July 2, 1998

Johns Hopkins Medical Institutions' news releases are available from the Office of Communications and Public Affairs' direct e-mail news release service. To enroll, call 410-955-4288 or send e-mail to [email protected].

EMBARGOED FOR RELEASE UNTIL TUESDAY, JULY 7, 1998
********************************************************************

"We know from previous research that regionalization improves outcomes for trauma and neonatal services, but this is the first study to show the same benefit for elective general surgery."

MARYLAND'S DEATH RATE FOR PANCREATIC CANCER SURGERY REDUCED BY MEDICAL REGIONALIZATION

A move in Maryland toward regionalization -- centralizing particular medical services at centers performing the greatest number -- decreased the death rates for one of the most complex operations for cancer, according to a study by Johns Hopkins researchers published in the July 7, 1998 issue of Annals of Surgery.

In what is believed to be the first study of its kind in the state, researchers examined medical data from 1984 to 1995 from all of Maryland's non-federal acute care hospitals to see if -- and how -- regionalization affected in-hospital deaths for patients undergoing a pancreaticoduodenectomy, a complex, high-risk operation for pancreatic cancer also known as the Whipple procedure.

Throughout the study period, regionalization gained momentum, with 795 of the operations performed at 43 hospitals over the 12-year period. However, one medical center (The Johns Hopkins Hospital) increased its annual share of Whipples from 20.7 percent to almost 60 percent during this period.

At the same time, Maryland's overall in-hospital death rate for the operation dropped from 17.2 percent to 4.9 percent. According to the Hopkins researchers, nearly two-thirds of this reduction (61 percent) was due to the concentration of cases at Hopkins, since the mortality rate for the procedure at Hopkins was 1.8 percent over the 12-year study period versus 14.2 percent for the other Maryland hospitals.

"This study strongly suggests that regionalization of pancreaticoduodenectomy to one or more high-volume specialized medical centers in a state saves lives," said Toby Gordon, Sc.D., associate professor of surgery and Johns Hopkins Medicine vice president. "It really underscores the importance of experience in improving patient outcomes for complex procedures."

In the study, hospitals were classified as either high-volume (20 or more procedures performed each year for six of the 12-year study period) or low-volume hospitals (those performing at least one, but less than 20 pancreaticoduodenectomies). Using that criterion, Hopkins, which averaged 51.1 pancreaticoduodenectomies per year, starting with six in 1984 and increasing to 72 in 1995, was the state's only high-volume provider. The 42 low-volume medical centers averaged slightly less than one case per year, with a range of 0 to 9 actual cases performed per year.

While the in-hospital mortality rates for Whipple procedures decreased for all hospitals, the relative risk of death at the low-volume providers compared to the high-volume provider more than doubled during the study time frame, Gordon said. According to the mathematical model used to calculate the risk of death related to the number of procedures performed by a medical center, an estimated 57 patient lives were saved because of regionalization.

The movement toward regionalization of pancreatic surgery may be due to savvy patients and their doctors, according to Gordon. "Since there are no formal polices encouraging regionalization, this shift toward centralization in Maryland occurred not as a result of providers, payers or government interventions, but rather was apparently based on patient or referring physician preference," he said.

The Whipple procedure is usually performed to remove portions of a cancerous pancreas, the duodenum and sometimes part of the stomach as well as the gallbladder and distal biliary tree. It is a procedure fraught with technical challenges and risks. In fact, according to medical studies, prior to 1980, the in-hospital mortality for this operation in the United States was more than 20 percent. --JHMI--

Media Contact: Gary Stephenson (410)955-5384 E-mail: [email protected]