Volume matters: Whipple success hinges on surgeon experience

MIAMI BEACH – Studies continue to confirm it: When it comes to complex surgical procedures, higher volume equals better outcomes. But like the chicken-or-egg conundrum, researchers are asking which factor comes first – surgeon or facility.

Data presented at the annual meeting of the Americas Hepato-Pancreato-Biliary Association suggest that personal experience makes the biggest difference, at least for the difficult Whipple procedure. Surgeons who performed the highest number of pancreaticoduodenectomies each year had the best outcomes; when they transferred to low-volume hospitals with historically poor results, these surgeons improved perioperative complications, mortality, and readmissions, and lowered costs.

"The salutary effects of being a high-volume hospital for pancreaticoduodenectomy are transferred when high-volume surgeons relocate," said Dr. Paul Toomey. "It seems that the benefits of a high-volume hospital are more due to who does the surgery rather than where it’s undertaken."

Dr. Toomey, a surgical fellow at the Florida Hospital, Tampa, had a unique opportunity to study what happens when two surgeons highly experienced in the Whipple procedure transferred from a busy hospital to low-volume facilities. The surgeons, who together performed more than 100 of the procedures each year, moved for personal reasons, Dr. Toomey said in an interview.

The study focused on perioperative outcomes, mortality, and readmissions in two groups of Whipple patients: the last 50 undertaken at the high-volume hospital (more than 12 pancreaticoduodenectomies per year), and the first 50 at the low-volume hospital where they worked afterward.

The patient groups were similar. Their mean age was 78 years, a little more than half were men, and the average American Society of Anesthesiologists class was 3. The rates of malignancy were similar in the high- and low-volume centers (88% vs. 82%, respectively).

Overall, the average operative time was 252 minutes, with an estimated blood loss of 300 mL. Patients were in intensive care for 2 days, with an average hospital stay of 9 days. The readmission rate was 19% and 30-day mortality, 5%.

But when Dr. Toomey compared the two time periods, he found significant differences in outcomes, which appeared to be associated with the transfer of the highly experienced surgeons. In fact, he said, outcomes were actually much better at the low- than the high-volume centers after the transfer.

The mean operative time at the low-volume centers was 205 minutes, compared with 305 minutes at the high-volume centers. Estimated blood loss was also less at the low-volume centers (350 vs. 255 mL). ICU stays were significantly shorter (1 vs. 3 days), as was total length of stay (7 vs. 12 days). Readmission rates over 30 days were similar (20% vs. 18%), as was 30-day mortality (4% vs. 6%).

"The salutary benefits of being a high-volume hospital for pancreaticoduodenectomy seem to be transferred when high-volume surgeons relocate," Dr. Toomey said. "The benefits of a high-volume hospital may be more due to who does the pancreaticoduodenectomy rather than where the pancreaticoduodenectomy is undertaken."

Dr. Thomas Wood, also of the Florida Hospital, found a similar trend in his study, which examined Whipple outcomes statewide over 20 years. He related the outcomes to concentration, rather decentralization, of care.

For his study, Dr. Wood examined data from the Florida Agency for Health Care Administration, collected over three 3-year epochs: 1992-1994, 2001-2003, and 2010-2012. The data were sorted by surgeon volume of pancreaticoduodenectomy during these periods and correlated to length of stay, in-hospital mortality, and hospital charges, which were adjusted to 2012 dollars.

Over the 9 years, 893 surgeons performed 3,531 pancreaticoduodenectomies. During each epoch, the number of surgeons went down as the number of operations increased.

In the first epoch, 363 surgeons performed 729 operations. In the second, 334 surgeons performed 1,233 surgeries, and in the third, 196 performed 1,569 operations.

"By 2010-2012, 46% fewer surgeons undertook 115% more surgeries compared to the first period," Dr. Wood said. "In 1992-1994, 62% of pancreaticoduodenectomies were undertaken by surgeons who performed one or fewer per year. This fell to 13% by 2010-2012."

At the same time, the number of surgeons who performed more than 36 procedures in each 3-year period (12 per year) grew significantly. In the first epoch, one surgeon alone performed 45 procedures. In the second, six surgeons performed a total of 361. And in the third epoch, 11 surgeons performed 806 Whipples.

