Aching Eyes and Ghosts from the Past

This month we jump into the HaVoc filled world of cannabis hyperemesis syndrome, wade through the ED management of hyperglycemia, discuss ghosts from the past and learn about a very unique approach to wearing contact lenses.

Take Post Test https://cmecourses.som.uab.edu/mod/quiz/view.php?id=6036
Aching Eyes and Ghosts from the Past
Featuring:
Matthew Delaney, MD | Charles Khoury, MD
Learn more about Matthew Delaney, MD 

Learn more about Charles Khoury,MD 

Disclosure Information:
Planners:
Ronan O’Beirne, EdD, MBA
Director, UAB Continuing Medical Education
Katelyn Hiden
Physician Marketing Manager, UAB Health System
The planners have no relevant financial relationships with ineligible companies to disclose.


Release Date: November 7, 2022
Expiration Date: November 6, 2025

Faculty:
Matthew Delaney, MD
Associate Professor, Emergency Medicine
Charles Khoury, MD
Associate Professor, Emergency Medicine

Drs. Delaney and Khoury have no relevant financial relationships with ineligible companies to disclose. There is no commercial support for this activity.



Transcription:

Dr. Matthew Delaney: Welcome back to episode three of EBM Breakdown. I am Matthew Delaney, joined as always by the one and only, Dr. Charles Khoury. And Charles, we've got a great month. We've got a big month on tap.

And I want to start out with a story about a friend of mine and you've met some of my friends and I would describe my friends group in general throughout my life as a bit eclectic. There are more than a few wild cards in my friend group. But one of the guys that I'd known for a while who's actually more reasonable told me a story that's really, really wild. So, he had corneal abrasions back when he was in college, and he just mentioned -- he's not medical -- he just mentioned to me like, "Yeah, those hurt really bad." I was like, "Oh, I didn't know you had that." He's like, "Yeah, I played hockey. And I went to like an end of season party and I was a bit overserved. And so, I went back to my dorm. And what I remembered my mom telling me before I went to college is that always take your contact lenses out each night. And that was the thing that I remembered I had to do. And so I spent a really long time trying to take my contacts out and it turns out that they weren't in my eyes to start with. And so, what I did is just destroyed both of my eyeballs. And when I slept it off and woke up the next day, it was unbearable. I was miserable." And he went to student health and saw a doctor who I'd love to meet because the doctor just looked and said, "Yeah, this is really messed up. I'm going to write you for a bunch of codeine syrup. And I want you to just drink all of this and then go home and sleep until your eyes heal." And that was this kind of management strategy, which we're just going to drink a bunch of codeine syrup, sleep and let this be a self-limited condition, which is a bold plan. It's certainly not something that I would recommend or prescribe for a patient. But I think the point is well put that this is a really painful condition where we don't have a lot of good options in terms of getting the patient comfortable.

Dr. Charles Khoury: Yeah. I kind of have my own contact lens story, and I was one of those people who took my contact lenses out every single night. I never missed a night, but I got contact lenses the day I went into seventh grade and never wore glasses again since that time. And I would take them out every single night, but I used the same contact lens case every day from the time I was in seventh grade until I was an attending physician. Now, I cleaned it every single day. I washed it with soap and water every day. It was immaculate. But it was a $4 piece of plastic and it became sentimental to me. It was the same contact case that got me through junior prom and the MCAT and medical school exams, and I used that thing until I got LASIK a couple years ago.

Dr. Matthew Delaney: Where is that thing now?

Dr. Charles Khoury: I still have it and I keep it next to like graduation certificates.

