesn't have the uh a real ring to it, so I don't think it's caught on.
• 6:10
  SPEAKER_01: It should, though, because it really encompasses that this culture could be positive, but that doesn't mean that it's causing infection to get back to both of what uh your points were.
• 6:19
  SPEAKER_01: Um and Anna, you also mentioned uh drivers.
• 6:22
  SPEAKER_01: And I know that both of your research programs have examined not just respiratory culturalization, but what are the reasons that Cliches decide to order these tests?
• 6:29
  SPEAKER_01: And what have you learned about the key drivers of respiratory culturing?
• 6:32
  SPEAKER_01: And were there any behaviors or assumptions that surprised you?
• 6:36
  SPEAKER_02: Yeah, so actually you're um referencing back to one of the first studies I ever did on the topic.
• 6:42
  SPEAKER_02: I was a fellow at the time, and I was just getting warmed up in this whole career of diagnostic stewardship.
• 6:49
  SPEAKER_02: And I wanted to understand why people were sending these cultures and what they were using them for.
• 6:55
  SPEAKER_02: So I actually had a little report that would tell me who ordered the culture that day.
• 7:00
  SPEAKER_02: And I would go and physically try to find the person who ordered the culture and ask them.
• 7:06
  SPEAKER_02: So I did like a verbal survey, and then I would follow up several days later to uh understand how the results of that culture impacted the patient's management.
• 7:16
  SPEAKER_02: And so this was prospective and they really couldn't escape me, I guess.
• 7:21
  SPEAKER_02: Um and what I learned, at least in our center, was two clinical drivers kind of came out of that.
• 7:28
  SPEAKER_02: One was the patient developed a fever, and there was a report of a change in their respiratory secretions being suctioned out of their airway.
• 7:38
  SPEAKER_02: And about half of the cultures really weren't acted on, or other clinical data ended up being more informative for the patient's course.
• 7:46
  SPEAKER_02: So roughly half of those cultures maybe were not very helpful in the patient's management.
• 7:52
  SPEAKER_02: And as this work has progressed, and I've had the opportunity to work with clinicians across the country now and also reevaluate our own practices in our institutions.
• 8:04
  SPEAKER_02: Um, one of the other things about that fever driver is it's often linked to a sepsis evaluation.
• 8:11
  SPEAKER_02: And so early recognition of sepsis is very important.
• 8:16
  SPEAKER_02: No one wants to miss a bacteria or treatable bacterial process.
• 8:20
  SPEAKER_02: And early on, there is some linking with temporality with that, getting antibiotics on board within an hour, meaning diagnostic tests ideally would be obtained within that time frame.
• 8:32
  SPEAKER_02: So there's some urgency to all of it.
• 8:34
  SPEAKER_02: And so I have found that in some cases, um, sepsis evaluations or kind of this panculture approach might be used sometimes where clinicians hear that their patient has an infectious sign and does the full workup of many things all in one kind of bundle: microbiology tests, lab work, a chest x-ray all obtained at once, rather than kind of stepping back and saying what's the most likely thing and doing a stepwise process.
• 9:03
  SPEAKER_02: So, from a uh clinical symptom perspective, those are definitely some of the drivers of uh the initial ordering or thought to get the culture.
• 9:14
  SPEAKER_00: I agree with uh with everything Anna said, uh, specifically the some of the specific clinical syndromes we see uh unnecessary diagnostic testing performed on and the kind of entire macro phenomenon of sepsis hypervigilance.
• 9:31
  SPEAKER_00: Um I think I'd also point to kind of two other larger issues in in what drives overdiagnosis of VAP.
• 9:40
  SPEAKER_00: Um, one is that our existing clinical definitions for VAP are rather blunt instruments.
• 9:46
  SPEAKER_00: So if you look at any kind of consensus guideline, what you'll read is you need a newer, persistent infiltrate on chest imaging.
• 9:54
  SPEAKER_00: And then it has to be accompanied by like something that makes you think the patient could have pneumonia, be it a fever or a white count or purulin secretions or even hypoxemia.
• 10:05
  SPEAKER_00: Um that is itself an incredibly broad constellation of different types of patients and phenotypes.