"From [the first through third periods,] there was an 11-fold increase in the number of high-volume surgeons and a corresponding 18-fold increase in the number of pancreaticoduodenectomies by surgeons who were performing at least 12 each year. They were performing more than 50% of these operations."

A separate analysis of just the 2010-2012 data evaluated outcomes by surgeons’ operation volume.

In the first epoch, in-hospital mortality was about 12%. In the second, it fell to 8%, and by the third, to 4%. Patients whose surgeons performed one to nine procedures (up to three per year) were hospitalized an average of about 14 days. Those whose surgeons performed 19 to 36 procedures (up to 12 per year) stayed an average of 8 days – a significant difference.

Mortality followed the same pattern. About 9% of patients whose surgeons performed one to three procedures per year died after surgery, compared with 2% of those whose doctors performed 12 procedures per year.

These more experienced surgeons also discharged more patients to home rather than nursing facilities (88% vs. 82%, a significant difference).

Both hospital and surgeon volume also affected costs. Overall, hospital costs increased significantly, even after adjustment for inflation, jumping from $93,000 to $133,000. But surgeons who performed at least 12 Whipples per year did so at a half the cost of those who performed 1-3 per year ($100,000 vs. $200,000).

"Suffice it to say, the busiest surgeons got the best results," Dr. Wood said.

Despite advances in centralizing this kind of specialized care, though, many low-volume centers are still performing the operations, said Dr. Jeffrey Sutton of the University of Cincinnati. "Research continues to show that high-volume centers have better outcomes," he said. "And yet a significant number of cases are still being done at hospitals that do less than one per year."

If research data aren’t enough to persuade hospitals to send patients to regional centers, Dr. Sutton wondered, could money be a motivating factor? He examined the records of almost 10,000 Whipple operations performed at 419 centers that are part of the University Health Systems Consortium clinical database. Of these, 120 were academic centers and 299 were affiliated hospitals. The procedures were performed from 2009 to 2011. Clinical outcomes included length of stay, mortality, and readmissions. He also assessed the cost of both the index admission and readmissions.

Hospitals were divided into volume quintiles of lowest, low, middle, high, and highest. The lowest-volume centers performed up to 21 cases/year in 2009 to up to 23 in 2011. The highest-volume centers performed up to 180 cases/year in 2009 and up to 216 in 2011.

Intraoperative mortality hovered around 2% at the middle-, high-, and highest-volume centers. At the low-volume centers, it reached 2.5% – not significantly different. But at the lowest-volume centers, 30-day mortality was significantly higher – nearly 4%.

Length of stay was similarly associated with volume. In low-, middle-, and high-volume centers, it was about 9 days. But in the lowest-volume centers, the average length of stay was 11 days – significantly longer than any of the others.

Readmission rates over the first postoperative month were also lowest in the highest-volume centers (16.5%). In the low-volume centers, 30-day readmission was just under 19%. But in the middle-, low-, and lowest-volume centers, it was significantly greater, hovering at nearly 20%.

"It’s not only that lower-volume centers are holding on to patients longer, they also are readmitting those same patients significantly more often," Dr. Sutton noted.

The cost analysis looked at Medicare charge data. "When both index and readmission costs were considered, the median per-patient cost at the lowest-volume centers was $23,005 – 11% more than at the highest-volume center."

But there was no significant difference in the cost of readmissions. "That means that the difference of about $2,263 extra per case was based solely on the index admission," Dr. Sutton said. "Essentially what we saw was that the more cases that are performed, the cheaper each individual case becomes."

"To put it bluntly, some low-volume centers are currently reimbursed higher sums of money for delivering suboptimal care to patients," Dr. Sutton said in an interview. In our current health care climate, which emphasizes improved outcomes at lower costs, this is a travesty. As health service researchers, it is our obligation to our patients to analyze and disseminate these data in an effort to urge policymakers to limit the financial reimbursements for poorer-performing providers."