Dr. Matthew Delaney: It's like Dumbo's magic feather, like there's no Charles Khoury without this wonderful 25, 30-year-old contact case. So, we're talking contacts, what not to do with contacts and corneal abrasions. And one of the topics that's come up has been why can't we use topical analgesics for patients who have corneal abrasions, right? We're comfortable using them at the bedside when we're making the diagnosis and doing the exam. But I was trained that these will melt your corneas, these will make you blind. And so, I actually used to make a show of it with the patients where I'd put them in and say, "You feel a lot better," I'd throw the bottle in the trash can and say, "But it's too dangerous for you to go home with. You can't go home with these." And number one, that bit is not very successful. It just kind of makes the patient frustrated. But also, that's not that scientifically valid. And in fact, if we look at the optha literature, ophthalmologists have been doing corneal procedures for years and actually sending patients home with short-term topical analgesics. Yet, despite the fact that the eye doctors are doing this, I think most ED docs, if you put a gun to our heads and say, "Can we send patients home with topical analgesics?" Most folks are going to say, "I don't think that's a good idea."

Dr. Charles Khoury: Yeah. Emergency physicians and ophthalmologists I've seen on medical Twitter get into wars about this, about whether or not you should do topical analgesics for corneal abrasions. And genuinely, the biggest fight I've seen on medical Twitter was related to this and the literature from the 1970s. So, there's not a ton of data on this. If you've ever had a corneal abrasion, you know that the life-saving drops are the tetracaine. But in this particular study, this was a single-center, prospective, double-blinded, placebo-controlled, randomized trial of adult patients who had uncomplicated corneal abrasions.

So, this study was a single-center, prospective, double-blinded, placebo-controlled randomized trial of adults with uncomplicated corneal abrasions. They looked at about 120 patients between 2015 and 2017, and the good news is all of these patients did receive antibiotic drops. They received polymyxin B trimethoprim sulfate with instructions to instill two drops every four hours into the affected eye. And these patients also received hydrocodone acetaminophen. There was the 7.5 milligram hydrocodone, about 12 tablets of them, and they were instructed to use these one to two tablets as needed for breakthrough pain. And then, patients were randomized into this one-to-one group of whether they received topical tetracaine, the 0.5% solution, which they were told to do one drop every 30 minutes as needed for 24 hours versus this topical placebo one drop every 30 minutes as needed for 24 hours, so it was really just an artificial tear solution.

Dr. Matthew Delaney: In terms of results, the tetracaine just worked. So, the baseline pain at the start in both groups was seven. And if you look at 24 to 48-hour pain scores, if you got tetracaine, you had one out of 10 pain. If you had placebo, your pain was eight out of 10. So, an absolute difference of seven. This was statistically significant. And if you look at other things, like how often were they using hydrocodone, in the tetracaine group, on average they took one tablet, whereas in the placebo group they took seven tablets. Complication rates, and these were not dramatic complications, but they were 3.6% in the tetracaine group versus 11 in the placebo group. And one of the fears about using these topical analgesics is it's going to prevent the cornea from healing. So, they looked at the presence of a small residual corneal abrasion on a repeat slit lamp exam, and they found that the tetracaine group was 18% and the placebo group it was 11%. Those numbers are different, but not statistically different, so same, same. The only difference was if you got tetracaine, you just actually felt better.

Dr. Charles Khoury: And in the study, additional drops were actually collected and disposed of because people were so worried about using too many drops, being numb for too long. The other major concern is that honestly, if your eye's completely numb, you may scratch it again and you wouldn't even notice. So, I can understand this initial fear of tetracaine.

Dr. Matthew Delaney: Yeah. I think one of the keys here is that tetracaine seems to be certainly effective and I would argue is safe when used for a short period of time. And if you pull the case reports about when have topical anesthetics hurt people's corneas, it's crazy stuff where they're doing two drops every 15 minutes for five, six days. And I think if we limit it to 24 to 48 hours, we're going to help patients feel better without exposing them to risk.