• 10:14
  SPEAKER_00: And so I think in the future, if if if clinical syndromes can be phenotyped in a more granular way, um, that can allow us to risk stratify folks based on diagnostic likelihood and perhaps refine diagnostic decisions a little bit more.
• 10:33
  SPEAKER_00: Another issue, though, is that when it comes to HAP and VAP, you know, we think probably the primary modifiable determinant of patient outcomes is early receipt of appropriate antibiotic therapy.
• 10:46
  SPEAKER_00: And these are very sick patients.
• 10:48
  SPEAKER_00: It's an ambiguous diagnosis to begin with.
• 10:51
  SPEAKER_00: And so many ICU providers just feel hamstrung sometimes by uh the entire idea that a patient is crashing.
• 11:02
  SPEAKER_00: Um, this is something that you could actually do for them.
• 11:06
  SPEAKER_00: Because if you miss the pneumonia, it can be catastrophic.
• 11:09
  SPEAKER_00: And if you intervene effectively based on a culture, it can be meaningful.
• 11:14
  SPEAKER_00: And so uh providers are really lack any incentive not to order these things whenever a patient is sick in any manner, shape, or fashion.
• 11:23
  SPEAKER_01: Well, those are both very good insights.
• 11:25
  SPEAKER_01: Oh, Ann, did you have any more to add?
• 11:27
  SPEAKER_01: Sorry.
• 11:28
  SPEAKER_02: Uh well, Owen's getting into this idea of like the cognitive biases also that we might have that make it harder for us not to get cultures.
• 11:37
  SPEAKER_02: This might be going to another question, but uh we can if you want me to keep going on that idea.
• 11:43
  SPEAKER_01: Oh, we can skip ahead to that because I think one of the recurring themes across the literature that you've both developed is that clinicians substantially overestimate the diagnosis of VAP and estimate the probability of VAP even further once an organism identified, even when the pretest probability is low.
• 11:59
  SPEAKER_01: How important is this diagnostic anchoring bias around positive cultures in VAP over diagnosis and antibiotic use?
• 12:04
  SPEAKER_01: And what strategies have you seen to be most effective for counteracting it?
• 12:09
  SPEAKER_02: Yeah, so I think there's two parts to it.
• 12:12
  SPEAKER_02: There's the first temporally, which is getting the test and then making the diagnosis.
• 12:17
  SPEAKER_02: So, in terms of getting the test, one of the things kind of giving it a name is that as clinicians, we sometimes there's a couple biases that may contribute to overtesting.
• 12:29
  SPEAKER_02: And this isn't specific just to pneumonia, it can be applied to other syndromes as well.
• 12:34
  SPEAKER_02: But uh often that we we feel an inclination to do something for our patient, and it feels awkward to just not do something and wait and watch.
• 12:45
  SPEAKER_02: And so sometimes we have this bias to act, and whether that be ordering the test or ordering an antibiotic to make a change in response to that clinical change rather than just seeing what direction it's going in, if we have the ability of time and safety to do that.
• 13:03
  SPEAKER_02: And another is um like a recall or outcomes bias, where perhaps the clinicians had a bad experience where a patient didn't fare so well, and maybe a bacterial process could have contributed to that.
• 13:17
  SPEAKER_02: And so that one bad outcome has much more weight than the actual probability that their patient has this infection right now.
• 13:27
  SPEAKER_02: And so the kind of concern to not miss it is out of proportion to the actual risk of infection the patient might really have in the moment.
• 13:37
  SPEAKER_02: And that also contributes to some of that misdiagnosis, or I see the positive culture.
• 13:42
  SPEAKER_02: So I'm gonna treat them for that bug now, that germ that I see.
• 13:45
  SPEAKER_02: Um, now that I've seen that a germ grew from their airway, I want to make sure I'm not gonna miss something and it leads to something bad.
• 13:54
  SPEAKER_00: You know, the the only other thing I'd add to what Anna said too is it's not uncommon for the person ordering the test to be different from the person who then later interprets and decides to act on it.