None of the researchers quoted in this article reported any financial disclosures.

msullivan@frontlinemedcom

MIAMI BEACH – Studies continue to confirm it: When it comes to complex surgical procedures, higher volume equals better outcomes. But like the chicken-or-egg conundrum, researchers are asking which factor comes first – surgeon or facility.

Data presented at the annual meeting of the Americas Hepato-Pancreato-Biliary Association suggest that personal experience makes the biggest difference, at least for the difficult Whipple procedure. Surgeons who performed the highest number of pancreaticoduodenectomies each year had the best outcomes; when they transferred to low-volume hospitals with historically poor results, these surgeons improved perioperative complications, mortality, and readmissions, and lowered costs.

"The salutary effects of being a high-volume hospital for pancreaticoduodenectomy are transferred when high-volume surgeons relocate," said Dr. Paul Toomey. "It seems that the benefits of a high-volume hospital are more due to who does the surgery rather than where it’s undertaken."

Dr. Toomey, a surgical fellow at the Florida Hospital, Tampa, had a unique opportunity to study what happens when two surgeons highly experienced in the Whipple procedure transferred from a busy hospital to low-volume facilities. The surgeons, who together performed more than 100 of the procedures each year, moved for personal reasons, Dr. Toomey said in an interview.

The study focused on perioperative outcomes, mortality, and readmissions in two groups of Whipple patients: the last 50 undertaken at the high-volume hospital (more than 12 pancreaticoduodenectomies per year), and the first 50 at the low-volume hospital where they worked afterward.

The patient groups were similar. Their mean age was 78 years, a little more than half were men, and the average American Society of Anesthesiologists class was 3. The rates of malignancy were similar in the high- and low-volume centers (88% vs. 82%, respectively).

Overall, the average operative time was 252 minutes, with an estimated blood loss of 300 mL. Patients were in intensive care for 2 days, with an average hospital stay of 9 days. The readmission rate was 19% and 30-day mortality, 5%.

But when Dr. Toomey compared the two time periods, he found significant differences in outcomes, which appeared to be associated with the transfer of the highly experienced surgeons. In fact, he said, outcomes were actually much better at the low- than the high-volume centers after the transfer.

The mean operative time at the low-volume centers was 205 minutes, compared with 305 minutes at the high-volume centers. Estimated blood loss was also less at the low-volume centers (350 vs. 255 mL). ICU stays were significantly shorter (1 vs. 3 days), as was total length of stay (7 vs. 12 days). Readmission rates over 30 days were similar (20% vs. 18%), as was 30-day mortality (4% vs. 6%).

"The salutary benefits of being a high-volume hospital for pancreaticoduodenectomy seem to be transferred when high-volume surgeons relocate," Dr. Toomey said. "The benefits of a high-volume hospital may be more due to who does the pancreaticoduodenectomy rather than where the pancreaticoduodenectomy is undertaken."

Dr. Thomas Wood, also of the Florida Hospital, found a similar trend in his study, which examined Whipple outcomes statewide over 20 years. He related the outcomes to concentration, rather decentralization, of care.

For his study, Dr. Wood examined data from the Florida Agency for Health Care Administration, collected over three 3-year epochs: 1992-1994, 2001-2003, and 2010-2012. The data were sorted by surgeon volume of pancreaticoduodenectomy during these periods and correlated to length of stay, in-hospital mortality, and hospital charges, which were adjusted to 2012 dollars.

Over the 9 years, 893 surgeons performed 3,531 pancreaticoduodenectomies. During each epoch, the number of surgeons went down as the number of operations increased.

In the first epoch, 363 surgeons performed 729 operations. In the second, 334 surgeons performed 1,233 surgeries, and in the third, 196 performed 1,569 operations.

"By 2010-2012, 46% fewer surgeons undertook 115% more surgeries compared to the first period," Dr. Wood said. "In 1992-1994, 62% of pancreaticoduodenectomies were undertaken by surgeons who performed one or fewer per year. This fell to 13% by 2010-2012."

At the same time, the number of surgeons who performed more than 36 procedures in each 3-year period (12 per year) grew significantly. In the first epoch, one surgeon alone performed 45 procedures. In the second, six surgeons performed a total of 361. And in the third epoch, 11 surgeons performed 806 Whipples.