Now Charles, one of the things that's interesting is in that kind of fight on Twitter, there's been a lot of noise about what strength of drops should you give. And in this study, they used tetracaine 0.5%, which is basically what you get out of Pyxis. Now, there are a lot of online guides to, "Okay. Take that, dilute it down by a factor of 10." And there are some ED-based studies that have shown that that approach of diluted down seems to still provide pain relief. But again, major risk seems to be from overuse. So, I think the concentration doesn't seem to matter. And in fact, we have optha studies where they give stronger concentrations than 0.5% and don't see adverse events. So, I think that this is a very reasonable thing to do. We can give them the drops out of the Pyxis and say, "You can use these, but only for a short period of time."

Dr. Charles Khoury: Yeah. So, the bottom line is that 24 hours of topical 0.5% tetracaine for uncomplicated corneal abrasions reduced pain scores at 24 and 48 hours, and was opioid-sparing for breakthrough pain compared to the placebo. Larger trials would be needed to rule out rare adverse events, but adding the study to previous study data indicates that, you know, 24 to 48 hours of topical anesthetic use for uncomplicated corneal abrasions is probably safe.

So, I want to talk about something a little more mundane to us. Let's switch gears a little bit and talk about hyperglycemia. And I don't know about you, but I always struggle to figure out exactly what to do with these patients who come in for one complaint. As I'm doing a workup, I find that they're hyperglycemic. It's not unusual for me to see a patient with a headache or with some unrelated belly pain, and then you find out their glucose is greater than 400. And we've looked at studies in the past that addressed whether or not asymptomatic hyperglycemia should be treated in the emergency department. But what we haven't really talked about is this gray area where a patient's symptoms could be affected by their hyperglycemia. So, you are still kind of hoping to bring it down. But in your quest to bring it down, how do you know the best dose of insulin? These previous studies have looked at what happens when we give these patients insulin, and generally found that as insulin doses increase, patients do have better blood sugars, but they also have increased lengths of stay.

Dr. Matthew Delaney: Right. When I worked in Maine, I swear to you, we never saw diabetes. We had plenty of other conditions, but the incidence of diabetes was nothing like it is down here where I currently practice. And so, I do struggle with this. Patients come in with an ankle sprain, their sugar's, you know, 490. I feel like I should do something. Again, we have to balance that and we want to tie up resources.

And so, this paper I think is interesting. It's by Koscal, et al., it's called Intravenous Insulin for the Management of Non-Emergent Hyperglycemia in the Emergency Department. This is from the American Journal of Emergency Medicine, July of 2021. And the clinical question here is, what is the impact of IV regular insulin glucose reduction and ED length of stay in patients who show up in the ED with non-emergent hyperglycemia?

Dr. Charles Khoury: This was a single-center, retrospective, chart review study that assessed adult patients 18 and over who received IV regular insulin for the treatment of hyperglycemia, and they were discharged from the emergency department at one large academic level one trauma center between the years 2015 and 2018. Patients were excluded if IV insulin was ordered, but discontinued prior to administration, obviously, if subQ insulin was administered, if the serum potassium was greater than five, if the initial blood glucose was less than 70, if admission lab values were not obtained, if they were in DKA, if they died during their emergency department visit and if they were admitted to the hospital. So, they're looking at pretty healthy, simple hyperglycemic patients who came into the emergency department.

Dr. Matthew Delaney: So, the primary outcome was two things. ED length of stay and blood glucose reduction between the two groups. So, high dose was you got greater than five units of regular insulin and low dose was less than or equal to five of IV insulin. And then, a couple secondary outcomes, so basically, are there other factors associated with your blood glucose reduction? They looked at what happened to your potassium and then, the percentage of patients that developed hypoglycemia, which for this study was a sugar of less than 70. And they also looked at hypokalemia and set it at potassium less than or equal to 3.5.

Dr. Charles Khoury: They started with a total of 1,770 patient encounters identified during the study period. And of these 1,770, 1365 were excluded. So, that resulted in just 405 patients included in this analysis. Of these 405 patients with hyperglycemia in the emergency department, 209 of them or about 52% of them received low-dose insulin, so less than five units. And then, 196 of them or about 48% of them received high-dose insulin, which they defined as greater than five units IV. They were managing a median glucose of about 447, which is actually a little higher than I would've expected for a study like this. And when they compared low dose and high dose insulin administration, they found that the ED length of stay was fairly similar between the cohorts. It was a median time around 240 minutes or three hours for both.

