Every year emergency departments are inundated with cases of influenza-like illness. Rapid flu testing (RFT) offers the promise of a quick and relatively noninvasive rapid diagnostic test. However, the use of this test has significant limitations that can lead to increased risk for both the patient and the provider.
Across various patient populations, RFT is limited in its ability to reliably identify cases of influenza. This poor performance can lead to situations where patients present with influenza yet have a negative RFT. Basing treatment decisions on a falsely negative RFT may lead to under-diagnosis and under-treatment of influenza. Failure to diagnose and treat influenza may increase the risk of a bad outcome for both patients and providers.
There are 4 reasons why providers should not order RFT in the emergency department.
1. RFT performs poorly
Objectively, most rapid flu tests have difficulty reliably identifying patients with influenza. The reported sensitivity of RFT ranges from 10-90% across various patient populations. Interpreting reported test characteristics is difficult, as the available data examines a wide variety of test types across a wide variety of patient populations.1,2
In a recent systematic review, Jacobus et al. evaluated the diagnostic accuracy of RFT. Using 159 studies across a wide range of patient populations, the authors found that RFT had a pooled sensitivity of 62.3%, specificity of 98.2%, positive likelihood ratio of 34.5 and negative likelihood ratio of only 0.38. Given this negative likelihood ratio, the authors concluded that a negative RFT should not be used to rule out the influenza.3
2. RFT misses influenza during the peak of flu season
The influence of disease prevalence becomes much more important during the peak of flu season when the overall disease prevalence is high. With a high rate of disease, the PPV of the RF becomes much more reliable, and the test is much more likely to accurately identify patients with influenza. Unfortunately when prevalence increases, there is a drop in the NPV of RF. With a low NPV, there is an increase in rate of false negative results, so providers run the risk of seeing a patient with influenza and misdiagnosing them after a RF test that was falsely negative.
Sintchenko et al. examined the potential for misdiagnosis and overtreatment across a range of disease prevalence. During a period of high prevalence (10-30%), up to 15% of patients would have a falsely negative RFT and could miss the opportunity to be treated with antivirals. The authors conclude that a strategy of empiric treatment of high-risk patients may be a more effective approach to patients during seasonal influenza outbreaks.4
3. Your clinical impression is sufficient
During the flu season, providers are able to accurately identify patients using clinical criteria with a performance that equals or surpasses most available RFTs. In a retrospective analysis, Monto et al. reported that during an outbreak of influenza, the presence of cough and fever had a PPV of 79%.5 In addition the presence of nasal congestion, and the fairly sudden onset of symptoms are both features that may increase the predictive value of a provider’s clinical assessment.
4. We should focus on treating rather than testing
The Centers for Disease Control (CDC) has issued broad guidelines for patients with potential influenza that emphasizes treatment of potential flu over testing. The biggest risk to providers and patients comes from cases of untreated influenza in high risk patients. To limit these cases, providers should not routinely base clinical decisions on the results of RFT.
Per the guidelines, patients with suspected influenza who are “high-risk” for developing complications from the flu should be treated empirically with antivirals “regardless of initial negative test results.” According to the CDC, this “high risk” subset encompasses a wide variety of patients including those who are obese, elderly, or have a “chronic medical condition.” The efficacy of antivirals for influenza is somewhat questionable especially in patients who are not considered to be high-risk and broad use of antivirals in a low risk population is not encouraged by the CDC or supported by the available literature.
Providers need to be aware of, but should not feel constrained by, the CDC guidelines. Every year, providers are going to see a large number of patients with possible influenza who are “high-risk” according to the CDC. While not all “at-risk” patients need to be treated with antivirals, in the event of an adverse patient outcome, it is likely that the provider’s care would be compared to these readily available guidelines.
One approach to minimize risk to the provider is to have a structured system for discussing and documenting shared medical decision making involving cases of potential influenza. For example:
I think the patient has an influenza type illness. Given the unreliable nature of rapid flu testing I do not think that they need further testing emergently. They (are/are not) considered high risk of having a bad outcome according to the CDC guidelines. I have discussed the role of empiric antiviral therapy and have offered them a prescription.
When can a RFT be helpful?