• 14:07
  SPEAKER_00: Um and it's a very different situation deciding whether to order respiratory culture versus being stuck footing the bill of a respiratory culture when you weren't sure it should have been ordered in the first place.
• 14:20
  SPEAKER_00: And so I think many providers, um, whether it's because there's a nocturnist who is admitting a patient, working them up and passing them off, or if you work in a teaching hospital, there's just a massive amount of turnover by trainees.
• 14:36
  SPEAKER_00: Uh, oftentimes the people who are acting on the results of cultures were not even the ones that greenlit testing in the first place.
• 14:42
  SPEAKER_00: And so they're kind of not following the entire diagnostic odyssey of the condition.
• 14:47
  SPEAKER_01: So these are both very great points.
• 14:49
  SPEAKER_01: And I think one of the things I wanted to kind of touch on too as we continue this discussion is um kind of discussing some of the similarities and differences in adult and pediatric patients when we talk about these drivers and diagnostic challenges.
• 15:02
  SPEAKER_01: So I know that NAO work is largely focused on children with artificial airways, and Owen, your work is focused predominantly on mechanically patients and adults.
• 15:09
  SPEAKER_01: Where do you see similarities and differences in these challenges in the adult and pediatric patients' populations?
• 15:17
  SPEAKER_00: Um, you know, I think one universal similarity is that, you know, every patient's obviously different, and children and adults are also different, but the very broad contours, I think, of what is likely to be VAP based on an amalgamation of signs and symptoms versus what is not supposed to be VAP is relatively consistent across age.
• 15:40
  SPEAKER_00: And so um I I know Anna might be getting in later to some of the work she's done on um deriving indications for ordering respiratory cultures in mechanically ventilated children, but many of the things that she has found there are equally applicable to adults.
• 15:59
  SPEAKER_00: What I would say uh a difference is between the two may be um more to do uh perhaps with the modality with which respiratory cultures are collected.
• 16:10
  SPEAKER_00: Um please tell me if I'm wrong.
• 16:13
  SPEAKER_00: I don't I don't live in Peds Land, I think, like like you do, Anna, but I think um uh there there's probably a little bit more distal respiratory culturing from Bronx and mini BALs, perhaps on the adult side, um uh than there may be uh in the pediatric world sometimes.
• 16:33
  SPEAKER_02: Yeah, so I fully agree with your first point.
• 16:36
  SPEAKER_02: I also think the general diagnosis of lung infection is similar across ages.
• 16:43
  SPEAKER_02: Um, some differences that in terms of risk for infection or other etiologies that might make it, you know, slightly different in the pediatric population.
• 16:56
  SPEAKER_02: Children generally tend to get more viral infections, the young, less than five-year-old group.
• 17:02
  SPEAKER_02: So if they're coming in and out of the hospital or they are patients at home, um, the viral etiologies are definitely a confounder when trying to interpret what's going on.
• 17:13
  SPEAKER_02: Is it viral or bacterial?
• 17:15
  SPEAKER_02: Another one is we have a special population of patients of babies who are born very prematurely.
• 17:22
  SPEAKER_02: They may have chronic lung disease or pulmonary hypertension related to that prematurity.
• 17:28
  SPEAKER_02: And so that's a particular type of patient that we do see struggle with respiratory exacerbations who might be ventilator dependent, for example, and at higher risk for both spectral infections and evaluating them for bacterial infections.
• 17:45
  SPEAKER_02: In terms of obtaining the cultures and what kind of specimens are obtained, I agree, Owen.
• 17:51
  SPEAKER_02: I think uh that tracheoaspirates are more commonly done in pediatrics.
• 17:56
  SPEAKER_02: And the threshold to do bronchoscopy is a little higher.
• 18:00
  SPEAKER_02: Often it requires uh requesting a consult by our pulmonologists who are not the critical care team, for example.
• 18:08
  SPEAKER_02: It's it would be a consulting team to come do that.
• 18:11
  SPEAKER_02: And so there's often a higher uh threshold and bar for when we get uh full bronchoscopies uh with bronchiobial lavage.
• 18:21
  SPEAKER_01: So great.