However, the use of high-dose insulin was associated with greater blood glucose reduction, so 213 versus 158. They also noted that folks who were getting greater than two liters of IV fluids had an increased length of stay and also had more of a drop in their blood sugar. Better renal function, GFR greater than 60, was associated with a smaller drop in blood sugar.

And here's a couple more interesting findings. The incidence of hypokalemia, low potassium, was 8.1% in the low-dose group versus 7.7% in the high-dose group. So, there wasn't actually a statistically significant difference between the two groups with regards to potassium levels.

So, to make a long story short, what they're saying is that the use of high-dose insulin didn't really increase the length of stay versus low-dose insulin, but it did seem to have tighter glucose control, but that giving more IV fluids obviously increased greater length of stay.

Dr. Matthew Delaney: Yeah, I mean that's, I think, helpful if you're someone who's going to reach for insulin in a patient who has marginally or asymptomatic hyperglycemia. You know, Charles, I don't really know what to do with this paper because when I see a patient who's got this incidental or this kind of marginal hyperglycemia. I have two goals. One is I want to make them feel better. So, if they're having polyuria or polydipsia, that's when I'll reach for insulin or fluids. But if they're asymptomatic, then that goal goes away.

And the second is I want to come up with an outpatient plan. And so, my real goal of treatment is not to hit any magic number. But I want you to get a blood sugar that your home meter can read. And for most patients, that's going to be about 500. And so beyond that, I'm not sure that insulin dosing for me really matters that much. I think it's nice to know that I can give more insulin from a length of stay standpoint. But I would actually challenge anybody that's listening to this who loves to correct to a certain number, to really think about, is that really something we need to do in the ED for this patient?

Dr. Charles Khoury: Well, you know, you make a really good point. But I struggle with whether or not high glucose will affect certain conditions. So if you present with abdominal pain and maybe you incidentally have hyperglycemia, is the hyperglycemia related to your abdominal pain? What about headaches? If you come in with a headache and your blood sugar is high, is it that maybe you're dehydrated and you need some more fluids, so that you no longer feel a headache? Or is it that incidentally you had a tension headache and you also had hyperglycemia? So, I think it's pretty simple with the ankle sprains, that we know that treating hyperglycemia acutely in the emergency department is not advised. There's really no point in doing it if someone comes in with an ankle sprain. But what about these conditions that maybe have something to do with the symptoms?

Dr. Matthew Delaney: Charles, I want to shift and talk about cannabis. And I've been reading up on this over the past couple of months, and it all started after a conversation with a patient, and this was a patient who had a chronic, painful condition. They had moved to a state that had recreational cannabis in additional to medicinal cannabis, so they had ready access to this. And whatever they were buying in this state, when they came back here to me, they said, "Hey, this really felt good. My symptoms were a whole lot better." As a doctor, what do you think?" And I thought that that was a simple enough question. And I went to the literature. And I'll tell you if you want to be confused as a doctor or as a clinician, just try to read cannabis literature. They're talking about different strains. There are different blends. There's terpenes. There's all this stuff used to describe cannabis that I can't find in any of our textbooks. And I found really the whole world of cannabis literature to be just truly puzzling. It was almost like the first time I read Lord of the Rings. I just couldn't really follow a lot of what they're talking about.

Even moving past my difficulty understanding what the authors are talking about, I'm left with one question about cannabis that I just can't seem to find any reasonable answer for. When did cannabinoid or cannabis hyperemesis syndrome become a thing? I mean, we both graduated in 2011 and this just kind of like appeared a couple years after that, and I don't think people just started to use cannabis in the mid like 20-teens.