"From [the first through third periods,] there was an 11-fold increase in the number of high-volume surgeons and a corresponding 18-fold increase in the number of pancreaticoduodenectomies by surgeons who were performing at least 12 each year. They were performing more than 50% of these operations."

A separate analysis of just the 2010-2012 data evaluated outcomes by surgeons’ operation volume.

In the first epoch, in-hospital mortality was about 12%. In the second, it fell to 8%, and by the third, to 4%. Patients whose surgeons performed one to nine procedures (up to three per year) were hospitalized an average of about 14 days. Those whose surgeons performed 19 to 36 procedures (up to 12 per year) stayed an average of 8 days – a significant difference.

Mortality followed the same pattern. About 9% of patients whose surgeons performed one to three procedures per year died after surgery, compared with 2% of those whose doctors performed 12 procedures per year.

These more experienced surgeons also discharged more patients to home rather than nursing facilities (88% vs. 82%, a significant difference).

Both hospital and surgeon volume also affected costs. Overall, hospital costs increased significantly, even after adjustment for inflation, jumping from $93,000 to $133,000. But surgeons who performed at least 12 Whipples per year did so at a half the cost of those who performed 1-3 per year ($100,000 vs. $200,000).

"Suffice it to say, the busiest surgeons got the best results," Dr. Wood said.

Despite advances in centralizing this kind of specialized care, though, many low-volume centers are still performing the operations, said Dr. Jeffrey Sutton of the University of Cincinnati. "Research continues to show that high-volume centers have better outcomes," he said. "And yet a significant number of cases are still being done at hospitals that do less than one per year."

If research data aren’t enough to persuade hospitals to send patients to regional centers, Dr. Sutton wondered, could money be a motivating factor? He examined the records of almost 10,000 Whipple operations performed at 419 centers that are part of the University Health Systems Consortium clinical database. Of these, 120 were academic centers and 299 were affiliated hospitals. The procedures were performed from 2009 to 2011. Clinical outcomes included length of stay, mortality, and readmissions. He also assessed the cost of both the index admission and readmissions.

Hospitals were divided into volume quintiles of lowest, low, middle, high, and highest. The lowest-volume centers performed up to 21 cases/year in 2009 to up to 23 in 2011. The highest-volume centers performed up to 180 cases/year in 2009 and up to 216 in 2011.

Intraoperative mortality hovered around 2% at the middle-, high-, and highest-volume centers. At the low-volume centers, it reached 2.5% – not significantly different. But at the lowest-volume centers, 30-day mortality was significantly higher – nearly 4%.

Length of stay was similarly associated with volume. In low-, middle-, and high-volume centers, it was about 9 days. But in the lowest-volume centers, the average length of stay was 11 days – significantly longer than any of the others.

Readmission rates over the first postoperative month were also lowest in the highest-volume centers (16.5%). In the low-volume centers, 30-day readmission was just under 19%. But in the middle-, low-, and lowest-volume centers, it was significantly greater, hovering at nearly 20%.

"It’s not only that lower-volume centers are holding on to patients longer, they also are readmitting those same patients significantly more often," Dr. Sutton noted.

The cost analysis looked at Medicare charge data. "When both index and readmission costs were considered, the median per-patient cost at the lowest-volume centers was $23,005 – 11% more than at the highest-volume center."

But there was no significant difference in the cost of readmissions. "That means that the difference of about $2,263 extra per case was based solely on the index admission," Dr. Sutton said. "Essentially what we saw was that the more cases that are performed, the cheaper each individual case becomes."

"To put it bluntly, some low-volume centers are currently reimbursed higher sums of money for delivering suboptimal care to patients," Dr. Sutton said in an interview. In our current health care climate, which emphasizes improved outcomes at lower costs, this is a travesty. As health service researchers, it is our obligation to our patients to analyze and disseminate these data in an effort to urge policymakers to limit the financial reimbursements for poorer-performing providers."