• 18:21
  SPEAKER_01: Um I think what I'm kind of hearing as we keep talking about this is we're kind of going between like talking about the individual behaviors around the patients and the provider interactions, whether there's different types of cognitive bias, fear of admission infections, uh, behaviors within individual uh providers and clinicians, as well as the workflow and culture of the institutions that they're talking that we're working in, and overall system level context.
• 18:47
  SPEAKER_01: A lot of things we think about when we're implementing different types of interventions.
• 18:52
  SPEAKER_01: And overall, I think there's some broad strokes that we can apply.
• 18:56
  SPEAKER_01: For instance, for stewardship teams, they might not have a dedicated diagnostic stewardship team, but either infection prevention or antimicrobial stewardship could start looking to improve respiratory cultures in their own hospital or health system.
• 19:08
  SPEAKER_01: And there's several points at which they can intervene on the diagnostic pathway, which we can talk more about, whether it's from ordering the culture, lab processing, all the way to reporting.
• 19:16
  SPEAKER_01: For teams just kind of beginning to think about this work, where in the pathway do you think they should focus first to help with this implementation of effective measures to improve respiratory culturing in both adult and pediatric patients?
• 19:28
  SPEAKER_00: Well, Kim, you mentioned um, you know, different phases of the diagnostic testing pathway.
• 19:33
  SPEAKER_00: And I think probably many listeners will be familiar with the broad contours of diagnostic stewardship.
• 19:39
  SPEAKER_00: Um, there's the decision to order a test where one can intervene.
• 19:43
  SPEAKER_00: There's the decision to collect the test where one can intervene.
• 19:46
  SPEAKER_00: And then there is how a test is reported by the laboratory to bedside providers, which also uh offers an avenue for intervention.
• 19:56
  SPEAKER_00: Um probably the lowest hanging fruit, both in terms of um how effective an intervention can be in lowering unnecessary testing, as well as um confronting minimal barriers to implementation, I think would be the ordering phase.
• 20:17
  SPEAKER_00: Um and uh that's for a couple of reasons.
• 20:21
  SPEAKER_00: One is that's where the bulk, I think, of work in this field has been done uh thus far.
• 20:28
  SPEAKER_00: Number two, um, the work that has been done suggests that that probably is a little bit more impactful than some collection or reporting interventions.
• 20:38
  SPEAKER_00: Um, and number three, I think we have identified some pretty clear phenotypes of patients who are commonly cultured that probably do not need to be cultured, uh, that you can very easily communicate to intensivists and bedside providers within ICUs.
• 20:58
  SPEAKER_00: Um, how you want to implement that, um, there's a bevy of options within diagnostic stewardship, from provider outreach, handshapes, handshake stewardship initiatives, um, prospective audit and feedback.
• 21:13
  SPEAKER_00: Our work largely has ignored trying to affect change in people's behavior by educating them and just trying to bluntly force them down a diagnostic pathway more than anything else uh through interruptive alerts and charts and uh uh gatekeeping restrictions on ordering tests.
• 21:34
  SPEAKER_02: Thanks.
• 21:35
  SPEAKER_02: Yeah, there's a lot of choices that you mentioned the gamut from the decision to order to how you interpret the result.
• 21:43
  SPEAKER_02: And I agree broadly that interventions that don't rely on individual clinician or behavior changes uh might kind of be uh have faster reach to people.
• 21:58
  SPEAKER_02: Um they also Also might be perceived differently.
• 22:02
  SPEAKER_02: So clinician autonomy is often a strong value, especially among really smart, really well-intentioned clinicians, often in practice for many years.
• 22:15
  SPEAKER_02: So there is a little bit of contextual understanding of what might make sense in your institution and whether folks at that facility will appreciate or support or find acceptable some of the more hard stop type interventions versus ones that are a more ground up trying to change their behaviors.
• 22:39
  SPEAKER_02: So, yes, changing behavior takes a lot more effort.
• 22:43
  SPEAKER_02: In the Brightstar work, a lot of what we've done is clinical decision support that what did not have hard stops.
• 22:51
  SPEAKER_02: It was algorithms and checklists on paper or digital versions that really required engagement from the ordering clinicians.
