Research Roundup (December 2023)

Welcome back to another edition of the research roundup, where we discuss an eclectic collection of articles selected through the rigorous process of whatever I happened to find most interesting in my recent reading (with a couple suggestions from Dr. Casey Parker).  The BroomeDocs podcast version can be found here: https://broomedocs.com/2023/12/first10em-journal-club-december-2023/

Our poorly managed hospitals might be killing patients

Roussel M, Teissandier D, Yordanov Y, et al. Overnight Stay in the Emergency Department and Mortality in Older Patients. JAMA Intern Med. 2023 Nov 6. doi: 10.1001/jamainternmed.2023.5961.PMID: 37930696

Everybody in emergency medicine already knows this, and it is very doubtful that any of the people responsible for this problem will ever read this paper, but boarding patients for prolonged times in the emergency department is very bad for outcomes, not to mention being inhumane. This is a retrospective cohort from 97 emergency departments in France looking at patients over the age of 75 who were admitted to the hospital from the emergency department. They compared patients who spent the night in the ED to those who were admitted to the ward prior to midnight. They include 1598 patients, and for their primary outcome of mortality by hospital discharge, mortality was significantly higher among patients boarded in the ED overnight (15.7% vs 11.1%). Baseline characteristics actually look very similar between the two groups, and if anything I would have expected the results to be biased in the opposite direction, with sicker patients rushed to the floor at the expense of the healthier ones. (However, I imagine this depends on the hospital system, as some systems might try to keep sicker patients in the ED where they have a resuscitationist available 24 hours a day.) Adverse events and hospital length of stay were also higher in the group spending the night in the ED. One big difference in these hospitals, as compared to the hospitals where I work, is that patients who remained in the emergency department remained the emergency physician’s responsibility. We have emergency nurses doing inpatient work (in addition to all their emergency responsibilities), but at least admitted patients are mostly cared for by the admitting teams, which could conceivably make a difference in outcomes. Of course, retrospective observational data like this is limited by confounders, but these results fit with multiple previous studies, and have face validity, so I tend to believe them. 

Bottom line: This retrospective cohort demonstrates an association between overnight boarding in the ED and increased mortality among elderly patients. Try printing a few copies of this paper and accidentally leaving them around the hospital’s c-suite.

The first trial of cryoprecipitate in trauma is a bust

Davenport R, Curry N, Fox EE, Thomas H, Lucas J, Evans A, Shanmugaranjan S, Sharma R, Deary A, Edwards A, Green L, Wade CE, Benger JR, Cotton BA, Stanworth SJ, Brohi K; CRYOSTAT-2 Principal Investigators. Early and Empirical High-Dose Cryoprecipitate for Hemorrhage After Traumatic Injury: The CRYOSTAT-2 Randomized Clinical Trial. JAMA. 2023 Oct 12. doi: 10.1001/jama.2023.21019. PMID: 37824155

This is a big study, looking at 1600 adult trauma patients requiring massive transfusion, and randomizing them to either empiric cryoprecipitate or standard care (it was an open label trial with no placebo). The logic is similar to the (potentially flawed) logic behind 1:1:1 transfusions, in that we know some patients develop traumatic coagulopathy, and some people think it makes sense ‘to get out ahead of it’ rather than wait for lab value to fall. (I will point out, as many seem to forget, that 1:1:1 transfusion does not have any good evidence of benefit.) As far as outcomes go, this was a definitively negative trial. The primary outcome was all cause mortality, and it was identical (25.3% vs 26.1%). All secondary outcomes, including functional outcomes, transfusion requirements, length of stay, and location of discharge, were also negative. However, there are a few reasons to be cautious in our interpretation of these negative results. First of all, all of the hospitals were already providing cryoprecipitate empirically for at least some patients, and the addition of more cryoprecipitate may not be nearly as important as the addition of some cryoprecipitate. Second, these hospitals were all using a 1:1:1 ratio, rather than 1:1:2, which has never been proven to help at all, but could influence hemostasis and therefore the impact of fibrinogen. There were also a lot of protocol violations, so the treatment differences between the two groups were not all that dramatic.

Full write up here. 

Bottom line: This is really the first trial looking at this question, so there definitely is no good reason to be providing cryoprecipitate or fibrinogen concentrates empirically in trauma. That doesn’t mean we should ignore fibrinogen altogether. I send levels early and often in massive hemorrhage, and replace when levels fall below 1.5 grams/L. 

