The September 2024 Research Roundup

Welcome back to another research roundup. This edition doesn’t have any immediate practice changers, but I think there are some very important topics. Enjoy. 

The BroomeDocs podcast: https://broomedocs.com/2024/09/first10em-journal-club-august-2024/

The BroomeDocs podcast on YouTube: (I am on vacation, so this will probably be inserted later.)

In a surprise to no one who reads First10EM, clinical judgment is better than all decision tools ‘for sepsis’

Knack SKS, Scott N, Driver BE, Prekker ME, Black LP, Hopson C, Maruggi E, Kaus O, Tordsen W, Puskarich MA. Early Physician Gestalt Versus Usual Screening Tools for the Prediction of Sepsis in Critically Ill Emergency Patients. Ann Emerg Med. 2024 Mar 25:S0196-0644(24)00099-4. doi: 10.1016/j.annemergmed.2024.02.009. PMID: 38530675

OK, the study has a bunch of flaws that we can get into, but the results are almost certainly true. These authors prospectively asked attending emergency physicians to rate the chances that 2484 critically ill adult patients had sepsis, and compared their judgement to screening tools for sepsis, including SIRS, qSOFA, SOFA, and MEWS. (None of these are actually decision rules for sepsis, which is an obvious fatal flaw.) Physician judgement was better than all the decision tools, both at 15 and 60 minutes. Physician judgement had an area under the curve (AUC) of 0.90 (95% CI 0.99-0.92), beating all of LASSO (0.84; 95% CI 0.82 to 0.87), qSOFA (0.67; 95% CI 0.64 to 0.71), SIRS (0.67; 95% CI 0.64 to 0.70), SOFA (0.67; 95% CI 0.63 to 0.70), and MEWS (0.66; 95% CI 0.64 to 0.69). There are a number of major limitations to this study, which can be read in the main post, but I think the biggest issue is the diagnosis of sepsis. What exactly is sepsis? They defined it based on ICD 10 codes at the time of discharge, but how accurate is that? What if the patient developed an infection later, while in hospital? The other major limitation of this study is that the various scores were only calculated retrospectively, based on chart review, and there was a ton of missing data. On top of that, none of these scores are actually meant to diagnose sepsis – they are prognostication tools – so the entire comparison is somewhat nonsensical. A proper comparison would have been asking clinicians about the patient’s risk of death, not about their underlying diagnosis. With multiple fatal flaws, I think it would be appropriate to ignore this trial completely, but seeing as none of the tools have good evidence to begin with, this trial emphasises that none of them should be used clinically, and that we should stick with clinical judgement, which is actually quite good for the diagnosis of sepsis. Mostly, I think the point of this trial is that these tools should not be used retrospectively to judge the actions of doctors. I, of course, think there are lessons here that can be applied to all decision rules. 

Bottom line: These decision instruments ‘for sepsis’ have never been shown to improve upon clinical judgement or improve patient care. Until we have evidence of benefit, they shouldn’t be used, and we should rely on clinical judgement, which this trial demonstrates is quite good. 

Do you still suggest emergency medicine to your students?

Rosenberg H, Syed S, Atkinson P. CJEM debate series: #StillTheOne-while more challenging than ever, emergency medicine is still the best career path available for medical students. CJEM. 2024 Jun;26(6):381-385. doi: 10.1007/s43678-024-00682-8. Epub 2024 Apr 9. PMID: 38592662

