Abstract
Background
Non-operative management (NOM) of traumatic solid organ injury (SOI) has become commonplace. This paradigm shift, along with reduced resident work hours, has significantly impacted surgical residents’ operative trauma experiences. We examined ongoing changes in residents’ operative SOI experience since duty hour restriction implementation, and assessed whether missed operative experiences were gained elsewhere in the resident experience.
Methods
We examined data from American College of Graduate Medical Education case log reports from 2003 to 2018. We collected mean case volumes in the categories of non-operative trauma, trauma laparotomy, and splenic, hepatic, and pancreatic trauma operations; case volumes for comparable non-traumatic solid organ operations were also collected. Solid organ injury operative volumes were compared against non-traumatic cases, and change over time was analyzed.
Results
Over the study period, both trauma laparotomies and non-operative traumas increased significantly (P < .001). In contrast, operative volumes for splenic, hepatic, and pancreatic trauma all significantly decreased (P < .001; P = .014; P < .001, respectively). Non-traumatic spleen cases also significantly decreased (P < .001), but liver cases and distal pancreatectomies increased (P < .001; P = .017). Pancreaticoduodenectomies increased, albeit not to a significant degree (P = .052).
Conclusions
Continuing increases in NOM of SOI correlate with declining resident experience with operative solid organ trauma. These decreases can adversely affect residents’ technical skills and decision-making, although trends in specific non-traumatic areas may help to mitigate such losses. Further work should determine the impact of these trends on resident competence and autonomy.
Introduction
The management of traumatic solid organ injury (SOI) has changed substantially in recent years. The increased use of computed tomography (CT) scans and the advent and validation of Focused Assessment with Sonography in Trauma has largely replaced diagnostic peritoneal lavage in the evaluation of both blunt and, in select cases, penetrating trauma.1-3 Based on these trends, and informed partially by successes in pediatric trauma patients, non-operative management (NOM) of SOI has gained widespread acceptance.3,4 Consequently, SOI operative trauma volumes have declined.1-4 Avoiding operations for SOI has the potential to reduce hospital length of stay, eliminate operative morbidity, and minimize resource utilization. However, the reduced exposure to traumatic SOI cases simultaneously decreases general surgery residents’ (GSR) operative trauma experience, which is critical for learning surgical techniques and the development of intra- and post-operative decision-making.1,5
Global trends in trauma management have evolved in the context of the resident work-hour restrictions mandated by the American College of Graduate Medical Education (ACGME) in 2003.3,6 While the reported impact of the 80-hour work week on GSR operative volume is variable, several studies have demonstrated an overall decrease in operative trauma cases in this time frame, including for SOI specifically.3,6-8 A limited exposure to operative SOI can impair residents’ familiarity and facility with key surgical techniques, potentially hampering their ability to confidently function independently at the conclusion of residency. 5 Much of the previous research examining this topic was performed during the time period surrounding the implementation of the 80-hour work week, and few studies have described ongoing trends in resident operative experience with SOI. As such, this study aims to describe current GSR operative SOI experience since the introduction of work-hour restrictions to minimize the effect of duty hours on trends in case volumes, and also to evaluate similar trends in elective solid organ operations to determine whether any lost SOI operative experiences are being compensated for elsewhere. We hypothesized that over time continued increases in NOM would be associated with sustained decreases in operations for SOI, and the declines in SOI cases would not be offset by non-traumatic operations.
Methods
This study was reviewed by the University at Buffalo IRB and determined to be exempt. We obtained ACGME case log data over a 15-year period, from academic year 2003-2004 to 2017-2018. These data are publicly available through the ACGME’s website, and represent the case logs of graduating GSR for each academic year, which are utilized for the purposes of application for the American Board of Surgery (ABS) qualifying examination. Data points obtained for each academic year included mean case numbers in the following categories: exploratory laparotomy for trauma; non-operative trauma; splenectomy/splenorraphy for trauma; repair/drainage hepatic laceration and hepatic resection for injury (combined); drainage and/or resection of pancreatic injury (combined); open non-traumatic splenectomy; non-traumatic liver cases (combined); Whipple procedure (pancreaticoduodenectomy); distal pancreatectomy; and pancreatic drainage/resection for abscess or necrosis (combined). Due to low numbers, the following categories were combined for ease of analysis: hepatic laceration and hepatic resection for trauma; drainage and resection of pancreatic injury; drainage of pancreatic abscess and resection for pancreatic necrosis. Since traumatic liver operations were combined, we utilized the subtotal (non-trauma) liver category for comparison. Elective laparoscopic splenectomies were excluded from analysis as there was no comparable traumatic spleen category.
