Can emergency physicians identify tests and procedures whose necessity should be questioned? In light of high-radiation, low-yield tests, we can – indeed we are obligated – to do more than that.   

The September issue of Emergency Physicians Monthly contained a debate over the merits of the Choosing Wisely campaign. ACEP President David Seaberg made the case for why ACEP leadership had decided to not participate in the Choosing Wisely campaign while EPM publisher Mark Plaster spoke out in favor of ACEP’s participation. I would like to use this column to add more context to the debate. Readers are urged to read both positions on and add their insight to the discussion.

The Choosing Wisely campaign asks medical specialty societies to each identify five tests or procedures commonly used in the field, but whose necessity should be questioned by physicians and discussed with patients (see It is not to say that these tests and procedures are never to be performed, but only that physicians ought to consider carefully their indications.

As of April, 2012, nine medical specialty societies had each designated five such tests or procedures and eight other specialties are slated to publish their lists this fall. In addition, 11 consumer advocate groups, including Consumer Reports, have joined to help disseminate information regarding the program to laymen. On August 28, 2012, Dr. Seaberg, on behalf of ACEP, sent an e-mail survey to members requesting suggestions to help identify areas where emergency medicine can reduce the cost of care without sacrificing quality. This is apparently the beginning of an effort on ACEP’s part to address similar concerns as the Choosing Wisely campaign.

Most emergency physicians will acknowledge that there are tests and procedures routinely performed that are not supported by the evidence, which generate unneeded costs and potentially subject the patients to harm (either directly or as the result of subsequent tests or procedures precipitated by the initial unneeded tests or procedures). Some of these unneeded tests and procedures add substantial costs to a patient’s bill but probably have little risk of harm. For example:

  • Routinely measuring a PTT in patients on warfarin (when only a PT is indicated)
  • Routinely ordering an amylase and lipase, when only a lipase is indicated
  • Routinely giving IV fluids to a child with mild to moderate dehydration (when the sippy diet is advised by the American Academy of Pediatrics and other organizations for mild to moderate dehydration). And, despite the fact that a bag of IV fluids likely costs the hospital about a dollar, charges are typically in excess of $100.
  • Routine administration of IV fluids to normally hydrated, non-vomiting patients with kidney stones (does not accelerate passage and may increase pain)
  • Routine administration of IV fluids in the setting of asthma exacerbations
  • Frequent use of unindicated Foley catheters (e.g., when an IV diuretic is given to a CHF patient or for “nursing convenience”).
  • Routine use of anti-nausea medications like promethazine (e.g., Phenergan) and hydroxyzine (e.g., Vistaril) when IM opiates are given
  • Routine use of IM ketorolac for noncolic, nondysmenorrheic pain (e.g., headaches, backaches or pretty much any other pain for which a patient can take oral ibuprofen). Hospitals typically charge over $100 just to administer a parenteral medication plus the inflated charge for the medication. Head-to-head studies comparing 60mg of IM ketorolac with 800mg of oral ibuprofen show no clinically significant difference in onset, duration or efficacy.
  • Prescription of antibiotics for children with asthma exacerbations or croup

But there are some unnecessary tests that are both very costly and which can harm patients. One example is the routine use of CT exams in suspected appendicitis in children. Even the American College of Radiology has advised considering doing ultrasound examinations in these cases. We’ve written extensively about the “ultrasound first” approach and its support in the literature. Specifically, the ACR, as one of its five items says, “Don’t do computed tomography (CT) for the evaluation of suspected appendicitis in children until after ultrasound has been considered as an option.” Three other items on its list of five also related to emergency care:

  1. Don’t do imaging for uncomplicated headache.
  2. Don’t image for suspected pulmonary embolism (PE) without moderate or high pre-test probability.
  3. Avoid admission or preoperative chest x-rays for ambulatory patients with unremarkable history and physical exam. Sometimes it is the request of other medical staff members that this test be done, however, emergency physicians should not “aide and abet” and should support discontinuing these needless x-rays.

Of the above three, clearly doing unnecessary pulmonary CT angiograms ranks among the tests that are most expensive and most associated with risk (a CT of the chest is the equivalent to the radiation received from about 500 chest X-rays).
How often are these tests performed without appropriate indications? The first study below found that 7% of low risk patients with a negative d-dimer received CTPAs while the second study found the incidence to be 14% and the third found an incidence of 11%.

Based on extensive literature on this subject, it would be the extraordinarily rare patient in whom a CTPA would be indicated who had low clinical risk and a negative d-dimer, yet, these costly and risky tests are performed in a substantial subset of patients in whom they are unindicated. Were some PEs found in patients who did not meet clinical indications – about 1% at three-month follow-up is customary. But it is acknowledged that no strategy will be 100% sensitive for the diagnosis of PE and in some of these cases, a PE at three months may not have been their when the original studies were performed.

Corwin, M.T., et al, Am J Roent 192:1319, May 2009

BACKGROUND: Protocols for the evaluation of patients with suspected pulmonary embolism (PE) generally call for evaluation of “low clinical probability” patients with a D-dimer assay, no further testing if the D-dimer result is negative, and imaging with CT scanning if it is positive. For “high clinical probability” patients, it is generally recommended to proceed directly to imaging without prior D-dimer testing. Rates of adherence to these recommendations in actual practice are uncertain.