I have personally been back and forth on this debate over the years – not just about emergency medicine, but medicine in general. Would you still recommend this career? I liked this short pro/con piece, but this type of debate isn’t ideal for a short summary, so if you are interested, I would suggest reading it yourself. These days, we are all intimately familiar with the many negatives of emergency medicine. The biggest issue, which is well laid out here, is that the majority of our problems are outside of our immediate control. We can’t fix the broken primary care system that drives too many patients into our safety net, nor the increasingly complex patients that can’t get an appointment with their specialists, nor the bedblock or medication shortages that make our jobs so much harder. The lack of agency is a massive driver of burnout in emergency medicine. On the flip side, I think it is important to keep the benefits of our job in mind. Emergency medicine is a fascinating specialty. You will see something new every shift. You will never get bored. (Can you imagine spending 30 years performing a repertoire of just 3 primary surgeries?) Our jobs are flexible. We get to leave our work at work. Medicine isn’t easy (it isn’t supposed to be), but it can definitely be rewarding. (I also think we tend to underestimate the frustrations that everyone else faces in their own jobs, even if they are just sitting at a desk selling ads.)

Bottom line: Would you be happy if you child follow-ed your footsteps into emergency medicine? I think it is important to be honest about the realities of our job for students considering the profession. We don’t do any favours by ‘selling’ the specialty. But I would still choose this path again for myself, so have no problem recommending it to others.

Are you being controlled?

Iserson KV. Reflexive control in emergency medicine. Am J Emerg Med. 2024 Jul;81:75-81. doi: 10.1016/j.ajem.2024.04.037. Epub 2024 Apr 23. PMID: 38677197

This topic might have warranted a stand alone write up. I found this essay fascinating and important, and I think it should be widely read for its implications both within medicine, but also in your day to day life. It discusses ‘reflexive control’, which sounds entirely nefarious and was actually born out of Soviet military research, but is the very simple concept that most human decisions are made reflexively (using heuristics), and that targeted interventions can easily influence those decisions. “Reflexive control incorporates psychological tactics, disinformation, and strategic messaging designed to induce the target to make decisions that serve the controller’s interests.” It sounds like sci-fi level mind control, but I think that anyone who follows politics in 2024 is aware of the profound influences of disinformation and psychological tactics. 

Reflexive control can be used for both good and bad. Medical educators use these techniques all the time when we simplify complex concepts into shorter slogans, such as “chest pain plus” or “scalpel, finger, bougie”. These educational memes spread and broadly influence the heuristics emergency physicians use to make decisions, and there is probably an overall benefit. (Although, I have also spent a lot of time illustrating examples that are not evidence based and probably cause harm.) Conversely, there are many examples of reflexive control that are definitely detrimental. This isn’t all about external control. Many examples come from within medicine, where heuristics are shaped by groups spreading concepts like “time is brain” or “time to antibiotics” based on misguided goals. (These examples often start as well meaning campaigns within medicine, but will predictably be taken advantage of by external players with skin in the game, if they can find any way to piggyback onto the issue.) Of course, the external influences are probably most damaging, as the large companies who make billions of dollars of our daily decisions can also afford to invest in psychological manipulation. An even more dystopian future is on the horizon, where these reflexive controls are not just targeting our personal heuristics, but also undermining the entire AI infrastructure that will eventually invade all of medicine. 

We like to believe that our heuristics are refined over time based on clinical experience, but I imagine we all overlook the many non-clinical contributors to our decision making. I mean, none of us really want a crappy pen, but we know that you will change your prescribing pattern if a pharmaceutical company gives you one. 

What are the keys to effectively managing reflexive control in medicine? “Navigating the complex healthcare landscape requires ED leadership to actively counter external pressures and prioritize patient well-being.” We did a really good job getting rid of drug company sponsorship, at least for a little while, in emergency medicine in the places I work. However, I have definitely noticed that influence creeping back in. I think the impacts are exactly as bad as you would expect from allowing people with a vested interest to pitch expensive products to you under the guise of education. There is no such thing as a free lunch. You are influenced by a free pen. Our little human brains cannot handle the massive marketing departments of the pharmaceutical companies. You need to eliminate the influence altogether. 