The 15-year study period was divided into 3 equal 5-year time periods (Period 1: 2003-2008; Period 2: 2008-2013; Period 3: 2013-2018), and these periods were compared within each category using both Student’s t test and ANOVA (SAS, version 9.4) to evaluate for significant changes over time; significance was determined to a level of P < .05. Average case numbers in each category were plotted over time and linear regression performed. Finally, elective spleen, liver, and pancreas cases were compared against SOI cases using Poisson regression for both the annual mean number of cases and the annual rate of change in each procedure category. Non-traumatic cases were used as a reference for both the mean and rate models, and the model fit was evaluated using previously described techniques.9,10
Results
Average Trauma Cases by Category.
Trauma cases within each category analyzed over time. The values in each column represent the mean number of cases for each academic year. P values based on t testing are reported as a comparison between the current and preceding period (ie Period 1 and Period 2 or Period 2 and Period 3); the boxed value at the bottom represents the P value of ANOVA testing including all 3 periods in a continuous fashion.
Average Non-Trauma Cases by Category.
Non-trauma cases within each category analyzed over time. The values in each column represent the mean number of cases for each academic year. P values based on t testing are reported as a comparison between the current and preceding period (ie Period 1 and Period 2 or Period 2 and Period 3); the boxed value at the bottom represents the P value for ANOVA testing including all 3 periods in a continuous fashion.
Similar to operations performed for trauma, non-traumatic open splenectomies decreased from an average of 2.3 to 1.4 cases (P < .001); these cases also decreased significantly between study periods (P < .001; Table 2). These data, and their linear regressions, are presented in Figure 1. Based on our Poisson regression model, the mean number of traumatic spleen operations was significantly greater than the elective operations (P = .02); however, we found no significant difference between the rate of decrease in traumatic and non-traumatic cases during the overall study period (P = .78). Average number of trauma spleen cases as compared with non-traumatic spleen cases. Mean values are depicted for each academic year, with trendlines depicting global trends. As shown, volumes decreased for both categories, although the rate of decrease was greater for trauma operations.
In contrast to operations on the spleen, traumatic liver operations significantly decreased while the overall number of non-traumatic liver operations-wedge resections, formal anatomic hepatectomies, and abscess drainage combined-rose significantly (8.3 to 9.9; P < .001; Figure 2). As shown in Table 2, non-traumatic liver operations consistently increased, including between each time period. On Poisson modelling, there were significantly more non-traumatic liver operations on average than traumatic cases (P <.001). Interestingly, despite the significant changes for the traumatic and elective liver cases separately, the model did not demonstrate a statistically significant difference in the absolute rate of change for the 2 trends (P = .70). Average number of trauma liver cases as compared with non-traumatic liver cases. Mean values are depicted for each academic year, with trendlines depicting global trends. While liver operations for trauma decreased, non-traumatic liver cases increased significantly.
For non-traumatic pancreatic operations, both average Whipple and distal pancreatectomy case numbers increased significantly from Period 1 to Period 2 (P = .010 and .040, respectively), with no significant change during Periods 2 to 3 (Table 2). The opposite was seen in the category of pancreatic abscess drainage/debridement for pancreatic necrosis (Table 2). Among pancreatic cases, only the Whipple category demonstrated no significant change over the entire study period (Table 2). Once again, Poisson modelling showed a highly significant difference between traumatic and non-traumatic operations (P < .001), but no difference in the rate of change for the pancreatic categories (P = .38). Linear regressions for traumatic and non-traumatic pancreas categories are depicted in Figure 3. Average number of trauma pancreas cases as compared with non-traumatic Whipple procedures, distal pancreatectomies, and pancreatic abscess drainage or debridement. Mean values are depicted for each academic year, with trendlines depicting global trends. Whipple operations rose, although not to a significant degree, while distal pancreatectomies increased and drainage/debridement operations decreased.