Choosing wisely for kids

From the American Academy of Pediatrics and the Canadian Association of Emergency Physicians, we have the Choosing Wisely recommendations for pediatric emergency medicine. (I don’t know if these have been published in a journal somewhere, but they are available in multiple locations online.) I don’t think there is anything controversial here (Choosing Wisely usually just states the way too obvious anyway), but I include them in case anyone is still doing any of these things: 

• Do not obtain radiographs in children with bronchiolitis, croup, asthma, or first-time wheezing.
• Do not obtain screening laboratory tests in the medical clearance process of pediatric patients who require inpatient psychiatric admission unless clinically indicated.
• Do not order laboratory testing or a CT scan of the head for a patient with an unprovoked, generalized seizure or a simple febrile seizure who has returned to baseline mental status. 
• Do not obtain abdominal radiographs for suspected constipation. 
• Do not obtain comprehensive viral panel testing for patients who have suspected respiratory viral illnesses.

The big problem with Choosing Wisely is they only ever seem to tackle really soft targets. Any doctor routinely doing any of these things is not going to read Choosing Wisely anyway, so what is the point? I wish they would tackle things that really make a difference. For example, in pediatrics they could tell us to stop paying so much attention to the damn urine, and that would help children out tremendously.

More antibiotic adverse events

Daneman N, Cheng Y, Gomes T, Guan J, Mamdani MM, Saxena FE, Juurlink DN. Metronidazole-associated Neurologic Events: A Nested Case-control Study. Clin Infect Dis. 2021 Jun 15;72(12):2095-2100. doi: 10.1093/cid/ciaa395. PMID: 32303736

A huge number of emergency presentations are from medication adverse events and drug interactions, and we miss this diagnosis all the time. In fact, I would guess that ‘medication adverse event’ is the most commonly missed diagnosis in medicine. (We really need wider availability of pharmacy consults in the ED, or follow-up clinics to help consider whether unexplained symptoms might actually be from drug side effects or interactions). I love the studies this group puts out, although the conclusions you can draw from database studies are obviously limited. They performed a case control study of patients over the age of 65 who presented to the hospital with any of cerebellar dysfunction, encephalopathy, or peripheral neuropathy within 100 days of a prescription for either metronidazole or clindamycin. They matched each of those cases to 10 controls. Metronidazole was associated with more neurologic complications than clindamycin (OR 1.7). The major confounder is that the underlying condition could be causing neurologic issues, which is why they compared metronidazole to clindamycin, which would be used in somewhat similar conditions. Although the risk is important to know, the absolute risk is small: among metronidazole recipients, the overall incidence of neurologic events at 100 days was approximately 0.25%.

Bottom line: For patients with unexplained neurological symptoms, medications should always be on the differential. Keep metronidazole in the back of your mind. 

We have 99 problems, but Rads aren’t one

Briggs B, Kalra S, Panacek E. Risk of Radiation Exposure to Emergency Department Personnel From Portable Radiographs. J Emerg Med. 2022 Nov;63(5):645-650. doi: 10.1016/j.jemermed.2022.08.002. Epub 2022 Oct 12. PMID: 36243611

I like this study, because it asks a simple but important question, and gives us a relatively definitive answer (as compared to most of the research we talk about). With all the portable x-rays done in emergency departments, are we being exposed to an important level of radiation? At a single trauma center, they had emergency medicine attendings, nurses, and residents volunteer to wear dosimeters while on shift during a 3 month period. During that time, 1464 portable x-rays were performed in the department. They also placed some dosimeters at stationary places around the department. Portable x-rays do produce radiation, as evidenced by a median of 0.18mSv measured on dosimeters in the resuscitation rooms. However, there was no radiation measured on dosimeters left in other clinical areas. Most importantly, none of the badges worn by clinical staff detected any radiation at all. Unfortunately, they don’t really give us any details about local practices, such as whether fluoroscopy is available or used regularly, or how often emergency staff stay in rooms wearing lead while x-rays are performed.

Bottom line: I have never been concerned about radiation exposure around the department, but this is a nice little study to confirm that that is one less problem we have to think about.

