Alas, Poor (Literally) Detroit

High-quality care can mean many things.  In the ideal sense, it conveys cooperation and coordination between the many facets of healthcare delivery – great physician care, the best possible translation of medical evidence to an individual patient, outstanding nursing, electronic systems that double-check and triple-check for safety, and an army of staff to support patients in all settings to ensure best possible health.

Or, high-quality care can just be a number.  A surrogate number, obliquely related to the ideals of quality care.  And the penalty for bad care can be monetary, a decrease in the level of reimbursement to a hospital.  Reimbursement, that, paradoxically, could be used to increase the quality of care.

And, so we see this paradox playing out in Detroit, as this letter in the NEJM points out.  Detroit, suffering publicly through bankruptcy and infrastructure collapse, simply does not have the resources to support public health.  Thus, its citizens, with poor access to preventative and primary care, and already burdened by the misery of the City, are forced to seek care – repeatedly – in its metropolitan and inner-city hospitals.  And, because of such repeated visits, these hospitals are subject to the Medicare Readmission penalty – and thus, fewer resources with which to care for those lacking adequate support to care for themselves outside the institutional setting.  And around we go again.

Are we truly measuring and encouraging quality?  Or are we punishing those systems who simply cannot afford to be further stressed?

“Medicare Readmission Penalties in Detroit”
http://www.ncbi.nlm.nih.gov/pubmed/25207786

For tPA, Is Delusion the Standard of Care?

Yet again, the tPA apologists dip into their bag of registry data in an attempt to defend tPA – and end up contradicting themselves.

In 2009, neurologists in India published a retrospective case series examining the outcomes following tPA at their institution.  Specifically, they divided up the cases between those with arterial occlusion present on CT angiogram of the cerebral vessels, and those with no demonstrated arterial occlusion.  For patients with demonstrated occlusion, there were significant differences in early NIHSS improvement favoring tPA, but no long term mRS improvements.  Conversely, there were 119 without occlusion present – and the early NIHSS improvement and late mRS improvement outcomes were similar.  There were, however, substantial baseline differences between those receiving tPA and those who did not – and retrospective studies are confounded by many biases – but there was at least a suggestion that some stroke subtypes might not benefit from tPA.

Clearly, that did not sit well with the authors of this study, a handful of whom are paid representatives of Boehringer Ingelheim.  They performed their own retrospective review of a multi-center registry to evaluate outcomes of patients without arterial occlusion demonstrated on initial angiography, but definitive acute stroke seen on follow-up MRI.  They also further subdivided stroke subtypes into lacunar (subcortical, thalamic, and pontine infarct <20mm) vs. non-lacunar (all others).

They identified 154 non-lacunar strokes, 49 of whom underwent thrombolysis, and 102 lacunar strokes, 54 of whom underwent thrombolysis.  Outcomes favored tPA for non-lacunar stroke syndromes, with 51% of tPA patients mRS 0-1 at 90 days, compared with 30% for those who did not receive tPA.  Symptomatic intracranial hemorrhage occurred in 6.1% of tPA patients, compared with 1% without.  Lacunar strokes, however, had identical 90 day mRS outcomes – 65% vs. 63%.  For lacunar strokes, sICH occured in 3.7% of tPA patients vs. 0% without.  The authors still try to statistically adjust their way out of this equivalency for the primary outcome – but fail.

So, again, this is a retrospective study confounded by many biases.  However, the authors have nicely demonstrated support for a hypothesis some stroke subtypes – particularly those for whom no arterial occlusion is demonstrated on angiography – might not benefit from tPA.  Thus, the conclusion:
“In conclusion, this retrospective study demonstrates the efficacy of intravenous thrombolysis in patients with ischemic stroke who have no radiographically demonstrated arterial occlusion at presentation. Both subgroups, nonlacunar and lacunar strokes, were found to have had better clinical outcome after receiving r-tPA.”
Or, the opposite of what their data suggests.

Is it really so impossible conceive tPA might not be magical?