For admitted patients, RFT is often ordered in an effort to identify patients with influenza in an attempt to cohort infected patients or provide appropriate isolation precautions. In addition RFT may be used by various health agencies in an effort to monitor the overall rate of influenza. While a positive RFT may be helpful in these efforts, unfortunately given it poor performance characteristics, a patient who with a negative RFT may actually have influenza..
A positive RFT may provide some diagnostic clarity in high-risk patients. For instance, a positive RFT in a febrile 65 year old female with COPD who has myalgias and nasal congestion may help confirm the high pre-test likelihood of influenza and may allow providers to focus their workup and treatment. Conversely, if the same patient had a negative RFT, providers should still provide empiric treatment for influenza given the poor performance of RFT and the risk of untreated influenza in high-risk patients.
Routine RFT does not appear to be necessary when evaluating a patient with potential influenza. For patients with a high pre-test probability of influenza, providers should evaluate their overall risk of complication and treat high-risk patients with antivirals. The available literature does not support broad testing or treatment in low-risk patients. In the era where we as providers are being asked to “Choose Wisely” when utilizing healthcare resources, RFT appears to be a commonly ordered, yet poorly performing diagnostic test that can be reasonably avoided in the large majority of patients in the emergency department.
Expert Peer Review
Overall a great article on a topic with important test saving and cost saving implications.
Most of my comments are on grammar and word choice. I then have a few thoughts that could be added to your discussion.[...]
I agree with your sentiments that there is little utility in doing RFT for the majority of patients that we see in the ED and that we should instead be treating them clinically. Your article seems supportive of giving antiviral treatment to patients that meet the case definition of influenza. You may want to put a disclaimer in your conclusion that studies suggest that starting antivirals within 48 hours of symptom onset only decreases symptoms by an average of one day and that the clinician may choose to NOT treat low-risk patients because of the duration of symptoms, small benefit, side effect profile of the medication, and the cost of the medication.
Another point that you may consider mentioning is the fact that hospitals may use influenza testing to cohort hospitalized patients. Therefore, while our treatment decisions in the ED may be independent of the RFT, where the patient is placed in the hospital may be dependent on the results.
Nice job Matthew.
Maine Medical Center, Portland Maine
Tufts University School of Medicine
Expert Peer Review
The point that a negative RFT is unlikely to provide reliable information is an important one. Again, the focus of the clinician should be on the patient for whom the diagnosis of influenza is truly useful. An elderly COPD patient, for instance, with unexplained fevers and myalgias would be a good candidate for RFT, as a method for confirming the likely etiology for the patient’s symptoms. While bacterial superinfection in the setting of influenza can occur, a positive RFT in this patient is still helpful in narrowing the differential diagnosis and guiding therapy.
It might be useful to reference this clinical scenario (or some variant thereof) as an example in which both empiric antiviral therapy as well as RFT may have clinical utility. As you point out, the CDC recommends that patients at high-risk for influenza-related complications, as in this elderly patient, should receive empiric therapy. Sintchenko, et al. stress that the mortality in high-risk patients with untreated influenza may approach 25%.
Treating these patients independently of the RFT result avoids missing high-risk patients who may benefit from therapy while a positive RFT result, as previously noted, may provide additional diagnostic certainty to the evaluation of the undifferentiated febrile patient.
I agree that the phrase in Section 2., “treating before testing” should be clarified. I think the phrase should read “testing before treating,” as the authors point out that empiric treatment may lead to overuse of antivirals in low-risk populations.
Your final recommendations regarding RFT are reasonable. The potential complications of untreated influenza in high-risk populations should likely receive greater emphasis, as part of a strategy that emphasizes treatment over testing.
Vice Chair for Academic Development
Co-Director, Division of Emergency Ultrasound
Expert Peer Review
Good comments on this Drs. Chan, Nelson, and Pigott. Is it worth including that some state health departments ask for flu reporting for surveillance purposes? I know you\'re trying to argue against it\'s use in the ED, but I think Drs. Nelson and Chan make great points in that we need to think outside the ED sometimes. This may be one of those times. Don\'t want EDs to completely abandon a practice that may help others downstream.