More Futile “Quality”, vis-à-vis, Alert Fatigue

The electronic health record can be a wonderful tool.  As a single application for orders, results review, and integrated documentation storehouse, it holds massive potential.

Unfortunately, much of the currently realized potential is that of unintended harms and inefficiencies.

Even the most seemingly innocuous of checks – those meant to ensure safe medication ordering – have gone rogue, and no one seems capable of restraining them.  These authors report on the real-world effectiveness of adverse drug alerts related to opiates.  These were not public health-related educational interventions, but, simply, duplicate therapy, drug allergy, drug interaction, and pregnancy/lactation safety alerts.  These commonly used medications frequently generate medication safety alerts, and are reasonable targets for study in the Emergency Department.

In just a 4-month study period, these authors retrospectively identified 826 patients for whom an opiate-related medication safety alert was triggered, and these 4,742 alerts constituted the cohort for analysis.  Of these insightful, timely, and important contextual interruptions, these orders were overridden 96.3% of the time.  And, if only physicians had listened, these overridden alerts would have prevented: zero adverse drug events.

In fact, all 8 opiate-related adverse drug events could not have been prevented by alerts – most of which were itching, anyway.  The authors do attribute 38 potentially prevented adverse drug events to the 3.7% of accepted alerts – although, again, these would probably mostly just have been itching.

Thousands of alerts.  A handful of serious events not preventable.  A few episodes of itching averted.  This is the “quality” universe we live in – one in which these alerts paradoxically make our patients less safe due to sheer volume and the phenomenon of “alert fatigue”.

“Clinically Inconsequential Alerts: The Characteristics of Opioid Drug Alerts and Their Utility in Preventing Adverse Drug Events in the Emergency Department”

Hypertonic Saline No Help for Bronchiolitis

There’s probably no common illness with quite so much flailing about in the evidence base than bronchiolitis.  Between bronchodilators, steroids, saline, and epinephrine, they’ve thrown the kitchen sink at it.  Some show potential benefit, while others do not.

This is another study following up the suspected benefit of nebulized hypertonic (3%) saline in reducing length of stay in hospitalized patients.  In a randomized trial, these authors enrolled 227 patients to receive either nebulized 3% saline or 0.9% saline placebo.  There were no important differences between patients randomized to each arm.  By all outcome measures – intensive care unit admission, objective deterioration of respiratory status, inpatient length-of-stay, or readmission – there was no difference between therapies.

The effect seen here – with seemingly contradictory results – is probably precisely as predicted by Ioannidis’ statistical theory regarding “Why Most Published Research Findings are False”.  Simply put, the prior probability of any of these treatments being helpful is quite low – and the extent of random bias associated most of these studies high.  The net effect, then, is the posterior probability associated with any significant finding is barely changed.

Nebulized hypertonic saline is probably harmless, but if it’s not helpful, it still doesn’t have a role.

“3% Hypertonic Saline Versus Normal Saline in Inpatient Bronchiolitis: A Randomized Controlled Trial”

The Unintended Harms of “Quality”

Harder!  Better!  Faster!  Stronger!  There is a proliferation of time-based measures in the Emergency Department – the glut of which funds a horde of administrative overhead, and for which the Center for Medicare and Medicaid Services will audit.  These measures must frequently seem relatively benign and commonsense when conceived – but their implementation is anything but.

This is a retrospective quality evaluation from the Christiana Health System, looking at their door-to-balloon time metric for STEMI.  Faster is better – of course – so, in 2009, an aggressive quality improvement intervention was performed to decrease delays and obstacles to cardiac catheterization.  As described in the article, this mostly seemed to consist of exhorting each step the process to be performed more rapidly, and providing additional feedback during the QI initiative.

And, it worked!  Median door-to-ballon time sank from 76 minutes to 61 minutes by 2010.

Unfortunately, this came at a cost: false-positive activations more than doubled.  Furthermore, the mortality rate of false-positive activations jumped from 5.6% to 21.6%.  The fatal alternative diagnoses included massive PE, intracranial hemorrhage, severe sepsis, and aortic dissection.

