Article Text

Download PDFPDF

Emergency department information system diagnosis: how accurate is it?
  1. J S Mitchell,
  2. F Westerduin
  1. Emergency Department, Royal Hospital for Sick Children, Glasgow, UK
  1. Dr J S Mitchell, Emergency Department, Yorkhill Hospital, Glasgow G3 8SJ, UK; joannemitchell{at}

Statistics from

Emergency department information systems (EDIS) have been heralded as a “must” for the modern emergency department (ED),1 and it is claimed that use of these systems will enhance patient care.

Over the past decade, information systems have been rolled out to the ED; however, flaws have been recognised. An American study found 90% accuracy for discharge diagnosis data held on their own EDIS.2 A more recent study comparing electronic records with hand-written records found them to be of equal accuracy,3 raising the possibility that our communication with the GP has always been questionable. Accurate communication with the GP is vital to maintain continuous patient care across primary and acute services. Any information passed on which turns out to be incorrect could have dramatic effects on patient care.

Our own ED audited with a discharge diagnosis accuracy of only 87%. The main reason for this was user carelessness in selecting a “close” but inaccurate diagnosis. However, one user error resulted in a completely random diagnosis being inserted into the system.

We should remember that, in this age of new technology, a single diagnosis can be difficult to reach from the ED, especially when the patient is being admitted for further investigations. Diagnosis is a dynamic process and the ED diagnosis will often differ from the hospital diagnosis because of the limited information available to the emergency physician. For this reason, a wide variety of diagnoses must be present on the EDIS menu and having a free text area to provide extra information is invaluable.

We wish to highlight the importance of regular audit of EDIS and improved education and training for staff using EDIS. This letter also questions the validity of EDIS alone as a data collection tool for research and policy-making purposes.



  • Competing interests: None.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.