They close with a short list of methods for emergency physicians to avoid being manipulated by reflexive control, which you can read in detail in the article, but I will reproduce short form here:

1. Use valid, independent information sources, particularly evidence based medicine
2. Be alert to educational bias
3. Beware of ghostwritten medical articles
4. Use clinical guidelines with caution
5. Avoid incurring reciprocal obligations to industry
6. Beware of deceptive and unethical marketing tactics
7. Keep industry relationships transparent
8. Exercise caution with exaggerated claims and off-label marketing
9. Advocate for independent research funding
10. Be leery of industry-funded patient advocacy groups

Turn down for what?

Adams C, Walsan R, McDonnell R, Schembri A. As loud as a construction site: Noise levels in the emergency department. Australas Emerg Care. 2024 Mar;27(1):26-29. doi: 10.1016/j.auec.2023.07.004. Epub 2023 Jul 31. PMID: 37532590

This is a simple study that placed noise monitors at 6 locations in a Sydney, Australia trauma center. The average noise level over the 24 hours was 56 decibels (dB). As a comparison, apparently the WHO recommends that noise levels should not exceed 35dB during daytime hours. The waiting room and treatment areas were slightly louder, averaging 60 dB, but even the “overnight stay area” averaged 50 dB. Peak noise levels get worse, hitting 100 dB in the waiting room and ambulance bay. They chose to look at a ‘typical weekday’, but my experience is that Saturday nights are particularly bad for patients screaming at each other and really ramping up the volume. From a workplace safety standpoint, these levels aren’t that bad. (This study would benefit tremendously from some comparisons to noise levels in other settings. For example, what is the noise level of the airport hotel room I tried to sleep in the last time I traveled?) They are a long way from the 85dB average which seems to be the threshold for needing hearing protection. The results confirm the futility of using stethoscopes in the emergency department, but otherwise probably don’t have much clinical relevance. Most of the conversation I have seen about this paper has focused on clinicians, but in an era where patients are forced to sleep in emergency department hallways for days at a time, I think this paper needs to be considered in a human rights framework. My understanding is that sleep deprivation using loud noise and bright lights have been used as a torture mechanism, and hospital food may not be much better than prison food. The emergency department isn’t a good place to sleep at the best of times. Can you imagine being forced to sleep in a hallway at your lowest moment of health? Is there any chance that this paper will be the one to finally wake up our administrators to the inhumane conditions they force upon our patients?

Bottom line: Emergency departments are noisy, although not ridiculously so.

Beneficial Effect of Calcium Treatment for Hyperkalemia Is Not Due to “Membrane Stabilization”

Piktel JS, Wan X, Kouk S, Laurita KR, Wilson LD. Beneficial Effect of Calcium Treatment for Hyperkalemia Is Not Due to “Membrane Stabilization”. Crit Care Med. 2024 Jul 24. doi: 10.1097/CCM.0000000000006376. PMID: 39046789

This paper is really only for people who are either very nerdy or very pedantic. When asked why we give calcium for hyperkalemia, most of us say “membrane stabilization”, even if we have no idea what that really means. It turns out, based on this lab based study using a canine myocyte model, that isn’t correct at all. They used a section of ventricle, and soaked it in either normal potassium (4mMol), high potassium (8-12 mMol), or high potassium plus calcium (3.6 mMol). Hyperkalemia slows conduction velocity and elevates the resting membrane potential. Calcium has no effect on the resting membrane potential (in other words it does not “stabilize” the membrane), but it does restore conduction velocity, which reverts the ECG from wide to narrow complex. Why would you need to know this? Obviously you don’t. It is completely irrelevant. I have just wasted your time. You give calcium because you are hoping it saves someone’s life, not because of anything it happens to do to the ECG. Perhaps there is an underlying lesson about surrogate outcomes in evidence based medicine there, but really this just tells you how much time I spend reading very pointless science.

Bottom line: If you want to be pedantic, calcium doesn’t not ‘stabilize the cardiac membrane’, but instead restores normal cardiac conduction velocity. However, voicing that bit of information will probably have you marked as a social pariah. 