The authors go on further to describe a follow-up QI intervention of education and feedback regarding the missed diagnoses, and, over time, the mortality rate has improved.  However, false positives persist around 20% of activations – triple the original rate.

So, they’ve saved 15 minutes of door-to-balloon time – a probably clinically insignificant amount – at the cost of scads new false-positives and at least one substantial bump in mortality.  And, you know this is but _one_ of many time-based metrics invading – and harming patients in – the Emergency Department.

Will the madness ever stop?

“Aggressive Measures to Decrease “Door to Balloon” Time and Incidence of Unnecessary Cardiac Catheterization: Potential Risks and Role of Quality Improvement”

The Thermometer Accuracy Round-Up

Every patient gets vital signs.  They’re … vital, after all.  And the presence or absence of fever is one of those vital signs – and may substantially alter the diagnostic approach to a number of patients.

However, I’m guessing few of us universally use central sources of temperature measurement – bladder, rectal, pulmonary artery, or esophageal.  We, instead, rely on peripheral measurements – axillary, temporal artery, tympanic membrane, or oral.  And, these authors are here to tell us – within the bounds of their low-quality studies – we are being randomly misled.

This is a meta-analysis and systematic review of 75 studies comparing various peripheral temperature measurement techniques to central gold standard.  These authors rate ±0.5°C as a reasonable, clinically acceptable limit of agreement – and, frankly, even that seems like enough deviation to be relevant.

In their pooled analysis, however, peripheral thermometers were nowhere close – and the ranges worsened both with fever and hypothermia.  In adults, for fever the 95% limit of agreement ranged from -1.44°C to +1.46°C, and for hypothermia the range was -2.07°C to +1.90°C.  This led to a pooled sensitivity and specificity for fever of 64% and 96%.

The authors acknowledge multiple limitations to their data, mostly related to biases in sampling and measurement.  However, this probably still the best evidence available – and it seems it’s just about a flip of a coin whether a peripheral temperature measurement will miss a true fever.

“Accuracy of Peripheral Thermometers for Estimating Temperature”

The Macrolide Scourge, Updated

We’ve had a couple prior alerts regarding the potential cardiovascular risks of macrolides.  These have been taken with quite the grain of salt, particularly considering macrolides were, at one point, prescribed in a trial thought to be cardioprotective.

This meta-analysis is probably the most reasonable data to date on the subject.  These authors (of which there are 17) gather 33 studies comprising 20,779,963 participants.  These studies show, reasonably consistently, small increases in ventricular arrhythmias and sudden cardiac death.  The relative risks reported are 2.52 for VT and 1.31 for SCD, and there was not much reliable variation between different macrolide antibiotics.
However, these relative risks translate into just a handful of additional cases per million prescriptions.  The number needed to harm with cardiovascular death is about 25,000.  On the flip side, across the included studies, the all-cause mortality was unchanged.  So, yes, perhaps macrolides are not entirely benign – but neither is the underlying condition for which they are prescribed.
Overall, this study doesn’t add much to our insight into appropriate macrolide usage.  There may be particular subgroups for whom they may be best avoided, but, the alternative agent must be equally effective against the existing pathology – and not have its own particular undesired interaction.

Of course, if macrolides are used in context where their benefit is minimal or zero – then there is only harm.

“The Role of Macrolide Antibiotics in Increasing Cardiovascular Risk”

The Great Reveal of Andexanet Alfa

The brave new world of “bleeding that doesn’t stop” is a little closer to ending today.  However, this is definitely just the smallest of baby steps in that direction – and hardly as straightforward and simple as the authors’ conclusion:

“Andexanet reversed the anticoagulant activity of apixaban and rivaroxaban in older healthy participants within minutes after administration and for the duration of infusion, without evidence of clinical toxic effects.”