Dr. Charles Khoury: Yeah, it is a phenomenon that I've seen more and more. And interestingly enough, it's more publicized in the literature in states where marijuana has become legalized. So, you know, one of the effects of having legalized marijuana is that more people are using it and using it regularly. And also, more people are probably presenting to the emergency department because of these complications.

Dr. Matthew Delaney: I see two flavors of this. One is the patient who comes in with belly pain, has nausea and vomiting from, you know, who knows what, and they just happen to have cannabis in their drug screen that the hospitalist orders.

Dr. Charles Khoury: And to be clear, we are not talking about those patients here.

Dr. Matthew Delaney: No, that's just a hospitalist nervous and looking for a diagnosis. What we're talking about with this study is the patient who is filling the emesis basin with noise. They've thrown up everything they've got. They're just putting demons out of their body and they report, "Hey, look, I do use cannabis regularly." And usually, there's some type of chronicity to this. It's rarely that this is a one-off for the first visit.

And this is an interesting paper. It's by Ruberto, et al., Intravenous Haloperidol versus Ondansetron for Cannabis Hyperemesis Syndrome, it's called the HaVOC Trial -- cool name -- in Annals of Emergency Medicine in 2020. The clinical question is very simple. None of our current tools seem to really fix this problem and so, maybe haloperidol works. And it's a very clunky, complicated study that ultimately didn't work, but does give us some helpful information.

So, it was a randomized, triple-blind crossover trial with up to three treatment periods per subject. And they were randomizing patients to get either ondansetron or they would get different doses of haloperidol and you had to have a working diagnosis of cannabis hyperemesis syndrome. So, they define that as greater than or equal to three episodes of emesis in a cyclical pattern separated by one month during the preceding two years, right? That's the patient we're talking about. They keep showing up. This keeps being a problem. And the patient had to endorse near daily to daily use of cannabis by inhalation, so not edibles, not tinctures, for greater than or equal to six months. And the primary outcome was if you got either haloperidol versus ondansetron, what was the change in the abdominal pain and nausea scores at two hours versus baseline?

And we talked about this being this triple-blind, crossover trial. And the initial goal was we will randomize them when they first show up, and then because this is often a cyclical condition, when they come back, we'll capture that patient and then potentially put them in a different treatment arm. So, they built this huge apparatus to say, "We're going to get a lot of information about what treatments work for these patients." But it didn't quite work out that way.

One of the issues here was just finding patients. So, they had 313 patients who accounted for 570 visits with presumed CHS. Again, this is a chronic condition. Out of that group, 62 patients were eligible, yet only 33% said they'd participate. We'll get back to that. Thirty of those patients got one treatment. And so, when you look at it, 17 patients got ondansetron of eight milligrams, seven patients got haloperidol at 0.1 mg/kg, and they say that's kind of the higher dose, six got lower dose haloperidol at 0.05 mg/kg. But the problem was, again, they had this plan of you have a cyclical condition, we're going to catch you in your next visit. Well, they only caught three down the road. So, what we're left with is a look at a small group of patients who have this hyperemesis syndrome associated with cannabis use. And when they looked at the outcomes for the patients, the mean difference in that primary outcome, the pain and nausea scores between ondansetron and haloperidol was 2.3 centimeters, so two points. And that was a statistically significant difference that did favor haloperidol over ondansetron.

Other things they looked at, so when a patient got haloperidol, there was a higher treatment success rate, 54% versus 29%. The need for rescue antiemetics was much lower. Now, lots of patients needed rescue antiemetics, 60% in the ondansetron group versus 30% in the haloperidol group. Reduced use of benzodiazepines. I'm not sure how often I'm reaching for benzos in these patients. But if you got ondansetron, 41% also got benzos versus 8% who got haloperidol. And then, the rates of needing any other medicine were also much lower in the group that got haloperidol.