METHODS: This study, from Brown University and Rhode Island Hospital, retrospectively evaluated patterns of diagnostic testing in 5,344 ED patients evaluated for possible PE.

RESULTS: The clinical suspicion for PE was low in 70% of the patients (3,716), in whom D-dimer testing was performed. The D-dimer result was negative in 61% of these patients (2,285) but CT scanning was performed in 7% of this group and was positive for PE in one (0.6% of those undergoing CT scanning, but 0.04% of those with negative D-dimer test results). D- dimer results were positive in 39% of the patients with a low clinical suspicion of PE (1,431), but CT scanning was performed in only 58% of this group and was positive for PE in 2% (1.3% of those with a low clinical suspicion but a positive D-dimer). CT scanning was performed in all 1,628 patients with a high clinical suspicion for PE, and was positive for PE in 9%.

CONCLUSIONS: These findings suggest that in actual clinical practice D-dimer screening for possible PE is not being utilized as recommended. 21 references (This email address is being protected from spambots. You need JavaScript enabled to view it. - no reprints)
Copyright 2009 by Emergency Medical Abstracts - All Rights Reserved 10/09 - #37

Teismann, N.A., et al, Ann Emerg Med 54(3):442, September 2009

BACKGROUND: Protocols for the evaluation of possible venous thromboembolism (VTE) include imaging without prior D-dimer testing in patients considered to have an intermediate or high pre-test probability of VTE and preliminary D-dimer testing in those with a low pre-test probability, with subsequent exclusion of the diagnosis when the D-dimer is negative and imaging if it is positive.

METHODS: The authors, from Alameda County Medical Center-Highland Hospital in Oakland, CA, reviewed electronic records of 553 ED patients who underwent D-dimer testing for possible VTE (either deep venous thrombosis or pulmonary embolism) during a 13-month period to determine if imaging studies were performed according to these guidelines.

RESULTS: The D-dimer result was negative in 48.1% of the patients. Imaging was performed in 13.9% of these patients, and was positive for VTE in 3/37 cases (1% of patients with a negative D- dimer). The D-dimer result was positive in 51.9% of the patients, of whom only about half (52.6%) subsequently underwent imaging for VTE; imaging was positive in 20 of these 150 patients (13% of those who were imaged after a positive D-dimer).

CONCLUSIONS: In this ED setting, guidelines for imaging after D-dimer testing for VTE were inconsistently followed. The authors speculate that, in some cases, this might have been related to performance of D-dimer testing before a proper clinical assessment of the pre-test probability of VTE, and that clinicians might have been guided by their clinical perspectives rather than D-dimer results when making decisions about the need for imaging. They note, as well, the potential for false-negative and false-positive D-dimer

Venkatesh, A.K., et al, Arch Intern Med 172(13):1028, July 9, 2012

BACKGROUND: It has been estimated that up to 25% of imaging for pulmonary embolism (PE) is avoidable. An imaging efficiency measure targeted at appropriate use of CT pulmonary angiography in ED patients with a low pretest probability of PE was endorsed by the National Quality Forum (NQF) in 2011.

METHODS: The authors, coordinated at Harvard University, analyzed prospectively collected data for 5,940 hemodynamically stable patients who underwent at least one diagnostic test for possible PE in the EDs of ten academic and two community hospital EDs in order to identify rates of avoidable imaging (CT or V/Q scan), defined as imaging performed in patients with a low pretest probability of PE according to the Wells score and/or physician gestalt, and those in whom D-dimer testing was not performed or was negative.

RESULTS: The Wells score was consistent with a low pretest probability of PE in 69% of the patients, an intermediate probability in 28% and a high probability in 3%. D-dimer testing was performed in 72% of the patients and imaging was performed in 65%. More than half of the stable patients who were imaged (54%) had a low pretest probability. The NQF criteria for avoidable imaging were met for 32% of the patients (based on failure to perform a D-dimer test in 67% of these patients and imaging despite a negative D-dimer in 33%). Imaging was consistent with PE in 58 of the 1,205 patients in whom imaging was considered avoidable (4.8%).

CONCLUSIONS: According to the NQF criteria, imaging was avoidable in nearly one-third of patients who were imaged for potential PE in the ED. 22 references (This email address is being protected from spambots. You need JavaScript enabled to view it. - no reprints) (PMID: 22664742)
Copyright 2012 by Emergency Medical Abstracts

The bottom line: There are literally dozens of tests, prescribing practices and procedures routinely performed by emergency physicians that are not supported by the evidence and which generate unneeded costs, can prolong ED visits and can even harm patients. It would seem that emergency physicians have a moral and ethical obligation to “do the right thing” for our patients. Substantial efforts should be made to narrow variability and focus on the provision of evidence-based care.

Richard Bukata, MD

Editor of Emergency Medical Abstracts (

Add comment

Security code

Popular Authors

  • Greg Henry
  • Rick Bukata
  • Mark Plaster
  • Kevin Klauer
  • Jesse Pines
  • David Newman
  • Rich Levitan
  • Ghazala Sharieff
  • Nicholas Genes
  • Jeannette Wolfe
  • William Sullivan
  • Michael Silverman

Earn CME Credit