None of the researchers quoted in this article reported any financial disclosures.

msullivan@frontlinemedcom

MIAMI BEACH – Studies continue to confirm it: When it comes to complex surgical procedures, higher volume equals better outcomes. But like the chicken-or-egg conundrum, researchers are asking which factor comes first – surgeon or facility.

Data presented at the annual meeting of the Americas Hepato-Pancreato-Biliary Association suggest that personal experience makes the biggest difference, at least for the difficult Whipple procedure. Surgeons who performed the highest number of pancreaticoduodenectomies each year had the best outcomes; when they transferred to low-volume hospitals with historically poor results, these surgeons improved perioperative complications, mortality, and readmissions, and lowered costs.

"The salutary effects of being a high-volume hospital for pancreaticoduodenectomy are transferred when high-volume surgeons relocate," said Dr. Paul Toomey. "It seems that the benefits of a high-volume hospital are more due to who does the surgery rather than where it’s undertaken."

Dr. Toomey, a surgical fellow at the Florida Hospital, Tampa, had a unique opportunity to study what happens when two surgeons highly experienced in the Whipple procedure transferred from a busy hospital to low-volume facilities. The surgeons, who together performed more than 100 of the procedures each year, moved for personal reasons, Dr. Toomey said in an interview.

The study focused on perioperative outcomes, mortality, and readmissions in two groups of Whipple patients: the last 50 undertaken at the high-volume hospital (more than 12 pancreaticoduodenectomies per year), and the first 50 at the low-volume hospital where they worked afterward.

The patient groups were similar. Their mean age was 78 years, a little more than half were men, and the average American Society of Anesthesiologists class was 3. The rates of malignancy were similar in the high- and low-volume centers (88% vs. 82%, respectively).

Overall, the average operative time was 252 minutes, with an estimated blood loss of 300 mL. Patients were in intensive care for 2 days, with an average hospital stay of 9 days. The readmission rate was 19% and 30-day mortality, 5%.

But when Dr. Toomey compared the two time periods, he found significant differences in outcomes, which appeared to be associated with the transfer of the highly experienced surgeons. In fact, he said, outcomes were actually much better at the low- than the high-volume centers after the transfer.

The mean operative time at the low-volume centers was 205 minutes, compared with 305 minutes at the high-volume centers. Estimated blood loss was also less at the low-volume centers (350 vs. 255 mL). ICU stays were significantly shorter (1 vs. 3 days), as was total length of stay (7 vs. 12 days). Readmission rates over 30 days were similar (20% vs. 18%), as was 30-day mortality (4% vs. 6%).

"The salutary benefits of being a high-volume hospital for pancreaticoduodenectomy seem to be transferred when high-volume surgeons relocate," Dr. Toomey said. "The benefits of a high-volume hospital may be more due to who does the pancreaticoduodenectomy rather than where the pancreaticoduodenectomy is undertaken."

Dr. Thomas Wood, also of the Florida Hospital, found a similar trend in his study, which examined Whipple outcomes statewide over 20 years. He related the outcomes to concentration, rather decentralization, of care.

For his study, Dr. Wood examined data from the Florida Agency for Health Care Administration, collected over three 3-year epochs: 1992-1994, 2001-2003, and 2010-2012. The data were sorted by surgeon volume of pancreaticoduodenectomy during these periods and correlated to length of stay, in-hospital mortality, and hospital charges, which were adjusted to 2012 dollars.

Over the 9 years, 893 surgeons performed 3,531 pancreaticoduodenectomies. During each epoch, the number of surgeons went down as the number of operations increased.

In the first epoch, 363 surgeons performed 729 operations. In the second, 334 surgeons performed 1,233 surgeries, and in the third, 196 performed 1,569 operations.

"By 2010-2012, 46% fewer surgeons undertook 115% more surgeries compared to the first period," Dr. Wood said. "In 1992-1994, 62% of pancreaticoduodenectomies were undertaken by surgeons who performed one or fewer per year. This fell to 13% by 2010-2012."

At the same time, the number of surgeons who performed more than 36 procedures in each 3-year period (12 per year) grew significantly. In the first epoch, one surgeon alone performed 45 procedures. In the second, six surgeons performed a total of 361. And in the third epoch, 11 surgeons performed 806 Whipples.