Dr. Charles Khoury: Now, I think it's important to mention that this was a really hard study to enroll in. The patients were really skeptical. About half of them were skeptical of the diagnosis itself, and I get it. If you're coming to the emergency department and you're vomiting a bunch and somebody just says, "It's because you're smoking marijuana," that doesn't intuitively make sense to you as a patient. I get it. This is a pretty big limitation on this trial. They were unable to achieve a pre-specified enrollment target. And they had overall a pretty small sample size causing wide confidence intervals in all outcomes.

I mean, think about trying to do this study. You build this complicated apparatus that's going to follow patients over months. And then, in reality you get patients who are likely high who are very frustrated because this is a problem that keeps happening and they can't seem to fix it, and then they're vomiting their brains out. So, I'm surprised they got anybody to agree to be in this study. These are just patients who feel miserable, and you mix in the cannabis and it's just an inherently hard population to study.

Yeah. And I'm sure you'll agree with me in that we have tried a lot of things for these patients. I remember when we were trying to place capsaicin cream on their abdomens, and it would just burn so much that they couldn't tolerate it. But even capsaicin cream has been advocated as a pretty good remedy for this.

Dr. Matthew Delaney: And that kind of gets down to my key takeaway for this paper, is obviously this is a small study. They tried, just didn't get there in terms of being big and with robust findings. But it does bring up the issue of when do we try something because we don't have other good options. And I think that the logistics of this are interesting to me. So, a lot of places I've worked, you are not allowed to give IV haloperidol. Actually, it's a never event. You can get trend tracked if you give it that way. And so, I look at this and would say I think that this is some evidence that haloperidol may be effective for these patients because very clearly the other things we're doing don't consistently work. But if you're thinking about reaching for IV haloperidol, I would say that this is evidence, this is a low level of evidence. And specifically, the friction point at your shop may be the IV use of it. So personally, I've started doing this for these patients. But I'm going to go IM. That's an apples to oranges comparison in some ways, but I think just be cautious about what protocols exist in your world, because this is kind of the tip of the spirit. This is, I think, something we can use, but really isn't quite ready for primetime from an evidentiary standpoint.

Dr. Charles Khoury: So, the bottom line on this, cannabis cessation is the most effective treatment modality, but a one-time dose of Haldol 0.05 milligrams per kilogram may be a reasonable option in patients with suspected CHS.

Dr. Matthew Delaney: We're going to shift gears and talk about heuristics, which number one is a really hard word to spell. Number two is a very hard word to say. Number three, when you brought this paper to me, I had to look up what a heuristic is. But this is really worth talking about, and I think one of the things that we don't talk about enough in medicine and specifically in the show is how we make decisions, why we make decisions. So, we talk about things like clinical decision tools and risk ratification. But if we put it all together just past the tools we have, I think the way that we make decisions or make an assessment is actually really interesting. And we have both worked with folks who are great physicians, but I would say that they're very motivated by anecdote. They'll tell you, "Hey, one time I had a patient with an ankle sprain. It turned out they had aortic dissection." So, now every time I see an ankle sprain, I'll do an ultrasound to look for AAA. And this seems extreme. And it seems kind of crazy, but the idea that we have these self-made rules of thumb or this heuristic is it's widely recognized in the world of cognitive science.

And in fact, one of the common heuristics is this availability heuristic, which is basically the idea that our estimate of an event's likelihood is influenced by how well can we remember a similar event happening previously. And previous studies have shown us that this availability bias is certainly real when it comes to how do we make medical decisions. There's an interesting study that looked at patients who had AFib and who were anticoagulated and had a bleeding event, which we know does happen in patients who are on anticoagulants. And the clinician who had prescribed the medication was less likely to recommend anticoagulation for future patients with AFib.

Again, seeing a complication influence future behavior, there's also a study that looked at OB-GYNs and found that in cases where there was an adverse event during vaginal delivery, going down the road, short-term, medium-term, maybe long-term, the clinicians were more likely to either recommend a different route of delivery or a C-section over a vaginal delivery. And this one makes sense. There was a third study that said that in situations where a report of injury was filed against the physician, whether this was lawsuit or a complaint or a trend track, that going down the road, that clinician was much more likely to order diagnostic imaging after that kind of slap on the wrist or that case was filed. But that was a short-term increase in that situation. And so, these studies have kind of all come together and asked this question of how influential is this availability bias? Do the things we see in the past really change our behavior going forward?

