More and more hospitals have instituted Electronic Medical Record (EMR) systems. The great majority of medical information about patient and their interaction with caregivers and administrative staff is recorded electronically and saved in medical information repositories. The information goes in but often does not come out. Clinicians (and patients) must repeatedly enter redundant information and reinterpret unstructured data multiple times, and cut and paste the information in multiple reports. EMR semantic search allows intelligent extraction of answers to directed questions easily.

Using QPID as a basic tool, we have built a number of workflow and clinical decision applications. To order interventional procedures, usually the referring physician needs to supply a great amount of supporting data, which then the radiologist must judge to properly plan the exam. Most of this information is already in the EMR either as explicit structured fields or as unstructured text, which QPID's semantic search and NLP tools are well geared to extract. In screening for MRI exams, many of the questions about metal in the body or claustrophobia should be already mentioned in PCP notes, or even prior radiology reports. Rather than ask the patient every time, especially for information they may not remember or be able to communicate, previously entered relevant data can be displayed to the screener or patient for confirmation.

A number of applications are being used around Partners using EMR semantic search as the basic tool. The applications use this information to present and collect information for increasing safety, optimizing workflow, and providing clinical decision support.

We present a number of workflow and safety enhancement applications driven by the QPID EMR semantic search engine.