Data-driven decisions are important for strategic planning in radiology and medicine just like in any business. Today, many healthcare organizations rely on information systems to collect data on specific metrics which are vital to understanding the entity’s performance. Extracting and presenting this important data is not always straightforward and may be impractical without the appropriate infrastructure in place or knowledgeable, dedicated personnel on staff.  Even with the engagement of both IT professionals and leadership, using this data for knowledge discovery and/or process improvement activities can still be challenging. Without the appropriate tools, very specialized IT professionals need to be involved every time this data has to be captured and processed.

On the other hand, typically individuals with leadership roles are those best suited to interpret the results in the context of the future directions of the organization. Often discoveries through data interpretation generate additional or new questions, requiring the discovery process of data collection, processing, presentation, and interpretation to repeat. If this iterative process requires frequent communication between the IT personnel (experts in data collection and analysis, but may not be fully aware of the implications around this information) and the organizational administration (experts in understanding the implications of the data but may not know how to access, collect and display the data) the activity can be prone to errors at multiple levels, may occur too slowly to be useful, and will likely fail to demonstrate its full impact.
 
The adoption of Business Analytics tools in the healthcare sector has the potential to dramatically improve organizational capabilities.  
 
Business Analytics (BA) refers to techniques employed to interactively explore massive amounts of data in order to gain insight on past and current performance to help with future strategic planning. Although the name implies that this activity is only used in the business sector, the concept can be adapted to many other areas, including healthcare. The main advantage of this technique over the old model is that it empowers individuals to interactively generate customized views of the data geared to answer specific and sometimes unpredictable questions. This puts the power of dynamic querying and display in the hands of those who are in the best position to interpret the information.
 
BA to Improve Radiology Efficiency

By allowing administrators to interactively slice and dice real-time as well as aggregate data with dynamic views of departmental performance metrics, problem areas can be identified and possible solutions suggested. For example, imagine that referring physicians are unsatisfied with the amount of time they have to wait for a report to be completed and communicated to them once an order for an imaging examination is placed in the system of a particular hospital. Once this issue is exposed to radiology administration and the findings are confirmed via performance measurement, this observation can potentially generate a cascade of questions that need accurate and prompt answers in order to fully identify which portion of the process from exam ordering to radiology report completion is deficient. What is the median time from ordering to exam completion? And from exam completion to preliminary report availability? And from preliminary to final signature? If the first seems to be the culprit, are these times the same for both MRI and CT? If MRI is driving these numbers, should another magnet be purchased? What is the volume of each magnet?
 
BA to Improve Radiology Quality

The same concept can be used to monitor quality initiatives within a department. As an example, imagine that a quality improvement application is used to report every case in which an exam was sub-optimally performed. Questions that come to mind could potentially be:  Is there a pattern? Does this happen during the week or during the weekend? Is this happening at an individual level or is this a systemic process? Are the suboptimal images coming from a single scanner? Has the scanner been recently calibrated?

One can envision a myriad of questions to be answered, with creative solutions elucidated through business analytics.
 
It is not as easy as it looks but you can do it and business analytics can help. Although business analytics tools have many benefits over the old model for process improvement efforts, it is not magical and data does not covert itself into beautiful graphics alone. To set-up standard aggregated views of the data can be a big lift requiring extensive multi-collaborative activities. However, once the ground work is completed one of the greatest advantages is that the tools can be automatically updated and augmented so that important quality, efficiency, and financial measures can be displayed in real-time.

Luciano Prevedello, MD, Katherine Andriole, PhD, FSIIM, and Ramin Khorasani, MD, MPH from the Brigham & Women’s Hospital, Harvard Medical School contributed this article. All three are part of the faculty for the Business Analytics Learning Track at SIIM 2010.