Dr. Charles Khoury: And even from like a primitive standpoint, this makes sense. You touch fire, you have a bad outcome and then you don't want to touch that fire again. And in some ways, we are kind of conditioned to that in medicine, that we have a really bad outcome or we see a difficult diagnosis. And it reminds you of another time that this happened. And I've worked with plenty of colleagues who say that I will scan this patient who presents with this because this one time this happened. And that does make sense almost on a genetic point-of-view where you feel it and for that reason you're going to do it. And so, I really like this paper. Let me intro it for us. It is by Dan Ly and the paper's called The Influence of the Availability Heuristic on Physicians in the Emergency Department. It was published in Annals of Emergency Medicine in November of 2021. And the clinical question they asked was does diagnosing a patient with pulmonary embolism alter emergency department testing, a D-dimer or a CT, in subsequent patient encounters?

Dr. Matthew Delaney: It's a cool study. So, it's a VA-based study. And they looked at pre and post-encounter testing behaviors, testing strategy. They got over 400,000 patient encounters across a hundred VA EDs. There were just over 7,300 ED physicians, and then there were 416,720 patient encounters for shortness of breath. This is a big study. And they looked at the testing strategy pre and post making a diagnosis of PE. And they compared it to the pre and post behavior for pneumothorax. So basically, they're saying if you find a PE and a patient with shortness of breath, what does your behavior look like going forward compared to if you found a pneumothorax? Does that influence your behavior going forward?

And the other thing they did is they also compared it to unrelated other tests. So, they were basically trying to say are you just ordering more in general? So, they looked at your rate of ordering TSH and a hand x-ray, so two totally unrelated symptoms. So, they looked to try to really drill down on does finding a PE really change our behavior going forward?

Dr. Charles Khoury: So, let's look at the pertinent results. Nine percent of these patients were tested with either D-dimer or CT angio, or a combination of the two. And 1% of these patients were diagnosed with a PTE. Then for the next 10 days, after that particular physician made that diagnosis, there was a 1.4% increase in PTE testing, which is about 15% more tests. It went from 9% before to 10.4% after. And this did regress to the means. So, they kind of showed that things went back to normal after about 40 days after you made this diagnosis.

Dr. Matthew Delaney: I really had not thought about the availability bias, probably because I'm too prone to letting it influence my thoughts before we found this paper. And there are a couple of things about the methods here that I think probably understate how big of an influence the availability bias can play.

So, one of the things is that this was a VA-based study, so the vast majority of patients are going to be young, fairly healthy men. And we know that young, fairly healthy men don't have as high risk of PE or DVT, right? They don't take birth control. That's one of the big risk factors here. And so, the incidence of PE in this patient population in this VA-based study, it's actually lower than I would expect to see in the general population. So, they are seeing PE, they're diagnosing PE, they're being influenced by PE. But what if you ran this in a non-VA study? What if you had a bunch of young women on birth control? I think you're going to be potentially much more influenced by making that diagnosis because your patients have a little bit higher baseline risk.

The other thing is that PE is one of the disease processes that I think we are drowning in potential tools when it comes to risk stratification, right? We have Wells', we have modified Geneva, we have PERC criteria. And so, I feel like when I'm looking at a patient with a potential PE. Even if I made that diagnosis recently, I do have a roadmap of how I think about the risk in those patients. If you shift this to something like vague abdominal pain, we don't have scores to tell us what's the chance that this is your appendix. We have nothing that's really nearly as ready for primetime as the PE risk stratification. So, my suspicion is if you take a disease process like weakness or nausea or abdominal pain where you don't have guardrails of really good validated clinical decision rules, I think the availability bias is going to be much more influential. I mean, I remember every time I've diagnosed a appendicitis, that next couple shifts I'm thinking, "This vague abdominal pain, is this an appy? Do I need to scan this patient?"