• 23:00
  SPEAKER_02: And in our recent uh work that we fortunately were able to publish, we did see uh different uptake in different contexts.
• 23:10
  SPEAKER_02: So in some centers, they saw a huge drop in their cultures and uptake of that clinical decision support.
• 23:17
  SPEAKER_02: In other centers, that was harder to achieve.
• 23:20
  SPEAKER_02: And we're looking formally into those contextual features and what was different from one place to another.
• 23:27
  SPEAKER_02: But um, it does rely on the individual changes to happen and how you engage people.
• 23:33
  SPEAKER_02: I would say if you are starting there, I definitely recommend having a partner in the unit, having a peer in that clinical group who will be an advocate and a champion for changing practices.
• 23:47
  SPEAKER_02: And I also recommend just coming at the whole thing with a little bit of curiosity.
• 23:52
  SPEAKER_02: What is going on in your center?
• 23:54
  SPEAKER_02: What are the actual drivers?
• 23:56
  SPEAKER_02: And sometimes there's some really low-hanging fruits.
• 24:00
  SPEAKER_02: For example, maybe trait aspirates are part of some order set, and all you have to do is unclick it from that order set.
• 24:08
  SPEAKER_02: And now it just got a little harder and less automatic for people to order because they need to intentionally order it.
• 24:15
  SPEAKER_02: Um, and so just understanding what are some of the things going on in your setting are is a great place to start.
• 24:23
  SPEAKER_02: And a second area that we hadn't talked about yet is partnering with the laboratory.
• 24:29
  SPEAKER_02: So I actually learned in this work that how laboratories process endotracheal specimens is quite variable across the country.
• 24:38
  SPEAKER_02: Um, when I first started working, actually many hospitals were doing quantitative reporting on endotracheal aspirate cultures.
• 24:46
  SPEAKER_02: And since then, there's been a little bit of a shift to semi-quantitative reporting, but that actually is an intervention in itself.
• 24:54
  SPEAKER_02: Also, specimen quality control.
• 24:57
  SPEAKER_02: Some labs may reject specimens that look like saliva and not like respiratory specimens.
• 25:03
  SPEAKER_02: So there are other things that we can do that may not change the ordering practice, but at least improve the quality of what we're reporting on.
• 25:12
  SPEAKER_00: I agree.
• 25:13
  SPEAKER_00: And I wonder if I could just piggyback a little bit more off of what Anna said.
• 25:17
  SPEAKER_00: Very much agree that partnership with intensivists is critically important because they are the ones that are actually taking care of the patients and they're the ones uh whose behavior is needed to affect change.
• 25:31
  SPEAKER_00: Um I wanted to uh emphasize just a couple of things as well.
• 25:35
  SPEAKER_00: I don't know if we're gonna get into this a little bit later, um, but really getting uh fairly specific about uh interventions on the ordering side of things.
• 25:48
  SPEAKER_00: Generally speaking, um, the way that these kinds of interventions will work, whatever modality of intervention you're going to use, is you want to try and define a clinical syndrome that you think is not VAP, that people order cultures for, and you want to tell them not to order when at the at the point of care.
• 26:09
  SPEAKER_00: And in our work, we've essentially found more or less four what we call VAP overdiagnosis syndromes that we believe incent unnecessary testing.
• 26:22
  SPEAKER_00: Anna mentioned two of them already.
• 26:24
  SPEAKER_00: I think far and away the most common one is a fever or a white count without any other clinical or radiographic evidence that that is referable to the lungs.
• 26:37
  SPEAKER_00: So I think if you're just looking for one piece of low-hanging fruit, that specific clinical syndrome right there is probably the biggest bang for your buck.
• 26:48
  SPEAKER_00: Anna mentioned another one, which is isolated changes to endotracheal secretions divorced from systemic uh constitu or constitutional signs or symptoms.
• 27:00
  SPEAKER_00: We see a lot of reasons for that, that Anna mentioned as well.
• 27:03
  SPEAKER_00: And then the two other syndromes for VAP overdiagnosis that we commonly see are acute hypoxic respiratory failure that resolves within six to 12 hours.