“Thrombolysis in Ischemic Stroke Without Arterial Occlusion at Presentation”

Sedation May Not Always Be the Answer in Shoulder Reduction

A guest post by Dr. Andrew Kirkpatrick, an Emergency Medicine resident at the University of Texas Medical School at Houston.

What happens when you combine Ultrasound guided nerve blocks and shoulder reduction?  In the hands of experienced operators, not surprisingly, great things!

This is a randomized control trial of 41 patients out of Turkey that compared shoulder reduction utilizing procedural sedation versus an ultrasound guided suprascapular nerve block.  Using the modified Kocher method for reduction, the authors found that the nerve block leads to a statistically significant decrease in pain quantified by a Visual Analog Score.  There was also non-significant trend toward first attempt success using the Ultrasound guided nerve block, which may have been significant if the authors had bothered to calculate the appropriate sample size.  No side effects were noted in the group receiving regional anesthesia, while the sedation group predictably suffered nausea/vomiting in 15%, hypoxia in 10%, and agitation in 15%.  Time to discharge in the nerve block group was 100 minutes less than the procedural sedation group, a finding which reached statistical significance.   Lastly there was no significant difference in patient or physician satisfaction between the two groups, although trends in both cases favored the nerve block group.

Despite all of its short-comings – including small sample size, no blinding, and no clear primary outcome – the paper makes a case for using a suprascapular block to avoid the risks associated with procedural sedation for the reduction of this common dislocation.  The authors show shoulder reductions can be done in the Emergency Department without the use of procedural sedation.  True, there are some patients that cannot tolerate shoulder reduction without being fully disassociated from the event, but in the appropriately selected patient, the technique outlined in this article could decrease patient discomfort and decrease the risk to the patient undergoing shoulder reduction all without starting an IV.  In the world of emergency medicine where we are pressured to “treat and street” patients as quickly as possible while attempting to minimize risk to the patient, the ultrasound guided suprascapular nerve block seems like a great addition to the shoulder reduction armamentarium.

“A Comparison of Suprascapular Nerve Block and Procedural Sedation Analgesia in Shoulder Dislocation Reduction”
http://dx.doi.org/10.1016/j.ajem.2014.02.014

Ondansetron vs. Metoclopramide in Hyperemesis

I hate to say it, but this is one of the first randomized trials I’ve stumbled across from Obstetrics and Gynecology – so, even though it's not terribly profound, I felt compelled to cover it.

This is a double-blind, randomized trial of women admitted to observation with hyperemesis gravidarum.  They received either 4mg of ondansetron or 10mg of metoclopramide every 8 hours for 24 hours, and patient were tracked for vomiting episodes and self-reported nausea.  With regard to these co-primary outcomes, there was no difference between study drugs – most had full resolution, and most in each group had 2 episodes of vomiting or fewer.  However, secondary outcomes and adverse effects – drowsiness, xerostomia, and persistent ketonuria – favored ondansetron.

The authors therefore conclude ondansetron ought probably be favored – all else being equal.  But, for these authors, all else is not equal – ondansetron is markedly more expensive.  Therefore, the ultimate conclusion of these authors is rather in favor of metoclopramide, considering their similar efficacy.

Of other interesting note, one patient, despite admission and supportive care, had 23 episodes of vomiting during the study period.  Oy.

“Ondansetron Compared With Metoclopramide for Hyperemesis Gravidarum
A Randomized Controlled Trial”
http://www.ncbi.nlm.nih.gov/pubmed/24807340

Farewell, CT Stone Protocol

Ureterolthiasis has become a poster child for over-utilization of advanced imaging.  Despite the relative level of distress kidney stones cause our patients, the use of computed tomography has never been associated with improved outcomes – yet, CT is widespread for its diagnostic utility, contributing substantially to $2 billion in annual healthcare expenditures for this condition in the U.S. alone.