More on peripheral vasopressors

Munroe ES, Heath ME, Eteer M, Gershengorn HB, Horowitz JK, Jones J, Kaatz S, Tamae Kakazu M, McLaughlin E, Flanders SA, Prescott HC. Use and Outcomes of Peripheral Vasopressors in Early Sepsis-Induced Hypotension Across Michigan Hospitals: A Retrospective Cohort Study. Chest. 2024 Apr;165(4):847-857. doi: 10.1016/j.chest.2023.10.027. Epub 2023 Oct 26. PMID: 37898185

This is low quality data that continues to demonstrate the somewhat sad fact (to someone like me who likes procedures) that it is perfectly OK to run vasopressors through a peripheral IV. It is a retrospective look at adult sepsis patients from 29 hospitals in Michigan, USA, who have data collected as part of a sepsis initiative. They looked only at those patients who were hypotensive and received vasopressors within the first 6 hours. The overall sepsis cohort had 7,039 patients, 549 of whom received vasopressors and are included in this study. 67% of vasopressors were initiated through a peripheral line. 12 of the 29 hospitals had policies requiring or recommended central lines for vasopressors, and care seemed to be stratified upon hospital lines (with peripheral vasopressor use ranging from 28% to 94% in different hospitals). This is what is sometimes considered a real world experiment, in that patients were ‘randomized’ to their treatment group based on hospital rules. Unfortunately, hospitals that require central lines for vasopressors by definition don’t follow evidence based medicine, and so we might assume that they are not following evidence in other areas of the patient’s care, and might have lower quality of care overall. On the other hand, the use of peripheral vasopressors seems to push people away from the standard care use of norepinephrine, although that might be because this data also includes the use of push dose pressors. For their primary outcome, there were no clinical differences in any clinical outcomes when comparing peripheral and central vasopressors, although the absolute differences were quite large in favour of peripheral (a 10% absolute lower rate of in hospital mortality). Complications were rare, but actually higher with central lines. There were no cases of necrosis or tissue ischemia from peripheral use, but 0.7% of patients had to have a central line removed due to complications. Most peripheral patients did have a central line placed eventually, but consistent with previous data, 33% managed to avoid a central line altogether.

Of course, we have known this for a long time, but I guess some places still struggle to catch up to the evidence.

Bottom line: If a patient needs a vasopressor, just start it through your peripheral IV, assuming you have one you trust. For the most part, central lines can be left to the ICU team during day time hours (and a good percentage of patients will never need one). 

Might PPIs actually have a role in critical care?

Cook D, Deane A, Lauzier F, et al. Stress Ulcer Prophylaxis during Invasive Mechanical Ventilation. N Engl J Med. 2024 Jul 4;391(1):9-20. doi: 10.1056/NEJMoa2404245. Epub 2024 Jun 14. PMID: 38875111 NCT03374800

This is not exactly core emergency medicine, although that often depends on how bad your ICU overcrowding is, and so I sometimes end up going down critical care rabbit holes. This is a large multicenter RCT with some very famous names in the author list asking the question of whether PPIs are useful for stress ulcer prophylaxis in intubated ICU patients. The trial included 4821 adult patients who were undergoing mechanical ventilation in the ICU, and randomized them to pantoprazole 40 mg IV or placebo. Their primary outcome was clinically important upper GI bleeding, which had to be clinically overt and result in either hemodynamic compromise or therapeutic intervention. Both the definition of hemodynamic compromise and the need for intervention are very subjective, which could be problematic, but that is limited by the fact that this is an appropriately blinded trial. Clinically important upper GI bleeding occurred in 1% of the PPI group and 3.5% of the placebo group (ARR 2.5%, 95% CI, 1.6 to 3.3, p<0.001) Mortality was not statistically different (29.1% vs 30.9%, p=0.25). Of course, the big problem is the “clinically important upper GI bleeding” may or may not be all that clinically important. The big thing that jumps out at me is that the median number of red cell units transfused was identical (and zero) in both groups. However, that number is probably skewed because of the low rate of GI bleeding, and in the appendix it does seem like 50-80% of these GI bleeds were real, as defined by a significant BP drop, a 20 point drop in hemoglobin, or the need to transfuse at least 2 units of PRBCs. They have a secondary outcome called “patient-oriented GI bleeding” but I am still not sure it is all that clinically relevant. On the other hand, there were no harms identified here (pneumonia and C diff were unchanged), so even very minor benefits might outweigh harms. There are also some changes between the final manuscript and the registered protocol, and although they might seem trivial on their surface (such as measuring outcomes at 60 versus 90 days), changing methodology on the fly increases researcher degrees of freedom, and is exactly how p-hacking occurs. 