• 27:15
  SPEAKER_00: As a general rule, um, pneumonia should not resolve that quickly.
• 27:20
  SPEAKER_00: Any cause of acute hypoxic respiratory failure that normalizes that quickly is far more likely to be from flash pulmonary edema, mucus plug, um, or aspiration pneumonitis.
• 27:31
  SPEAKER_00: And then last but not least, there's a lot of imaging that takes place in the ICU.
• 27:38
  SPEAKER_00: And while it's not the most common reason for VAP di uh uh overdiagnosis, we do see uh respiratory cultures ordered just for new gunk on a chest x-ray that is not accompanied by any other clinical signs to the patient themselves.
• 27:53
  SPEAKER_00: So those four syndromes are what we believe are areas where people don't necessarily need to order respiratory cultures, but commonly do.
• 28:03
  SPEAKER_02: Oh, and this is perfect because actually our clinical consensus guidance that we developed along the way with the Bright Star group uh very much echoed what you just said, in that the core of the guidance is that the patient should have signs that there's a respiratory tract infection with some kind of respiratory signs.
• 28:26
  SPEAKER_02: So in this case, perhaps the ventilator change.
• 28:29
  SPEAKER_02: And we use the word sustained change to indicate it shouldn't be a transient thing.
• 28:36
  SPEAKER_02: Plus, some kind of inflammatory infectious sign, a fever or the inflammatory marker.
• 28:43
  SPEAKER_02: So one of both to suggest there is an infection, and that infection is in the respiratory system.
• 28:50
  SPEAKER_02: Um, and so that's kind of the core of the guidance and what I think makes sense for most patients with pneumonia.
• 29:00
  SPEAKER_01: Well, this is awesome insight.
• 29:01
  SPEAKER_01: And I think really hitting on the ruling out and when you should avoid getting cultures is key.
• 29:08
  SPEAKER_01: And uh also um, and is that the most that's the recent publication you were referencing earlier that we can cite?
• 29:14
  SPEAKER_01: No?
• 29:14
  SPEAKER_01: There's a newer one?
• 29:15
  SPEAKER_01: That citation too.
• 29:17
  SPEAKER_01: Awesome.
• 29:18
  SPEAKER_01: Um yeah, I think whenever you can have guidance that's published that is geared from experts to um clinicians is always helpful.
• 29:27
  SPEAKER_01: And I really appreciate these insights.
• 29:29
  SPEAKER_01: Um you've already touched on a lot of those implementation strategies as well in terms of clinical decision support on feedback, the importance of local champions.
• 29:38
  SPEAKER_01: Is there anything else that you've learned about implementation uh that you want to share on this podcast about implementing the intervention and even sustaining those changes over time?
• 29:48
  SPEAKER_00: You know, um a lot of our work has involved other phases of the diagnostic testing pathway, including the collection and reporting phases.
• 30:00
  SPEAKER_00: And I have been surprised to learn just what a labyrinth uh testing clinical biospecimens is.
• 30:14
  SPEAKER_00: Um and uh so particularly um I think intervening after someone has ordered a test, both for the collection and reporting phase, is a much more complicated undertaking than I initially thought.
• 30:32
  SPEAKER_00: So I'll give you an example of that.
• 30:34
  SPEAKER_00: Um, in some of our work, we've attempted to use BAL cell counts and differentials, somewhat akin to how we use pyuria to obviate the need for uh culturing urine.
• 30:48
  SPEAKER_00: It's a great idea and concept, but the problem is respiratory secretions for some of these patients are really tenacious, and you can't put them through a hemocytometer very well, oftentimes.
• 31:02
  SPEAKER_00: And so the samples will be basically untestable for some patients in in situations like this.
• 31:10
  SPEAKER_00: Um, there's also issues of um the volume of sample that you get to do uh the requisite testing.
• 31:18
  SPEAKER_00: Um, it's very variable, um, especially in um BAL as uh the dilutional effect can be a real thing and probably varies across providers and institutions.
• 31:34
  SPEAKER_00: And so any kind of uh reporting phase intervention that's done for VAP really needs to involve a number of uh stakeholders from requisite uh leadership positions in the clinical laboratory space.