This, however, is a comparative effectiveness evaluation promoting ultrasound for the diagnosis of ureterolithiasis in the Emergency Department, a three-pronged evaluation comparing CT, formal ultrasonography by radiology technicians, and bedside Emergency Department ultrasonography.  Essentially, the objective of this study was to compare safety – regarding, in a sense, whether the additional information supplied by CT was valuable for the detection of life-threatening alternative diagnoses.  And, with respect to this outcome all strategies had, essentially, the same number of “misses” during the follow-up period – mostly acute cholecystitis, one case of appendicitis, and a smattering of other thoracoabdominal diagnoses.  And so – ultrasonography, even our amateur sort in the ED, is "just as good".

Of course, there are a few oddities associated with this publication.   There are, bizarrely, three “primary outcomes”, and the authors explicitly choose to report only two of them.  Total costs of care was intended to be an outcome, but the authors simply state those results will not be reported in this paper.  This study also has an interestingly low incidence of ~33% confirmed ureterolithiasis – which may result from their lack of a “gold standard” for diagnosis of stone, relying on patient-reported stone passage or follow-up for stone removal.  Or, it could be enrollment of a population with an oddly low incidence of hematuria – only ~63% of enrolled patients exhibited this common finding with a sensitivity of >80%.  I’d be curious to see the incidence of hematuria in the cases with alternative diagnoses, although there would likely be too few to draw any substantial conclusions.

There was also substantial crossover from the ultrasonography cohorts.  40.7% of those randomized to ED ultrasonography and 27.0% randomized to radiology ultrasonography ultimately underwent CT.  And, this crossover reveals the limitation of ED ultrasonography: a sensitivity of 54%, compared with a sensitivity of 88% for CT.  There was also a major exclusion relevant to the U.S. population: women over 250 lbs and men over 285 lbs.  Unfortunately, a substantial portion of the U.S. exceeds such superlative mass – and the generalizability of these results to that population is open to reasonable variability.

The take home point, however, is reasonable – if you suspect ureterolthiasis, start with bedside ultrasound.  If you see hydronephrosis, you've made your diagnosis, and you can rest easy regarding the relative incidence of missed alternative diagnoses.  Alternatively, however, one can make a very reasonable argument ureterolithiasis can be adequately diagnosed on clinical grounds, and advanced imaging is required only in the minority of cases, regardless of the findings on bedside ultrasound.  The vast majority of ureterolithiasis carries a relatively benign prognosis of recurring and remitting pain of a few weeks duration, and return precautions or outpatient follow-up for persistent symptoms is likewise a reasonable course of action.

“Ultrasonography versus Computed Tomography for Suspected Nephrolithiasis”
http://www.nejm.org/doi/full/10.1056/NEJMoa1404446

Are a Third of Research Conclusions Wrong?

As I covered last year, half of what you’ve been taught in medicine is wrong – we just don’t know which half.

And, it turns out, sometimes even the same authors taking a second look at the same data as before, can come up with new – and wildly different – conclusions.

This is a review of 37 randomized-controlled trials published after 1966 paired with 37 “re-analyses” of the same data.  These trials span the entire medical domain, from mycophenolate therapy after cardiac transplantation to homeopathy for fibrosis.  Of these 37 re-analyses, 32 of them involved authors from the original research group.  These re-analyses differed by changing statistical techniques, outcome definitions, or other study interpretation methods.

Following re-analysis, 13 (35%) changed the original conclusions – either suggesting more, fewer, or even entirely different patients should be treated.  The implication regarding the reliability of our evidentiary basis for medical practice is obviously profound – if even just the original authors and data can result in conflicting conclusions.

In his editorial, Harlan Krumholz argues the solution is clear: open the data.  Independent verification of findings – whether by erasing bias or undesired mathematical pathology – is critical to ensuring the most complete understanding of the evidence base.  If our highest duty is to our patients, we must break down the barriers created by self-interest and institutional policies in order to promote data sharing – and serve patients by improving the clarity and transparency of medical practice.

“Reanalyses of Randomized Clinical Trial Data”
http://www.ncbi.nlm.nih.gov/pubmed/25203082