LAS VEGAS — Adverse outcomes are significantly more common in hospitals that perform fewer than 50 bariatric surgeries annually, Edward H. Livingston, M.D., said at the annual meeting of the North American Association for the Study of Obesity.
“There is a very, very steep drop-off [in adverse outcomes] above 50 cases per hospital per year,” said Dr. Livingston, chair of the Division of Gastrointestinal/Endocrine Surgery at the University of Texas, Dallas.
He presented data from the 2001 National Inpatient Survey (NIS), which included 10,672 cases. Adverse outcomes fell from roughly 15% of bariatric cases in low-volume hospitals to about 5% in hospitals performing 50–75 and 75–100 bariatric surgeries per year.
In higher-volume hospitals, adverse outcomes continued to decline but at a much less dramatic rate, Dr. Livingston said.
He calculated 656 adverse outcomes (6.1%) and 36 deaths (0.3%) in the NIS sample. He described this as “about the range for most serious complications of bariatric surgery.”
The study used three proxy criteria to identify complications in the database: death during surgery, discharge to a long-term care facility, and a length of stay twice the median length of stay (7 days or more, based on a median of 3 days). Dr. Livingston said he chose these criteria because he had found in a previous study that many obviously adverse outcomes of bariatric surgery were not coded as complications.
He also reported that the procedure-volume relationship remained statistically significant after he corrected for age and sex, two known risk factors in bariatric surgery. The overall odds ratio of procedure volume being a risk factor for adverse outcomes was 0.98, he said. Being male produced the highest odds ratio (1.59) followed by age (1.04).
The study overcame two challenges, Dr. Livingston said at the meeting, cosponsored by the American Diabetes Association. First, it needed a large administrative database that would also reflect the full range of bariatric surgeries performed. The NIS data are drawn from hospital (both academic medical center and non-academic) admissions in 29 states.
“Most series outcomes come from individuals, in university practices, with a high technical degree of expertise,” he said. “We tried to look at the community at large.”
Second, although volume-outcome relationships are common in technically complex operations, Dr. Livingston said they are difficult to calculate in low-mortality procedures. To that end, he focused on adverse outcomes, as well as mortality.
“Clearly, we can demonstrate a volume-outcome relationship for low-mortality procedures if adverse outcomes are used in addition to mortality to find the effect,” he said.