• 31:56
  SPEAKER_01: I 100% agree.
• 31:57
  SPEAKER_01: We have also done a reporting level intervention that was published, I think, last year.
• 32:04
  SPEAKER_01: And it was using an algorithm, a clinical algorithm, to help guide whether or not uh results were reported.
• 32:11
  SPEAKER_01: We definitely uh collaborated with the Clinical Microbiology Lab because that affects their workflow significantly.
• 32:16
  SPEAKER_01: But also, as we were doing that project, we learned how challenging it was to actually implement, because unlike urine cultures, where again this type of research is fairly common, patients here are much more critically ill, much more dynamic, and uh harder to really diagnose, and you have that fear of missing diagnosis.
• 32:36
  SPEAKER_01: So we had to have uh intensivists and infectious disease physicians really reviewing on a daily basis to make sure we were not withholding results that were clinically valuable.
• 32:47
  SPEAKER_00: Yeah, I I'd add one more thing on top of that, Kim, which is um my my initial uh reluctance to engage in some of this work was for precisely that reason.
• 33:00
  SPEAKER_00: Severing ties with ICU colleagues and making them feel that we're we're handicapping their ability to autonomously practice.
• 33:09
  SPEAKER_00: And I think what was surprising, at least in the work we've done in our institution, is how receptive intensivists were to this kind of work.
• 33:15
  SPEAKER_00: I think many of them are frustrated by just the diagnostic morass that is VAP.
• 33:20
  SPEAKER_00: And I think they're eager for something akin to a permission structure to allow them to interpret a culture in a way that makes sense to them, but that they don't feel that they can do under current circumstances.
• 33:34
  SPEAKER_02: Yeah, that's great.
• 33:36
  SPEAKER_02: Both of you talked about ways that basically we can help that frontline clinician get to the optimal management and diagnosis for the patient.
• 33:46
  SPEAKER_02: I think everyone's on the same page there of wanting to accurately diagnose and treat their patients.
• 33:52
  SPEAKER_02: One additional implementation strategy I'll add is engagement of the non-ordering staff.
• 34:00
  SPEAKER_02: So your nurses and respiratory therapists.
• 34:03
  SPEAKER_02: We have found uh working in our institution and working with the Bright Star uh PICUs that engaging nurses and respiratory therapists were key in making changes happen.
• 34:17
  SPEAKER_02: And often they were eager to participate, eager to be part of the education, to contribute to design or receive, um, be part of the process.
• 34:27
  SPEAKER_02: In some cases, we even have sites where the respiratory therapist became a champion for it and really took on the leadership.
• 34:37
  SPEAKER_02: So I would say uh think about the folks actually put her in the order, but there's many other team members that are also thinking about that patient, critically appraising their signs and symptoms and very actively engaged, particularly in the clinical care setting.
• 34:54
  SPEAKER_01: I'll very much agree.
• 34:55
  SPEAKER_01: Um, I know for sake of time, I'm gonna just kind of move into thoughts about the future.
• 35:00
  SPEAKER_01: So both of your programs really have moved beyond simply describing the problem towards implementing these diagnostic stewardship interventions.
• 35:06
  SPEAKER_01: We've had a great discussion around implementation considerations, clinical considerations.
• 35:10
  SPEAKER_01: Uh, looking forward, what do you see as some major unanswered questions in BAP diagnostic stewardship?
• 35:15
  SPEAKER_01: Um, and can you kind of provide us a little more insight on where you see the future going?
• 35:23
  SPEAKER_02: Ideally, um, in some magic land, we had a more binary or uh less uh a test that was subject to less risk for misinterpretation.
• 35:34
  SPEAKER_02: So perhaps in the future, a combined inflammatory and microbiology test type of situation might be an advantage.
• 35:43
  SPEAKER_02: And I know there's colleagues around the country looking um at uh this question in more detail to perhaps inform uh future diagnostics.
• 35:52
  SPEAKER_02: So I would say that would be one area is improved diagnostic choices.
• 35:57
  SPEAKER_00: Yeah, I think a couple ideas.