This is ANNEXA-A and ANNEXA-R, two healthy-volunteer studies evaluating the utility of Andexanet Alfa for reversal of Factor Xa inhibitors.  Andexanet is a recombinant Xa decoy protein with a higher afinity for the Xa inhibitors than endogenous Factor Xa, having the net effect of restoring thrombin production.  In these studies, the 40-odd participants in each were housed at the study site for 8 days, loaded with either apixaban or rivaroxaban, and then assigned in a 3:1 or 2:1 ratio to receive either Andexanet or placebo.  Then, the participants were assigned to single-bolus dosing or bolus plus infusion.

The general gist of the outcomes is best visualized by the following graphic:

Andexanet, as expected, rapidly binds circulating Factor Xa.  However, there are two quite obvious, clinically important considerations.  First, reversal is very short-acting if only the bolus is given.  Then, it seems clear while the Factor Xa inhibitor is bound, it cannot be eliminated – and there is a rebound phenomenon to near-baseline levels once the infusion is stopped.  It appears this may be a very tricky drug to use in certain clinical scenarios, considering the duration of required reversal – especially if full and permanent effects are desired.

And, as noted before, this is just baby steps.  This handful of healthy volunteers did not have any specific adverse events related to the infusion, but the small numbers prevent any reliable conclusions regarding safety.  This is also just a pharmacokinetic study without clinical outcomes, and we don’t truly know its real-world effectiveness based on this publication.

It is quite an interesting innovation, but it still lands somewhere between “promising” and “minimally useful” – and almost certainly expensive.

“Andexanet Alfa for the Reversal of Factor Xa Inhibitor Activity”

It’s Not Time, It’s Brain

The evidence continues to mount that, simply, we’re doing it wrong.

Part of this is an acceptable confusion: we didn’t have the computational power or acquisition resolution capable of evaluating the intracranial circulation or perfusion when treating stroke.  However, as the evidence continues to mount from advanced imaging, it seems clear the “time” aspect is not the best determinant of outcome in stroke.

This is a small review of patients from DUST and MR CLEAN with MCA stroke and CT perfusion imaging available.  CTA is frequently performed to evaluate intracranial vasculature and estimate collateral circulation, but CT perfusion provides a dynamic look at contrast flow throughout the brain.

Based on this small 70 patient cohort, the reliability of the observations is hardly bulletproof, with wide 95% confidence intervals.  However, there is a reasonable, linear decrease in good outcomes as poor perfusion characteristics proliferate.  The best perfused collateral circulation led to 15 of 26 patients recovering to mRS 0-2.  While the flow rate and quality of the collaterals degraded, those with good recovery dropped to 12 of 26, then 1 of 12, and finally to 0 of 6.

The horizon is not far now where, hopefully, we’ll look back at the “time is brain” mantra as the infantile scramblings of the dark ages of acute stroke treatment.  Better imaging technology continues to demonstrate some brain is unrecoverable, regardless of timeframe, while others are likely excellent candidates for preservation of some function.  We ought to have much better options for tailoring treatment to the individual in the near future – and as this technology spreads.

“Impact of Collateral Status Evaluated by Dynamic Computed Tomographic Angiography on Clinical Outcome in Patients With Ischemic Stroke”

The Best Defense is … Overtesting?

In a muchhyped study, these authors attempt to address the important mythology of “defensive medicine”.  As the self-preservationists have taught us, there’s nothing to be gained from judicious use of resources – particularly in any sense that outweighs the patient satisfaction and perceived diagnostic certainty resulting from overtesting.

But … is it true?  Does healthcare spending at an individual level really equate to fewer lawsuits?


These authors looked just at the state of Florida, and looked specifically at the costs accumulated for the hospital discharges.  Then, they correlated this data with closed malpractice claims, and broke it all down by clinician specialty.  Impressively, the lowest quintile of spending was less than half the highest quintile.  Then, when comparing clinicians with the lowest levels of hospital spending to those with the highest, the malpractice claim rate steadily decreased for each specialty involved.

