Cancer Surgery Survival Better at Busy Hospitals : Two strategies—selective referral and quality improvement—might address the disparities.

SAN DIEGO — A new study has found that foregut cancer patients are significantly more likely to live 5 years after surgery if their operations are performed in hospitals doing a large number of procedures annually, Dr. John D. Birkmeyer said at a symposium sponsored by the Society of Surgical Oncology.

To determine the association between hospital volume of surgical procedures and long-term surgical mortality, Dr. Birkmeyer and his coinvestigators analyzed 10 years of information (1992–2002) in a linked Surveillance, Epidemiology, and End Results (SEER)-Medicare database.

The researchers divided hospitals into high-, medium-, and low-volume terciles based on the number of procedures performed each year. They adjusted data for patient age, acuity, comorbidities, income, stage, and adjuvant therapy to ensure they were not making unfair comparisons.

The greatest differences in the data were produced by esophageal resections, Dr. Birkmeyer reported. Five-year survival was twice as high for high-volume hospitals as for low-volume centers: 34% vs. 17%.

For gastric surgery, the survival curves were similar but showed a smaller difference at 5 years: 32% vs. 26%.

For pancreatic cancer, high-volume hospitals started out with a large advantage in postoperative survival that lasted for about 2 years. Although the difference was still significant at 5 years, it was narrower: 16% vs. 11%. “Quality in this particular cancer may help you run, but it won't help you hide,” Dr. Birkmeyer said at the meeting, where he presented results for three of six cancers in the new study.

“For foregut cancer, hospital volume has a huge effect in terms of hospital mortality, bigger than on almost any other operation,” he said. “High-volume hospitals have better outcomes largely because they have higher-volume surgeons,” he added.

An earlier study by his group and others established that high-volume centers had lower short-term surgical mortality (in hospital or within 30 days) than did hospitals doing relatively few procedures (N. Engl. J. Med. 2002;346:1128–37). When his group compared highest- and lowest-volume hospitals in that study, pancreatic resection and esophagectomy produced the greatest differences in absolute risk of mortality, at 12.5% and 11.9%, respectively. Pneumonectomy had a 5.4% difference in risk. Gastric cancer operations shared fourth place with cystectomy; both had a difference of 2.9%.

The new work shows that the benefit of high volume persists over time, said Dr. Birkmeyer, the George D. Zuidema professor of surgery at the University of Michigan, Ann Arbor. Relatively few previous studies have compared long-term survival, and most studies were small or dated, he said.

Dr. Birkmeyer suggested two strategies—selective referral and quality improvement—to address the disparities revealed by his work. He was not optimistic that either could achieve results any time soon, however.

Selective referral would focus on directing patients to the best hospitals. The media's publication of best hospital lists, the posting of comparative information on Internet databases, and payer efforts to steer patients to selected centers of excellence provide information that can enable selective referral, Dr. Birkmeyer said.

Part of this effort could include closing low-volume centers, he added. “The problem of low-volume surgery is in population-dense areas,” he said, citing previous research showing that patients would not have to travel long distances for care if low-volume centers were closed.

But pushback from angry providers and the lack of an authority to enforce such changes are greater obstacles to such plans, according to Dr. Birkmeyer. “Private payers have the will. They don't have the leverage,” he said. “Medicare has the leverage. It doesn't have the will.”

The second strategy—quality improvement—would “raise the tide” so that all hospitals performed optimally. Proponents would “systematically track surgical outcomes, identify practices associated with optimal outcomes, and assure those practices are implemented as broadly as possible.”

This might not be possible for cancer surgery, however, Dr. Birkmeyer said. First there is the challenge of identifying processes that matter, and then of finding a way to broadly implement them. For example, he cited findings that patients at high-volume hospitals get a wider variety of preoperative tests, are likelier to see a specialist before surgery, and tend to have slower operations. They also receive more invasive monitoring, and for some cancers are more likely to receive adjuvant therapy.

