Emergency Department Patient Datasets Available for Medical Informatics Research

De-identified patient data collected in a busy emergency department over a one-year period is now available to researchers seeking to study and improve patient monitoring technologies. The main PERSEUS dataset contains recorded vital signs, electrocardiograms and oxygen saturation waveform recordings, and red alarm information from 15 urgent care spaces. The ATOMICS data subsets include original, red alarm patient datastreams with annotations to support machine learning for experimental patient monitoring research. The resources, developed through an AHRQ grant awarded to researchers at Rhode Island Hospital/Alpert Medical School at Brown University, may be openly accessed and used to investigate topic areas such as more effective filtering of noise distractions or development of monitoring algorithms to predict patient deterioration. Access the datasets or email the project’s principal investigator, Leo Kobayashi, M.D., at lkobayashi@lifespan.org, for questions.