If you have it, you use it; If you don’t, you don’t

Philpott L, Clemensen E, Lau GT. Droperidol versus ondansetron for nausea treatment within the emergency department. Emerg Med Australas. 2023 Aug;35(4):605-611. doi: 10.1111/1742-6723.14174. PMID: 36755492

Based on prior evidence, and my experience with the drug overseas, I have always been a fan of droperidol. That being said, I think the drug may have been overhyped in recent years. (I am not sure it is really much better than haloperidol in most settings, when both are dosed appropriately). This is a single-blind RCT from two emergency departments comparing 2.5 mg of droperidol to 8 mg of ondansetron in 120 adult patients rating their nausea as at least 4/10 on a scale from 0-10. (This is a bigger dose of both drugs than is needed, but is an especially big droperidol dose.) Their primary outcome was the percentage of patients who had at least a 0.8 point improvement in nausea on that 11 point scale, and there was no statistical difference (93% with droperidol and 87% with ondansetron). The mean change in VAS score did marginally favour droperidol (38mm vs 29 mm, p=0.31). Perhaps the outcome I care most about, patients who subjectively felt that the treatment had the desired effect, also significantly favoured droperidol (85% vs 63%, p=0.006) Additional antiemetics were received by 16% of the droperidol group and 37% of the ondansetron group. Although most antiemetic trials have shown no difference between agents, I believe there are at least a few prior RCTs that suggest droperidol is the best agent available to us, so these results fit with prior research. That being said, this is a small trial with incomplete blinding, so I wouldn’t bet the farm on the results. Only the patients were blinded, but honestly, the patients might be the least important people to blind, as they mostly won’t know what these drugs are. It is the clinicians who will have preconceived notions that could bias the results. It is also worth noting that without a placebo arm, we have no idea whether either of these agents was superior to placebo. Ultimately, this trial won’t change my practice, both because I have no access to droperidol, but also because I think almost all of our antiemetics are pretty inter-changeable and at most moderately effective.

Bottom line: When I had droperidol, I was happy to use it, but I don’t think the benefit is so dramatic that we need to be clamoring for the drug to be returned to formularies for this reason in Canada. 

Prophylactic antibiotics for chest tubes

Freeman JJ, Asfaw SH, Vatsaas CJ, et al. Antibiotic prophylaxis for tube thoracostomy placement in trauma: a practice management guideline from the Eastern Association for the Surgery of Trauma. Trauma Surg Acute Care Open. 2022 Oct 25;7(1):e000886. doi: 10.1136/tsaco-2022-000886. PMID: 36312819

This is a new guideline on chest tubes from the Eastern Association for the Surgery of Trauma that states “we conditionally recommend that antibiotic prophylaxis be given at the time of insertion to reduce empyema in adult patients who require [tube thoracostomy] for traumatic hemothorax or pneumothorax.” I was actually surprised by the quality of studies presented. There are actually double blind RCTs dating back to 1977. So why isn’t this common practice? Well, the studies have all individually been negative, but the point estimates have all been promising. When combined in this meta-analysis, there does appear to be a benefit. Of course, a common pattern in medicine is that small early trials suggest benefit, and larger higher quality trials find none. These trials are still all small, and this meta-analysis mixes RCTs with observational data, so this conclusion is far from certain, but I think the conditional recommendation seems reasonable. Usually, I think prophylactic antibiotics are silly. Giving 100 patients antibiotics just so 5 can avoid antibiotics in a few days doesn’t make any sense. However, when the infection is serious enough or hard enough to treat, such as an empyema, the logic shifts, and so I think prophylaxis makes some sense when placing chest tubes. In most of the studies, antibiotics were started in the ED, but continued until after the chest tube was removed, so this isn’t something you will be able to start entirely on your own. It will require buy-in from the admitting team. 

Bottom line: Although this meta-analysis is far from certain, and I certainly wouldn’t delay a critical procedure just to wait for antibiotics, it probably makes sense to give prophylactic antibiotics when placing chest tubes.