"From [the first through third periods,] there was an 11-fold increase in the number of high-volume surgeons and a corresponding 18-fold increase in the number of pancreaticoduodenectomies by surgeons who were performing at least 12 each year. They were performing more than 50% of these operations."

A separate analysis of just the 2010-2012 data evaluated outcomes by surgeons’ operation volume.

In the first epoch, in-hospital mortality was about 12%. In the second, it fell to 8%, and by the third, to 4%. Patients whose surgeons performed one to nine procedures (up to three per year) were hospitalized an average of about 14 days. Those whose surgeons performed 19 to 36 procedures (up to 12 per year) stayed an average of 8 days – a significant difference.

Mortality followed the same pattern. About 9% of patients whose surgeons performed one to three procedures per year died after surgery, compared with 2% of those whose doctors performed 12 procedures per year.

These more experienced surgeons also discharged more patients to home rather than nursing facilities (88% vs. 82%, a significant difference).

Both hospital and surgeon volume also affected costs. Overall, hospital costs increased significantly, even after adjustment for inflation, jumping from $93,000 to $133,000. But surgeons who performed at least 12 Whipples per year did so at a half the cost of those who performed 1-3 per year ($100,000 vs. $200,000).

"Suffice it to say, the busiest surgeons got the best results," Dr. Wood said.

Despite advances in centralizing this kind of specialized care, though, many low-volume centers are still performing the operations, said Dr. Jeffrey Sutton of the University of Cincinnati. "Research continues to show that high-volume centers have better outcomes," he said. "And yet a significant number of cases are still being done at hospitals that do less than one per year."

If research data aren’t enough to persuade hospitals to send patients to regional centers, Dr. Sutton wondered, could money be a motivating factor? He examined the records of almost 10,000 Whipple operations performed at 419 centers that are part of the University Health Systems Consortium clinical database. Of these, 120 were academic centers and 299 were affiliated hospitals. The procedures were performed from 2009 to 2011. Clinical outcomes included length of stay, mortality, and readmissions. He also assessed the cost of both the index admission and readmissions.

Hospitals were divided into volume quintiles of lowest, low, middle, high, and highest. The lowest-volume centers performed up to 21 cases/year in 2009 to up to 23 in 2011. The highest-volume centers performed up to 180 cases/year in 2009 and up to 216 in 2011.

Intraoperative mortality hovered around 2% at the middle-, high-, and highest-volume centers. At the low-volume centers, it reached 2.5% – not significantly different. But at the lowest-volume centers, 30-day mortality was significantly higher – nearly 4%.

Length of stay was similarly associated with volume. In low-, middle-, and high-volume centers, it was about 9 days. But in the lowest-volume centers, the average length of stay was 11 days – significantly longer than any of the others.

Readmission rates over the first postoperative month were also lowest in the highest-volume centers (16.5%). In the low-volume centers, 30-day readmission was just under 19%. But in the middle-, low-, and lowest-volume centers, it was significantly greater, hovering at nearly 20%.

"It’s not only that lower-volume centers are holding on to patients longer, they also are readmitting those same patients significantly more often," Dr. Sutton noted.

The cost analysis looked at Medicare charge data. "When both index and readmission costs were considered, the median per-patient cost at the lowest-volume centers was $23,005 – 11% more than at the highest-volume center."

But there was no significant difference in the cost of readmissions. "That means that the difference of about $2,263 extra per case was based solely on the index admission," Dr. Sutton said. "Essentially what we saw was that the more cases that are performed, the cheaper each individual case becomes."

"To put it bluntly, some low-volume centers are currently reimbursed higher sums of money for delivering suboptimal care to patients," Dr. Sutton said in an interview. In our current health care climate, which emphasizes improved outcomes at lower costs, this is a travesty. As health service researchers, it is our obligation to our patients to analyze and disseminate these data in an effort to urge policymakers to limit the financial reimbursements for poorer-performing providers."

None of the researchers quoted in this article reported any financial disclosures.

msullivan@frontlinemedcom