Re:Neat-patienting for RSV in the emergency department
Proposed response Dear Editor, We thank the writers for their interest in our work and for the most part agree with them. (1)(2) We fully agree that bronchiolitis is a clinical diagnosis; it is because the viral aetiology cannot be determined clinically that testing is contemplated. We also agree that specific treatment is not yet available; although RSV specific drugs are being developed and it behooves emergency departments to at least start considering how they will respond when these treatments do become available. Cohorting is actually a controversial topic. Because different strains of respiratory syncytial virus (RSV) circulate during the same season, and dual infection with different viruses does occur, antigen testing to cohort patients risks re-infecting infants with a different RSV strain or another virus. We recognize that such limitations are unavoidable in some settings, but felt that such discussion was beyond the scope of our manuscript. In presenting a single positive and negative predictive value Beattie et al (1) allude to a key point of our paper. Prevalence is dynamic because both RSV and bronchiolitis are seasonal and sometimes have more than a single peak. Predictive values provide this information but are susceptible to prevalence which is dynamic. For example, a patient using an assay with the characteristics described in the letter writers' 2011 paper(3) who tested positive when RSV prevalence is 10% would have only a 35% chance of actually having RSV. Our modeling addresses the dynamic nature of RSV prevalence in a way that a static average measure cannot. Such contextual interpretation maximizes the value of testing, including when such testing is being used for cohorting in the manner Beattie et al describe. It also seems reasonable to assume that some antigen tests may perform better than others. However study methodology may also affect results. The writers' own paper (using PCR) appears to have evaluated the antigen test over a four month period in infants whom either senior nurses or doctors anticipated hospital admission. This is a legitimate clinical practice, but as a research design it diminishes both the external validity and generalizability. When measuring a test's diagnostic performance it is generally best to include a broad range of disease severity over a number of seasons (to address potential problems of antigenic drift affecting test performance (4)). Although confidence intervals were not presented in the letter writers' 2011 paper they can be calculated and for sensitivity overlap with ours. The sensitivity 83.4% (95% CI 78.9%, 87.3%) was not different from ours (79.4%, 95% CI 73.9%, 84.2%). Our measurements of specificity 83.1% (95% CI 77.9%, 87.5%) compared with 67.1% (95% CI 61.9%, 72%) do differ. This likely reflects the differences in inclusion criteria between our two studies. The letter writers' additional unreferenced set of test characteristics does not include enough information to allow their evaluation. We had to be selective in the references we cited because of space limitations; this does not diminish the value of cohorting papers that were not cited, rather it reflects the focused nature of the research question we addressed and (an unrequited) desire to avoid tangential controversies. Yours sincerely, Paul Walsh, Christina Overmyer, Christine Hancock, Jacquelyn Heffner, Nicholas Walker, Thienphuc Nguyen, Lucas Shanholtzer, Enrique Caldera, James Pusavat, Eli Mordechai, Martin E Adelson, Kathryn T Iacono (1) Beattie TF, McLellan K, Templeton K. Near-patient testing for RSV in the emergency department. Emerg Med J 2013, Rapid Response available at http://emj.bmj.com/content/early/2013/08/20/emermed-2013- 202729.full/reply#emermed_el_16338 (2) Walsh P, Overmyer C, Hancock C, Heffner J, Walker N, Nguyen T, et al. Is the interpretation of rapid antigen testing for respiratory syncytial virus as simple as positive or negative? Emerg Med J 2013 (20 August 2013, 10.1136/emermed-2013-202767) (3) Mills JM, Harper J, Broomfield D, Templeton KE. Rapid testing for respiratory syncytial virus in a paediatric emergency department: benefits for infection control and bed management. J Hosp Infect 2011 Mar;77(3):248 -251. (4) Anonymous. Evaluation of 11 commercially available rapid influenza diagnostic tests - United States, 2011-2012. Morb Mortal Wkly Rep, 2012 (61) 873-876.
Conflict of Interest:
No competing interests