These processes have not been identified as important mediators, according to Dr. Birkmeyer, and he questioned whether they might simply reflect stylistic differences.

SAN DIEGO — A new study has found that foregut cancer patients are significantly more likely to live 5 years after surgery if their operations are performed in hospitals doing a large number of procedures annually, Dr. John D. Birkmeyer said at a symposium sponsored by the Society of Surgical Oncology.

To determine the association between hospital volume of surgical procedures and long-term surgical mortality, Dr. Birkmeyer and his coinvestigators analyzed 10 years of information (1992–2002) in a linked Surveillance, Epidemiology, and End Results (SEER)-Medicare database.

The researchers divided hospitals into high-, medium-, and low-volume terciles based on the number of procedures performed each year. They adjusted data for patient age, acuity, comorbidities, income, stage, and adjuvant therapy to ensure they were not making unfair comparisons.

The greatest differences in the data were produced by esophageal resections, Dr. Birkmeyer reported. Five-year survival was twice as high for high-volume hospitals as for low-volume centers: 34% vs. 17%.

For gastric surgery, the survival curves were similar but showed a smaller difference at 5 years: 32% vs. 26%.

For pancreatic cancer, high-volume hospitals started out with a large advantage in postoperative survival that lasted for about 2 years. Although the difference was still significant at 5 years, it was narrower: 16% vs. 11%. “Quality in this particular cancer may help you run, but it won't help you hide,” Dr. Birkmeyer said at the meeting, where he presented results for three of six cancers in the new study.

“For foregut cancer, hospital volume has a huge effect in terms of hospital mortality, bigger than on almost any other operation,” he said. “High-volume hospitals have better outcomes largely because they have higher-volume surgeons,” he added.

An earlier study by his group and others established that high-volume centers had lower short-term surgical mortality (in hospital or within 30 days) than did hospitals doing relatively few procedures (N. Engl. J. Med. 2002;346:1128–37). When his group compared highest- and lowest-volume hospitals in that study, pancreatic resection and esophagectomy produced the greatest differences in absolute risk of mortality, at 12.5% and 11.9%, respectively. Pneumonectomy had a 5.4% difference in risk. Gastric cancer operations shared fourth place with cystectomy; both had a difference of 2.9%.

The new work shows that the benefit of high volume persists over time, said Dr. Birkmeyer, the George D. Zuidema professor of surgery at the University of Michigan, Ann Arbor. Relatively few previous studies have compared long-term survival, and most studies were small or dated, he said.

Dr. Birkmeyer suggested two strategies—selective referral and quality improvement—to address the disparities revealed by his work. He was not optimistic that either could achieve results any time soon, however.

Selective referral would focus on directing patients to the best hospitals. The media's publication of best hospital lists, the posting of comparative information on Internet databases, and payer efforts to steer patients to selected centers of excellence provide information that can enable selective referral, Dr. Birkmeyer said.

Part of this effort could include closing low-volume centers, he added. “The problem of low-volume surgery is in population-dense areas,” he said, citing previous research showing that patients would not have to travel long distances for care if low-volume centers were closed.

But pushback from angry providers and the lack of an authority to enforce such changes are greater obstacles to such plans, according to Dr. Birkmeyer. “Private payers have the will. They don't have the leverage,” he said. “Medicare has the leverage. It doesn't have the will.”

The second strategy—quality improvement—would “raise the tide” so that all hospitals performed optimally. Proponents would “systematically track surgical outcomes, identify practices associated with optimal outcomes, and assure those practices are implemented as broadly as possible.”

This might not be possible for cancer surgery, however, Dr. Birkmeyer said. First there is the challenge of identifying processes that matter, and then of finding a way to broadly implement them. For example, he cited findings that patients at high-volume hospitals get a wider variety of preoperative tests, are likelier to see a specialist before surgery, and tend to have slower operations. They also receive more invasive monitoring, and for some cancers are more likely to receive adjuvant therapy.

These processes have not been identified as important mediators, according to Dr. Birkmeyer, and he questioned whether they might simply reflect stylistic differences.