Cuff size and BP measurements

Ishigami J, Charleston J, Miller ER 3rd, Matsushita K, Appel LJ, Brady TM. Effects of Cuff Size on the Accuracy of Blood Pressure Readings: The Cuff(SZ) Randomized Crossover Trial. JAMA Intern Med. 2023 Aug 7:e233264. doi: 10.1001/jamainternmed.2023.3264. PMID: 37548984

Not emergency medicine, and not a perfect study, but it is probably valuable to have a sense of how important it is to use an appropriately sized blood pressure cuff. This study recruited 195 adult patients from the community, and measured their blood pressure with an appropriately sized cuff, one too small, and one too big. (For some reason they made everyone walk for 2 minutes before starting, and between each measurement, which seems strange to me, and likely to add variability between the measurements as compared to just doing them back to back.) As expected, BP measurements were somewhat low when the cuff as too large and somewhat high when the cuff was too small, but these differences were small (less than 5 mmHg), and I think my major take away from this paper is that there doesn’t seem to be a clinically relevant difference unless you are using a cuff 2 sizes off. I find this reassuring, because we are generally guessing about cuff size in the ED, and I would not be surprised at all if we are off but 1, but it would be rare (unless you just can’t find the right cuff) to be off by 2. They call this study randomized, in that they randomized the order of small/appropriate/large cuffs, but the randomization provides absolutely no value here. The randomization is completely negated by the fact the trial is not blinded, so the research nurses might just be seeing what they expected to see. (They don’t talk about procedures for faulty measurements, but it would not be the least bit surprising if a study nurse repeated a BP measurement if it ‘seemed wrong’.) In other words, this is a very low quality study, but my guess is any biases would make the differences between the groups seem bigger than they really are. Clinically speaking, I almost never care about the precise blood pressure, and any estimate within about 10 mmHg is good enough. For the few patients in whom this really matters, I wouldn’t be relying on a blood pressure cuff for long, and would move to an arterial line as soon as possible. (I suppose you can think of a few exceptions, such as in the assessment of preeclampsia).

Bottom line: Blood pressure cuff size has an impact, but especially for emergency work, as long as you are within 1 cuff size, you should be OK. 

Gestalt for PE (sort of)

van Maanen R, Martens ESL, Takada T, et al. Accuracy of physicians’ intuitive risk estimation in the diagnostic management of pulmonary embolism: an individual patient data meta-analysis. J Thromb Haemost. 2023 May 30:S1538-7836(23)00438-5. doi: 10.1016/j.jtha.2023.05.023. PMID: 37263381

At first glance this was an interesting paper, but I actually don’t think it tells us anything all that surprising or clinically useful. It is a systematic review and meta-analysis using patient level data to look at the accuracy of physician gestalt in the diagnosis of PE. Unfortunately for emergency department work, they excluded low risk patients, and so the rule in rates between 7 and 40% don’t really apply to us. Their definition of gestalt was whether the “PE is the most likely diagnosis” aspect of Well’s was ticked yes, and so the study wasn’t really looking at nuanced gestalt, but instead just asking the importance of this question in the decision rule (and we have long known that this is the most influential aspect of the Well’s score). I will also note that their definition of “positive for PE” was based on objective diagnosis or 1 month follow up. Obviously 1 month follow up has severe limitations, but even ‘objective diagnosis’ has significant limitations given how often radiologists disagree about these scans. They included 16 studies with over 20,000 patients, and physician gestalt is excellent. Rule in rates were 29% when physicians thought PE was the most likely diagnosis and 9% when we thought it was not the most likely. (This is obviously not a real measure of gestalt, because it is possible for me to think that PE is the second most likely diagnosis, but still have a gestalt that the risk is high enough for a work up. That is a win for gestalt that would be counted as a loss in this analysis.) The relative risk for gestalt was about 3, and remained consistent across the various patient subgroups and settings that they studied. (Perhaps gestalt is a little worse in patients with CHF). What this study really needed was a comparison. How did the relative risk or rule in rate compare to other measures of gestalt? How much better was the overall Well’s score as compared to just the gestalt aspect. Obviously, with a rule in rate of 9% in the gestalt negative group, this data doesn’t really help up clinically in the ED, but they excluded most low risk ED patients. This is an excellent reminder that the most important aspects of many decision rules are subjective, which is one of the many reasons that decision rules, as they are currently being used, are probably hurting medicine.

Bottom line: The gestalt aspect of the Well’s score is highly accurate, but this data can’t really be used for anything clinically, so I apologize if you read the entire summary rather than just skipping to the bottom line.