Bottom line: It doesn’t really affect my day to day practice, the true benefit is questionable, and I am nervous about protocol changes, but it is likely that PPIs will be the standard for stress ulcer prophylaxis in the ICU, and that seems fine to me. 

There is more on this topic in PulmCrit. 

A clinic for patients with persistent symptoms without clear pathology

Burton C, Mooney C, Sutton L, et al. Effectiveness of a symptom-clinic intervention delivered by general practitioners with an extended role for people with multiple and persistent physical symptoms in England: the Multiple Symptoms Study 3 pragmatic, multicentre, parallel-group, individually randomised controlled trial. Lancet. 2024 Jun 15;403(10444):2619-2629. doi: 10.1016/S0140-6736(24)00700-1. PMID: 38879261  

I found this study interesting, both because persistent unexplained symptoms are difficult for us to manage in emergency medicine, but also because this study has many of the fundamental flaws that are used in ‘alternative medicine’ research to demonstrate ‘benefit’ even when benefit is scientifically impossible. It touches on the broader issue of the acceptability of placebo in medicine, although that depends on how you define placebo, as compared to ‘therapeutic interaction’. They enrolled adult patients with persistent physical symptoms (also known as functional disorders) from 108 general practices in the UK, and randomized them to either usual care or usual care plus a symptom clinic intervention that involved 4 consultations focused on making sure the patient was ‘fully heard and validated’ about their symptoms. They label this approach “REAL” for recognition, explanation, action, and learning. Selection bias will be an issue, as almost 7000 patients were invited to participate, only 1250 responded, and only 354 were enrolled. Patients choosing to respond to a trial of this sort are probably more likely to respond to therapy than those who do not respond. There was a statistical improvement in symptom burden at 52 weeks, as measured by the PHQ-15 questionnaire. Both groups started at about 15. Usual care ended at 14 while the treatment group was around 12. I am not overly familiar with this score, but a quick search seems to indicate that the minimum clinically important difference is 3, so this doesn’t actually look like a clinically important difference despite being statistically significant. They also measured 11 other scores as secondary outcomes, and most of them showed no difference. 

The key problem with this trial is that it is unblinded and just compared to usual care, which undermines results when your primary outcome is completely subjective. This is the same problem, or exploit, that is often used by ‘alternative medicine’ research. The intensive act of being heard may be all these patients need, but is indistinguishable from placebo, and so does nothing to say that this specific approach is best. A much better study would have compared multiple active therapeutic options. Another option would have been to compare this REAL approach provided by the patients’ regular doctor as compared to this specific clinic, to see if the second opinion is helpful. There were many ways that this trial could have provided valuable information, but I don’t think this design does it. This just says that patients with long term unexplained physical symptoms need to be heard, validated, and provided with plans to manage their symptoms. We already know that, even if many family doctors struggle to find the time, compassion, and validation these difficult patients required. . 

Bottom line: I would love to have a ‘symptom clinic’ to refer these patients to, but this trial is very unconvincing, with the hallmarks of bias often present in pseudoscience. Additionally, despite reporting a statistical difference, it isn’t clear that there is a clinically important difference in symptoms. 

Cheesy Joke of the Month

A guy walks into his psychiatrist’s office completely wrapped in saran wrap, but wearing nothing else.

The psychiatrist says, “well, I can clearly see your nuts”.