EDITOR'S NOTE: Studies for this issue were identified using the Clinical Queries feature of PubMed, “hand” searching JAMA Pediatrics, Pediatrics, and The Journal of Pediatrics, and from customized EvidenceUpdates alerts.
EBM PEARL: THE 95% CONFIDENCE INTERVAL (CI), PART 2 - CLINICAL USE: The 95% CI has clinical use. As an example, consider the number needed to treat (NNT) and its 95% CI (see piece by Freedman below regarding article by Danewa et al; J Pediatr 2016;169:105-9). You are not particularly impressed with the use of ondansetron in dehydration with vomiting, and have decided that your personal NNT cutoff is 5. You need to be 95% confident that you would not have to treat more than 5 patients to see benefit in one. The results of the study found the ondansetron NNT 4 (95% CI, 3 to 7). As the upper end of the 95% CI crosses your treatment threshold of 5, you cannot be 95% confident that you may have to treat as many of 7 patients to a benefit one patient. The study demonstrated statistical significance, but it did not meet your personal criteria for clinical significance.
LITERATURE SEARCH PEARL: THE H-INDEX: The h-index, first described by Jorge Hirsch (Proc Natl Acad Sci USA 2005;102:16569-72), is a researcher-level productivity measure. The h-index combines, in a specific way, the total number of papers published and the number of citations for each paper. If the researcher has published “h” number of papers, the h-index is the number of papers cited at least “h” number of times. The h-index is not significantly increased by a single highly-cited paper or many sparsely-cited papers. The h-index may be used to compare individual researchers and groups of researchers (eg, research departments) (J Pediatr 2016;169:272-6). For further information and to calculate your h-index, go to researchguides.uic.edu/if/hindex.
Jordan Hupert, MD