Push dose Keppra

Summerlin JA, Scaturo N, Lund JA, Wang KM, Frank MA. Adverse events of undiluted intravenous push levetiracetam. Am J Emerg Med. 2023 Nov;73:182-186. doi: 10.1016/j.ajem.2023.08.046. PMID: 37708595

Whether or not it is truly supported by great data, levetiracetam appears to be the current antiepileptic of choice in medicine. It is usually diluted in a minibag of saline and given as an infusion over 15 minutes, but there is apparently lots of previous data that suggests that pushing the undiluted dose straight from the vial is safe. If true, that would obviously be hugely beneficial in status epilepticus. This is a prospective observational trial from a single ED that changed their process so that levetiracetam was primarily given as an undiluted push. Their primary outcome was all adverse events, and among the 250 patients with data, these occurred in 5.6% of patients. Obviously research like this needs a broad definition of adverse events, but it sounds like almost all of these were minor, primarily being pain at the IV site. (I would have loved a little more information about all the adverse events.) There were only 5 adverse events that were considered significant (hypotension, hypertension, tachycardia), but again I would love more information to know how important those adverse events really were. Without a control group, we can’t say how many of these patients would have had an adverse event even if the drug was given slower, but we know it isn’t 0%. Of the 19 actively seizing patients, 15 (79%) were seizure free when reassessed, which is impressive if those were status epilepticus patients, but much less impressive if they just jumped to Keppra in the first few minutes of a simple seizure. The biggest problem with this paper is that although push dose levetiracetam was used in 321 patients, they only managed to collect data on 250 of those patients, for a variety of reasons. Obviously, 250 patients is also not enough to comment on rare adverse events. 

Bottom line: This data is not perfect, but considering the potential upside in status epilepticus, I think the limited risk is worth it. However, if the patient is not actively seizing, there is really no reason to rush, and the standard diluted dose over 15 minutes seems more than reasonable. 

Coffee and Real-time Atrial and Ventricular Ectopy (CRAVE) trial

Marcus GM, Rosenthal DG, Nah G, Vittinghoff E, Fang C, Ogomori K, Joyce S, Yilmaz D, Yang V, Kessedjian T, Wilson E, Yang M, Chang K, Wall G, Olgin JE. Acute Effects of Coffee Consumption on Health among Ambulatory Adults. N Engl J Med. 2023 Mar 23;388(12):1092-1100. doi: 10.1056/NEJMoa2204737. PMID: 36947466

This is a single center RCT looking at the effects of coffee on health. The fact that we have an RCT looking at coffee is amazing, but their definition of ‘health’ (cardiac ectopy, activity levels, sleep minutes, and glucose levels) leaves a lot to be desired, and mostly undermines the value of the study. Using an app, they randomly assigned people to 2 straight days of coffee consumption or avoidance of caffeine. They recruited 100 adult participants through social media, flyers, and word of mouth. Selection bias will be a massive issue, because only a very specific type of person would volunteer for a study that will force you to abstain from caffeine. It was a pretty intrusive study, with participants wearing a continuous ECG monitor, as well as allowing geolocating so the researchers could tell if they were visiting coffee shops. The primary outcome was (in my mind) completely irrelevant: the number of premature atrial contractions per 24 hour period. Adherence to the study protocol was (predictably) imperfect, and it was worse among people who were more regular coffee users at baseline. There was no difference in PACs (58 vs 53 per day, p=0.1). There were more PVCs with caffeine (154 vs 102 per day). There was no difference in arrhythmias. Coffee made people more active (about 1,000 extra steps recorded on a fitbit). Coffee caused less sleep (397 versus 432, or a mean of 36 minutes less of sleep). Glucose levels were unchanged. This is a lot of data that basically amounts to nothing. I don’t think we needed science to tell us that coffee results in less sleep but probably more physical activity. None of this information really translates into health. I saw a lot of strong takes on this study in the media, but it doesn’t really seem to say much at all.

Bottom line: Like all things taken in moderation, coffee probably doesn’t have dramatic effects on health. Drink it if you like it. 