Dr. Charles Khoury: Yeah, and I want to blow your mind a little bit here. So, this all makes sense, but it's a little more concerning when we think about these things occurring in parallel rather than in series. So, let's say you see a PE and then a few days later you're kind of probably ordering more CT angios. But at the same time, maybe you saw a non-traumatic subarachnoid hemorrhage a couple of days ago, and it's made you order more CTs for headache. And then, there was this really weird presentation of an acute appendicitis the other day, so maybe you're thinking about that in all your patients with nausea and abdominal pain. These things kind of add up over time and they influence your practice in a way that may not necessarily be bad, but it's probably a little evidence-based than we would've liked. And I do see this phenomenon a lot in residents since part of our job is having them explain why they're ordering the tests that they're ordering. And since reading this paper, I know I've noticed a lot more residents will answer the question, "Why'd you order the test with...?" "Well, I saw the patient with this the other day and it kind of freaked me out." And that does make sense. But over time, when you add all these things, when they compound, you find yourself becoming a completely different physician, a different clinician, because of the things you saw in the past few weeks.

Dr. Matthew Delaney: Right. That's the art of medicine. And you always talk about, in medicine, you got to get enough touches. You got to just have enough encounters to really get your gestalt down. And so, I don't want to have the takeaway from this paper be that we should totally ignore previous experience because, again, that's how we learn, that's how we develop this clinical instinct.

But for me, I think the bottom line here is that no matter how I practice, this would suggest that this availability bias is real. And so, the next time I see a case that's alarming or "Oh my gosh, I didn't think I would see a PE in this patient", I'm going to intentionally try to not change my practice patterns, specifically short-term to really try to stay out from under this type of bias.

Dr. Charles Khoury: So, clinical bottom line after seeing a high-risk diagnosis, testing, and concern for this seems to increase transiently. All right. So that's it for this month. Delaney, why don't you give us some overall clinical bottom lines here from all the things we've reviewed.

Dr. Matthew Delaney: If you see a patient who has hyperglycemia in the ED, my first challenge would be, do you actually need to do something for that patient if they're truly asymptomatic? But the paper tells us that if you're going to treat a patient for hyperglycemia, giving higher insulin doses will help drop their sugar without causing increased length of stay.

Dr. Charles Khoury: Okay. Give me another one.

Dr. Matthew Delaney: It's important how we take care of our contact lenses. It's probably not safe to use the same container for decades, but you are living proof that that's not the most dangerous approach out there. But if you've got a patient like my buddy with corneal abrasions, using topical analgesics short-term, so no more than 24 hours, is certainly effective and seems to be safe.

Dr. Charles Khoury: All right. What else you got?

Dr. Matthew Delaney: All right. Cannabis hyperemesis syndrome, I hate it, patients hate it. We don't really fully understand what it is, but basically nothing that we use works. There is a small glimmer of hope from a very flawed study that haloperidol might be an effective option. Just know how you can give it at your shop, because a lot of places cannot give it IV.

Dr. Charles Khoury: All right. Then, take us home with this availability heuristic.

Dr. Matthew Delaney: So, fancy way to say when you see something that freaks you out, your behavior is likely to be influenced. It actually is likely to change short-term the next time you see a similar patient.

Dr. Charles Khoury: All right. That was a good wrap up.

Dr. Matthew Delaney: All right. So, thanks for listening to EBM Breakdown. If you like the show, go to your favorite site to get podcasts. Give us a five-star review. It'd be very helpful to let other people find the show. If you didn't like the show, if you have feedback, questions, concerns, drop us a line. You can find us on Twitter, shoot me an email, but we will be back next month with a next round of papers.

Dr. Charles Khoury: And if one person gives me a contact case in the next month, I'm not coming back.