You have accessJournal of UrologyThis Month in Pediatric Urology1 Oct 2022This Month in Pediatric Urology Julian Wan Julian WanJulian Wan More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000002899AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookTwitterLinked InEmail Diagnostic Code-Based Screening for Identifying Children with Primary Hyperoxaluria Before hypotheses could be generated, observation and data were needed. These conjectures could then be tested and their results collated. The interpretation of these new data restarts the cycle; hypotheses are discarded, revised, and reformulated. New tests are run and so on and so forth. Arthur Conan Doyle’s Sherlock Holmes, the famous fictional detective notable for his deductive acumen, identified a key weakness to this method: it needed reliable observations. In “The Adventure of the Copper Beeches,” he exclaims “Data! Data! Data! … I can’t make bricks without clay.”1 The near ubiquity of the computers, the Internet, and electronic medical records offer the hope that data could be gathered from vast physical and virtual networks. Reports of rare and unusual conditions that would take hundreds of years to collate by any single practitioner or institution could be gathered easily. Common conditions could likewise be accumulated and subjected to new analysis. In a multi-institutional study, Tasian et al (page 898) investigated this idea by evaluating the use of diagnostic clinical codes for primary hyperoxaluria.2 They obtained data from PEDSnet, a clinical research network that gathers electronic medical records comprising millions of children from a consortium of large children’s hospitals. They found, unfortunately, that there are weaknesses in the accuracy of coding and discovered the data not to be as useful as they might hope. The diagnostic codes for primary hyperoxaluria have poor positive predictive value; only those patients who had additional coding were most predictive. This suggests caution and preparations are needed when using large administrative data sets—one cannot simply presume accuracy and validity despite the relative ease of acquisition and transmission.