SIIM: Users guide to business analytics
MINNEAPOLIS—During the Society of Imaging Informatics in Medicine (SIIM) conference last week, Katherine Andriole, PhD, an associate professor of radiology at Harvard Medical School, identified the challenges associated with business data and presented a template for a business intelligence solution.
“Radiology information is distributed through multiple databases that can provide vital operational, clinical and financial information. The problem is that reports can’t be created ad hoc. Data reside in different systems and in different formats,” said Andriole, who is also assistant medical director of imaging IT and director of imaging informatics for the Center for Evidence-Based Imaging at Brigham and Women’s Hospital in Boston.
The first step to a business intelligence system is to extract data from various databases, such as RIS, hospital information systems and PACS. After the data is transformed into a unified format, it can be loaded into a business intelligence database, she said.
Data integrity is a potential roadblock during this stage: “Attend to data integrity from the beginning," cautioned Andriole. "Otherwise, the efforts may not provide the anticipated dividends, or the system may not perform as expected.”
Andriole recommended storing data using online analytical processes, a multi-dimensional cube format that makes data more manageable and representational than tables typical of relational databases.
The people side of the process is critical as well. All stakeholders--including radiologists, referring physicians, clinicians, technologists and administrators--need to develop performance indicators or variables that serve as metrics for achieving goals, she said. Typical goals range from better service and decreased patient wait time to decreased turnaround time.
“Be clear about the indicators. Make sure they measure what you need to meet your goals,” stated Andriole. For example, a simple study count can be misleading and fail to provide data about whether or not a scanner is operating at capacity.
Business intelligence is a multi-step process, she concluded. Formulating a solid plan that addresses technical, human and business components prior to launching the project provides a firm foundation for ultimate success.
“Radiology information is distributed through multiple databases that can provide vital operational, clinical and financial information. The problem is that reports can’t be created ad hoc. Data reside in different systems and in different formats,” said Andriole, who is also assistant medical director of imaging IT and director of imaging informatics for the Center for Evidence-Based Imaging at Brigham and Women’s Hospital in Boston.
The first step to a business intelligence system is to extract data from various databases, such as RIS, hospital information systems and PACS. After the data is transformed into a unified format, it can be loaded into a business intelligence database, she said.
Data integrity is a potential roadblock during this stage: “Attend to data integrity from the beginning," cautioned Andriole. "Otherwise, the efforts may not provide the anticipated dividends, or the system may not perform as expected.”
Andriole recommended storing data using online analytical processes, a multi-dimensional cube format that makes data more manageable and representational than tables typical of relational databases.
The people side of the process is critical as well. All stakeholders--including radiologists, referring physicians, clinicians, technologists and administrators--need to develop performance indicators or variables that serve as metrics for achieving goals, she said. Typical goals range from better service and decreased patient wait time to decreased turnaround time.
“Be clear about the indicators. Make sure they measure what you need to meet your goals,” stated Andriole. For example, a simple study count can be misleading and fail to provide data about whether or not a scanner is operating at capacity.
Business intelligence is a multi-step process, she concluded. Formulating a solid plan that addresses technical, human and business components prior to launching the project provides a firm foundation for ultimate success.