Unfortunately, all this depends on the black box of statistical adjustment:

“We accounted for differences in patient characteristics and admission diagnoses between physicians by estimating a patient level multivariable linear model of total hospital charges during the hospital stay as a function of patient age, sex, race, diagnosis related group, indicator variables for Charlson-Deyo comorbid conditions, and year and physician indicators.”

One bit I do think quite strong in this paper is the dramatic association between caesarean delivery rates and diminished malpractice risk.  However – is it all “defensive”?  Or is there just a level of appropriate vigilance that avoids poor outcomes?  Regardless, the trend is quite pronounced.

There is some bit of a leap to associated increased expenditures with defensive medicine, particularly given the various statistical adjustments for disease severity and patient characteristics.  Whether defensive or not, however, the interaction between spending and litigation is quite consistent.  I do think these data indicate what they are supposed to show, but it’s not as airtight as the hype.

“Physician spending and subsequent risk of malpractice claims: observational study”

What Did We Find On Stress Testing?

The so-called “provocative” testing advocated by the American Heart Association for every patient with chest pain of possible ischemic origin has been the thistle in our salad for many years.  They are the nidus for the crystallization of countless chest-pain observation units in our Emergency Departments.

And, they are one of lowest-yield venues in medicine.

Last year, I reported on a set of 1,754 biomarker-negative patients undergoing stress testing in Rhode Island, only 20 of which were true positives.  This latest report follows-up the stress testing outcomes from the ASPECT and ADAPT cohorts in New Zealand.  Of the 1,483 patients enrolled with negative biomarkers, 749 ultimately underwent exercise tolerance testing – 32 of which were positive, 22 of such had stenosis of greater than 70%.  Including index admission and 1-year follow-up of the study cohort, 20 of these received some form of revascularization (PCI or CABG).  In addition, there were 66 equivocal and 103 submaximal stress tests contributing to a total of 70 episodes of invasive coronary angiography, 32 of which identified stenosis >70%.

This is slightly higher yield, overall, than the previously reported U.S. cohort.  Most likely, this is due to patients in this study being almost a decade older – almost 60 years on average – and having median TIMI score of 2, compared with the U.S. stress test population having a TIMI of 0.  But, clearly, a cohort referred for stress testing resulting in only a 4.8% initial revascularization, including just 36 of 93 referred for coronary angiography, is not an effective use of healthcare resources.  And, this presumes the interventions performed as a result of stress testing provide some incremental morbidity or mortality benefit.

There are certainly patients for whom the various types of stress testing, non-invasive coronary angiography, and invasive angiography are each appropriate – but clearly we need to dramatically improve the specificity of our selection criteria.

“The incremental value of stress testing in patients with acute chest pain beyond serial cardiac troponin testing”

Death From Pills

Now that physicians can be convicted of murder for irresponsible opiate prescribing, it’s worth reviewing precisely how we’re fulfilling our felonious destiny.

This is a brief look at the prescription drug-related deaths in San Diego County for the year 2013.  These authors identified each patient death, performed a pharmacy query for all the medications prescribed in the year prior, and then compiled descriptive statistics regarding the providers and specialties involved.  The results are disturbing, yet unsurprising.

  • Primary care provided the bulk of prescriptions related to the 254 deaths, whether opiates, benzodiazepines, or sleep aids.
  • Regarding opiates, 190 patients amassed 2,350 prescriptions totaling 205,700 pills – over 1,000 per patient.
  • Emergency providers were not terribly over-represented, and provided among the fewest pills per prescription, at 22.9.
  • The sample sizes of many subspecialties were small, but the 22 orthopedists captured by this analysis provided an average 169 pills each refill.
  • “Doctor shoppers” constituted 28% of patients, but received 51% of the total prescriptions.  Various surgical specialties seemed to be their primary target.

As irritating as unnecessary antibiotic prescriptions and their ilk may be, this is far more gruesome and disappointing a spectacle.

“Who is prescribing controlled medications to patients who die from prescription drug abuse?”