• 35:59
  SPEAKER_00: One is um with onset, you know, technological advancements, particularly AI, um I think it is more possible than ever to tell providers what the actual concrete likelihood of pneumonia is in their patient.
• 36:18
  SPEAKER_00: Um and so we can actually provide data to providers at the point of care that, you know, patient X, you're about to order a respiratory culture on the likelihood of pneumonia uh based on uh autopsy-based studies would be this.
• 36:38
  SPEAKER_00: Do you still want to go ahead with the test?
• 36:41
  SPEAKER_00: I think that is uh an interesting way to apply advancing technology to the ordering phase of the diagnostic testing pathway.
• 36:50
  SPEAKER_00: Um, I think as Anna mentioned, um biomarkers are critically important.
• 36:56
  SPEAKER_00: I think people were relatively um enthusiastic about procalcitonin initially, filling that space.
• 37:04
  SPEAKER_00: But as we've seen with the real-world implementation of that test, it really has not proven to be of widespread utility, particularly in ruling in or out the diagnosis of HAPFAP at the point of care.
• 37:17
  SPEAKER_00: Um, and then I think lastly, again, a lot of the work in this field, I think, has been done in the ordering phase.
• 37:25
  SPEAKER_00: And when providers interpret respiratory culture tests, it very much is a binary thing to them.
• 37:32
  SPEAKER_00: Is there a germ or not?
• 37:34
  SPEAKER_00: But there's really a bounty of information on a respiratory culture that could conceivably be used to inform diagnostic stewardship practices.
• 37:44
  SPEAKER_00: So, for example, on gram stain, does what grows concord with what was seen on gram stain and the bio burden of organisms, the semi-quantitative bioburden of organisms on gram stain?
• 37:56
  SPEAKER_00: Are there leukocytes in there?
• 37:58
  SPEAKER_00: It's pretty darn hard to have pneumonia if you're not, you know, neutropenic on chemotherapy without a neutrophilic alveolitis.
• 38:05
  SPEAKER_00: That's evident.
• 38:06
  SPEAKER_00: Um, and then the actual amount of the germ that grows in the respiratory cultures is relatively important too.
• 38:12
  SPEAKER_00: We see a lot of providers here who technically get a positive culture, but it's, you know, less than whatever the quantitative or semi-quantitative threshold is based on the level of the culture, whether it's an endotracheal aspirate or more distant or more distal.
• 38:27
  SPEAKER_00: And even if it's just a couple of germs, they're going to treat it regardless.
• 38:30
  SPEAKER_00: So I think there's also a lot of interventions that haven't been explored using information that's routinely available on the test result itself.
• 38:39
  SPEAKER_01: So that's great insight from both of you.
• 38:41
  SPEAKER_01: Um, before we wrap up, I just want to make sure.
• 38:44
  SPEAKER_01: Is there any other um tidbits of information or insights you want to share through the podcast to wrap things up before we end?
• 38:54
  SPEAKER_00: If anyone is listening to this and is interested in this work and would like to participate in a multi-center study of something like this, uh, you can you can reach out to myself or Anna.
• 39:10
  SPEAKER_02: Life eyes, yeah.
• 39:12
  SPEAKER_02: We um we continue to appreciate and see the value in stewarding these cultures and these tests.
• 39:20
  SPEAKER_02: I'll say molecular tests, not those cultures too.
• 39:24
  SPEAKER_02: Um and ultimately the goal is to get the right diagnosis for our patients.
• 39:28
  SPEAKER_02: So if they need the antibiotics, they get them appropriately.
• 39:31
  SPEAKER_02: But if they can do without them safely, then that would be ideal.
• 39:35
  SPEAKER_02: So uh folks interested in diagnostic stewardship accuracy and please reach out to us.
• 39:41
  SPEAKER_02: We're happy to continue the conversation.
• 39:45
  SPEAKER_01: Excellent.
• 39:46
  SPEAKER_01: Thank you both so much for sharing your experience and perspectives today with us.
• 39:49
  SPEAKER_01: And thank you to everyone for listening and joining us for the SHA podcast.