Antibiotic Myths

McCreary EK, Johnson MD, Jones TM, Spires SS, Davis AE, Dyer AP, Ashley ED, Gallagher JC. Antibiotic Myths for the Infectious Diseases Clinician. Clin Infect Dis. 2023 Jun 13:ciad357. doi: 10.1093/cid/ciad357. Epub ahead of print. PMID: 37310038

Antibiotic myths? Of course I am in. This paper covers 8 excellent myths, but I just wanted to cover the 3 that are most impactful for emergency physicians:

1. Oral fosfomycin is an excellent drug for uncomplicated cystitis. I have seen a huge increase in the use of this antibiotic, especially since pharmacists have been allowed to independently treat patients in my province. (Don’t get me started.) They simply point out that the evidence is not perfect. There were some early trials demonstrating excellent efficacy with fosfomycin (equal to normal choices), but there is a more recent RCT that demonstrated that it was significantly worse than a 5 day course of nitrofurantoin (70% vs 58% cure rate, p=0.004). Personally, given that we have so many good options for UTI, I reserve fosfomycin for the ESBL UTIs that would otherwise require IV courses. Unfortunately, rising resistance from overuse may soon eliminate that option.
2. Trimethoprim-sulfamethoxazole does not have in vitro activity against S. Pyogenes. With the rise of MRSA, and that one RCT demonstrating some benefit from antibiotics in abscesses, trimethoprim-sulfamethoxazole is seeing a lot more use these days. It is often prescribed in combination with cephalexin, with the logic that trimethoprim-sulfamethoxazole will only cover staph and we need cephalexin to cover strep. It turns out that this is a misunderstanding based on the fact that agar used to include thymidine, which some species of strep can use to inactivate trimethoprim-sulfamethoxazole. They present data from a large RCT and a systematic review that suggests that trimethoprim-sulfamethoxazole can be used as a solo agent in the treatment of skin and soft tissue infections, and will be effective against both staph and strep species. 
3. Doxycycline is contraindicated in pregnancy and pediatric patients less than 8 years old. This ‘contraindication’ never bothered me much in my early career, but with the dramatic increase in Lyme disease, I feel this more and more often. Tetracycline is associated with maternal hepatotoxicity, inhibition of bone growth, and tooth discoloration. These warnings were applied to doxycycline as well without evidence, as a potential class effect, even though doxycycline has structural changes that should limit all of these impacts. They present some weak data on the lack of harms from doxycycline in children and pregnant women. The data is not definitive, but definitely demonstrates that most patients will be fine, and so they suggest the use of doxycycline in conditions (such as Rickettsial diseases) where there are not great alternatives. I would love to see a harm/benefit conversation regarding single dose prophylactic doxycycline for Lyme disease, because that is my biggest use case. It could go either way, because harms from a single dose are likely to be very small, but most patients given prophylaxis don’t really need it, so the benefits are also limited.

Intensive treatment of cerebral hemorrhage

Ma L, Hu X, Song L, Chen X, et al; INTERACT3 Investigators. The third Intensive Care Bundle with Blood Pressure Reduction in Acute Cerebral Haemorrhage Trial (INTERACT3): an international, stepped wedge cluster randomised controlled trial. Lancet. 2023 Jul 1;402(10395):27-40. doi: 10.1016/S0140-6736(23)00806-1. PMID: 37245517

This is a large multicenter stepped-wedge cluster RCT looking at a bundle of care for patients with acute cerebral hemorrhage presenting within 6 hours and without an underlying structural cause. It is a slightly complex trial, so you might want to read the details in the full post. Essentially, they did a before and after design, comparing baseline care at low and middle income hospitals specifically selected because they had no protocol for managing the physiology of these patients to a bundle of care that aggressively managed blood pressure, glucose, temperature, and warfarin reversal. They enrolled 7,036 patients, and demonstrated an improvement in their primary outcome of functional status based on an ordinal analysis of the modified Rankin scale. There are two big problems with this trial. First, trials of bundles are always difficult to interpret, because we have no way of knowing which aspect of care actually helped. (If you look at the numbers, nothing really changed between the groups, with basically identical blood pressures, glucose, and temperatures, so it seems unlikely that any of these individual aspects is the important factor). More importantly, although this is a cluster RCT, each individual hospital essentially acted as a non-blinded before and after observational study. The only thing that was randomized was timing of the swap from before to after. That means that all of the issues with before/after studies, such as the Hawthorne effect, come into play here, and there is a very good chance that the difference in the primary outcome is based on bias rather than the bundle itself. (Patients receiving the bundle would have to have much closer nursing observation, and good nursing care in stroke has previously been shown to improve outcomes.)

Bottom line: Despite being a very large trial, and having the word randomized in the title, I think this data is essentially equivalent to a large before/after study, and as such really isn’t practice changing. 

Cheesy Joke of the Month

In Canada, we call it an “elevator”, but in England they call it a “lift”.

I guess we are just raised differently.