SAN DIEGO — A new study has found that foregut cancer patients are significantly more likely to live 5 years after surgery if their operations are performed in hospitals doing a large number of procedures annually, Dr. John D. Birkmeyer said at a symposium sponsored by the Society of Surgical Oncology.

To determine the association between hospital volume of surgical procedures and long-term surgical mortality, Dr. Birkmeyer and his coinvestigators analyzed 10 years of information (1992–2002) in a linked Surveillance, Epidemiology, and End Results (SEER)-Medicare database.

The researchers divided hospitals into high-, medium-, and low-volume terciles based on the number of procedures performed each year. They adjusted data for patient age, acuity, comorbidities, income, stage, and adjuvant therapy to ensure they were not making unfair comparisons.

The greatest differences in the data were produced by esophageal resections, Dr. Birkmeyer reported. Five-year survival was twice as high for high-volume hospitals as for low-volume centers: 34% vs. 17%.

For gastric surgery, the survival curves were similar but showed a smaller difference at 5 years: 32% vs. 26%.

For pancreatic cancer, high-volume hospitals started out with a large advantage in postoperative survival that lasted for about 2 years. Although the difference was still significant at 5 years, it was narrower: 16% vs. 11%. “Quality in this particular cancer may help you run, but it won't help you hide,” Dr. Birkmeyer said at the meeting, where he presented results for three of six cancers in the new study.

“For foregut cancer, hospital volume has a huge effect in terms of hospital mortality, bigger than on almost any other operation,” he said. “High-volume hospitals have better outcomes largely because they have higher-volume surgeons,” he added.

An earlier study by his group and others established that high-volume centers had lower short-term surgical mortality (in hospital or within 30 days) than did hospitals doing relatively few procedures (N. Engl. J. Med. 2002;346:1128–37). When his group compared highest- and lowest-volume hospitals in that study, pancreatic resection and esophagectomy produced the greatest differences in absolute risk of mortality, at 12.5% and 11.9%, respectively. Pneumonectomy had a 5.4% difference in risk. Gastric cancer operations shared fourth place with cystectomy; both had a difference of 2.9%.

The new work shows that the benefit of high volume persists over time, said Dr. Birkmeyer, the George D. Zuidema professor of surgery at the University of Michigan, Ann Arbor. Relatively few previous studies have compared long-term survival, and most studies were small or dated, he said.

Dr. Birkmeyer suggested two strategies—selective referral and quality improvement—to address the disparities revealed by his work. He was not optimistic that either could achieve results any time soon, however.

Selective referral would focus on directing patients to the best hospitals. The media's publication of best hospital lists, the posting of comparative information on Internet databases, and payer efforts to steer patients to selected centers of excellence provide information that can enable selective referral, Dr. Birkmeyer said.

Part of this effort could include closing low-volume centers, he added. “The problem of low-volume surgery is in population-dense areas,” he said, citing previous research showing that patients would not have to travel long distances for care if low-volume centers were closed.

But pushback from angry providers and the lack of an authority to enforce such changes are greater obstacles to such plans, according to Dr. Birkmeyer. “Private payers have the will. They don't have the leverage,” he said. “Medicare has the leverage. It doesn't have the will.”

The second strategy—quality improvement—would “raise the tide” so that all hospitals performed optimally. Proponents would “systematically track surgical outcomes, identify practices associated with optimal outcomes, and assure those practices are implemented as broadly as possible.”

This might not be possible for cancer surgery, however, Dr. Birkmeyer said. First there is the challenge of identifying processes that matter, and then of finding a way to broadly implement them. For example, he cited findings that patients at high-volume hospitals get a wider variety of preoperative tests, are likelier to see a specialist before surgery, and tend to have slower operations. They also receive more invasive monitoring, and for some cancers are more likely to receive adjuvant therapy.

These processes have not been identified as important mediators, according to Dr. Birkmeyer, and he questioned whether they might simply reflect stylistic differences.