Diagnostic imaging business managers have access to a wide variety of data from a collection of clinical information systems; however, because of the disparate nature of these systems’ databases, developing meaningful metrics for practice management is difficult.
A team of developers from the Center for Evidence-Based Imaging at Brigham and Women’s Hospital in Boston, using open-source software tools, have created a prototype system that permits the aggregation and analysis of data from multiple systems. Their schema, published online before print in the Journal of Digital Imaging, allows for the application of business intelligence tools to the practice of diagnostic imaging.
“Business intelligence refers to the tools needed to integrate, store, analyze and present data from non-integrated sources,” wrote the authors.
They noted that radiology information is available from a variety of information systems: PACS, RIS, electronic medical record (EMR) or hospital information system (HIS), computerized physician order entry (CPOE), as well as financial systems such as billing or accounts receivable applications.
The challenge, from a practice manager’s perspective, is accessing this wealth of information to discriminate valuable departmental metrics or key performance indicators (KPIs)--such as MRI scanner throughput or report turn-around times--from the data repositories.
“Virtually merging relevant data from various sources into one single format and location could contribute to improved speed and efficiency of knowledge discovery,” they wrote.
According to the authors, the process of integration is known as Extract, Transform and Load (ETL). Their prototype system consisted of Microsoft Windows-based PC and the open-source MySQL server 5.0 as supporting database. The MySQL database served as data repository for all transformations done by the ETL application.
For ETL software, researchers selected Penataho Data Integration 3.0, an open-source application with reporting, analysis, dashboard and data-mining capabilities. It allows for input information to come from multiple database systems such as Oracle, Microsoft.
“Pentaho Data Integration is a JAVA-based application that allows administrators to create complex transformations and jobs in a graphical, drag-and-drop environment without having to generate any custom code,” the developers wrote.
Reports were output in a dashboard-style format via Microsoft Excel, using an open-source connection tool between the database and spreadsheet application.
“One important observation that became clear through testing is that two distinct databases can be combined using this process,” the authors reported. “For instance, selected fields from the RIS and CPOE databases could be potentially mapped and combined into one single resource using this process.”
For practice managers, the deployment of such an application will allow the generation of summary reports from disparate data sources in a quick, time-efficient manner.
“In addition, discovery of problematic processes within the department may become more apparent with analytic tools requiring new KPIs to monitor progress of behavior change for systematic and sustainable improvements,” the developers observed.
The authors reported that open-source tools have the potential to create a micro-database from various departmental information sources and that this database can be used to select KPIs that allow the monitoring of activities and processes within a practice.
“Business intelligence tools can thus enable improvements in healthcare quality, safety, efficiency and financial performance once operational,” the authors wrote.