One of the greatest clinical benefits of PACS is online access to prior imaging studies. For administrators charged with managing these clinical systems, one of their challenges is determining how long completed imaging studies need to be available in the PACS and when they can be migrated to other archive solutions.
“A general concept is that an average storage time of 2 years is enough to handle the majority of the requests,” said Peter Van Ooijen from University Medical Center Gronigen (UMCG) at the University of Gronigen in the Netherlands.
Van Ooijen and colleagues recently completed a study at their institution to investigate how many old studies were actually requested by physicians outside the radiology department. He presented the results at the presented at the 93rd scientific assembly and annual meeting of the Radiological Society of North America in Chicago last month.
The UMCG houses 1,300 beds, employs more than 8,500 staff, is the only university medical center in the northern part of the Netherlands, and is the final point of referral for many patients. Van Ooijen’s study analyzed the utilization of seven years of digital radiology images from studies conducted at the facility.
Physicians from outside the radiology department utilize a web-based application (Web1000 from Agfa Healthcare) with a one-year storage capacity for evaluation of radiological images that is connected to the UCMG electronic medical record (EMR). According to Van Ooijen, image data older than one year can be requested digitally from the radiology department’s Rogan-Delft PACS product
Van Ooijen’s team analyzed the PACS database for requests for image data older than one year made from the web-based viewing application. He noted that only the first request for a particular study is recorded; after the study is transferred to the web server it will stay there and can be accessed multiple times without additional requests being made to the PACS server.
They conducted their study of requests recorded from September through December 2006 and analyzed the data with the use of Microsoft Excel. The results were expressed in terms of the percentage per year of the age of the requested studies against the total of all image requests from the Web1000 application.
The team found that a total of 3,268 requests were made for data acquired more than one year ago at UCMG, an average of more than 30 requests per day. Van Ooijen said that 1,592 requests (48 percent) were for studies between one and two years of age, 711 (22 percent) were for studies between two and three years old, 418 (13 percent) were for studies between three and four years old, 296 (9 percent) were for studies between four and five years old, 188 (6 percent) were for studies between five and six years old, and 63 (2 percent) were for studies between six and seven years old. He noted that the average age of all image requests was 29 months.
“These results clearly show that 52 percent of the requests for data acquired more than one year ago were actually acquired more than two years ago, and 30 percent of the requests were for data more than three years old,” he said. “Furthermore, the results also show that in a small percentage (2 percent), even the oldest data in the PACS [that between six and seven years of age] were still requested.”
Van Ooijen observed that the number of requests clearly indicated that the availability of older data is required for clinical routine; not only for the most recent data, but also for data up to seven years old.