CHICAGO – The involvement of surgical residents in seven common general surgery procedures was associated with higher morbidity, but lower mortality in an analysis of 37,907 operations.
Absolute 30-day morbidity for all cases was 3% in those with surgical residents and 1% in those without surgical resident involvement (SRI).
Absolute 30-day mortality rates for all cases with and without SRI were 0.1% and 0.08%, respectively, Dr. Warren Tseng and his colleagues at the University of California, Davis, reported at the annual meeting of the Western Surgical Association.
In risk-adjusted analyses, SRI was associated with a significant 14% increase in morbidity (odds ratio, 1.14) and a 58% decrease in the risk of death (OR, 0.42).
Not surprisingly, mean incision-to-closing operative times were significantly longer for all procedures with SRI at 98 minutes, compared with 75 minutes without SRI.
Surgical residents were involved in 71% (or 27,039) of the 37,907 procedures identified from 2005 to 2007 in the American College of Surgeons’ NSQIP (National Surgical Quality Improvement Program) database.
The seven procedures and corresponding percentages of SRI were 12,829 open hernia repairs (69%), 3,404 laparoscopic hernia repairs (65%), 6,017 laparoscopic cholecystectomies (76%), 731 open right colectomies (74%), 2,614 total thyroidectomies (82%), 2,940 laparoscopic Nissen fundoplication cases (79%), and 9,372 laparoscopic Roux-en-Y bypass procedures (68%).
In risk-adjusted analysis, SRI was associated with significantly lower mortality for open right colectomy (OR, 0.32), said Dr. Tseng, a general surgery resident at the UC Davis.
"Surgical resident training is safe with respect to these seven procedures," he said, but suggested that future studies are needed to investigate the potential causes for the observed differences.
During a discussion of the study, Dr. Steven Stain of Albany (N.Y.) Medical College said that if it’s commonly believed that resident involvement increases morbidity, there will be patients, payers, and others who won’t want residents involved with cases.
"I’m sure your results are supported by the data, but my concern is that if the implications of this study are taken to the extreme, it’s going to hurt our ability to train residents," he said. "And ultimately, I believe that patients get better care with residents’ involvement, and I’m not sure that your study would support that assertion."
Dr. Stain added that although the NSQIP database is robust, it does not fully account for differences among hospitals, and thus it is not the same surgeon performing, for example, a thyroidectomy – with or without a resident.
Dr. Karen Borman of the Abington (Penn.) Memorial Hospital observed that resident involvement can be a bit of a proxy for other things at an institution, such as increased interruptions and traffic into the OR that could affect infection rates and outcomes. She also questioned whether there may be a selection bias behind cases without SRIs, as they may have been straightforward cases involving uncomplicated patients.
Dr. Richard Thirlby of Virginia Mason Medical Center, Seattle, took issue with the lack of procedure-specific mortality and relative risk calculations.
Invited discussant Dr. Amalia Cochran of the University of Utah, Salt Lake City, described the study as an important extension of previous studies, and said that it stands out for its sheer magnitude. She asked the authors to explain how they chose the seven procedures and how the working relationship between the attending surgeon and resident might influence operative times.
Dr. David Wisner, the study’s senior author, said that they selected procedures with minimal variability, whereas complicated procedures like a Whipple surgery could skew the results. He described the relationship between attending physician and resident as complicated, and said that one of the virtues of the NSQIP database is that some variables are weeded out in an analysis of more than 30,000 patients.
Dr. Wisner agreed that factors other than direct resident involvement could influence outcomes, and said that several exclusion criteria were applied to reduce the risk of selection bias. The study excluded cases with concomitant procedures, nonelective cases, moribund ASA (American Society of Anesthesiologists) category 5 patients, and cases with unknown SRI.
Residents of all levels were included as long as they were part of an Accreditation Council for Graduate Medical Education–accredited training program.
Dr. Tseng and Dr. Wisner disclosed no conflicts of interest. None of the discussants disclosed conflicts.