Discussion
Our analysis of ACGME case log data confirms our hypothesis that since duty hour restriction implementation, nonoperative trauma cases have continued to increase, to the detriment of SOI case volumes for GSR. Contrary to expectations, however, the number of nontraumatic liver and pancreas operations has risen, suggesting that some of the lost operative SOI experience may be offset in certain cases. Even so, these specific cases likely do not replace the opportunities for learning peri- and intra-operative decision making, technical skills, and team management techniques lost as a result of decreased SOI operations.
Surgical management of the trauma patient has changed substantially over the last several decades, and the GSR trauma experience has changed accordingly. Successful application of NOM for SOI has reduced the need for operative management in many patients. Furthermore, work-hour restrictions mean residents spend less time in the hospital; although the published impact of work-hours on case volumes has been equivocal, the GSR operative experience has nevertheless changed.8,11 Single institution studies performed shortly after these rules went into effect demonstrated no difference in overall case volumes.6,11 One study found that total numbers of emergency general surgery operations remained constant following work-hour restrictions; however, the number of advanced procedures decreased significantly, as did cases logged as either “first assistant” or “teaching assistant”. 11 Our results also appear to demonstrate declining case complexity, as the number of trauma laparotomies increased significantly over the study period. This finding may be attributable to increases in damage control management, inaccurate case logging, or, less likely, negative laparotomies. Long-term analyses of ACGME case logs have been similarly conflicting. One study reported an increase in overall case numbers following duty hour restrictions, and another found that although trauma case numbers did not change, total case numbers dropped significantly.3,7 Both studies, however, reported a general trend of decreasing resident trauma experience while confirming decreases in first assistant and teaching assistant cases specifically. 7 In analyzing the time period following duty hour restrictions, our goal was to distill the effect of NOM on operative SOI management and the resident experience.
Previous studies suggest that NOM has decreased the operative trauma experience for residents, and our results confirm that this trend has continued since the implementation of work hour restrictions. Although initial single institution reports were conflicting, a recent Canadian study found that only 40% of graduates over a 10-year period had been present for more than 10 trauma laparotomies.12-14 Furthermore, 1 multi-institutional group showed that approximately 3 operations for liver or spleen injury were performed for every 1000 blunt trauma patients admitted. 2 While duty hour regulations may have exacerbated the pre-existing effects of NOM on operative SOI experience, as Strumwass, et al 8 argued, our results show continuing decreases in SOI case volumes, suggesting that the primary driver for decreasing case volumes is the ongoing application of NOM. Additionally, surgeons accustomed to managing SOI non-operatively, may neither recognize the need for nor feel comfortable performing an operation for SOI. To this end, at least 1 study has indicated that increased NOM is associated with a lack of resident familiarity with operative hemostatic techniques. 5 However, our data do suggest that residents may gain back some of the lost experiences through increasing numbers of non-trauma liver and pancreas operations. Although the scenarios are not completely equivalent, increased exposure to such low-volume, high-complexity cases may help develop confidence and operative autonomy. It is important to note that, despite the statistical significance in many of our trends, the absolute difference between case numbers is small, and there may be little practical difference between performing 2 or 3 trauma splenectomies. Case numbers alone may therefore not provide a complete or accurate representation of the GSR operative experience.
Although graduating GSR generally report high levels of confidence to operate independently, several studies have raised concerns about their readiness for independent practice. In one survey, 80% felt confident to take call at a Level I trauma center, but confidence with specific procedures varied inversely with operative complexity. 15 Surveys of fellowship directors, however, tell a different story: two-thirds of fellows were perceived to be unable to operate independently for more than 30 minutes at the outset of fellowship. 16 These findings are corroborated by a multi-institutional study utilizing data from the System for Improving and Measuring Procedural Learning (SIMPL) application, which showed that almost 25% of graduating residents never attained meaningful autonomy in core ACGME procedures at the completion of residency. 17 Authors have proposed various strategies to overcome deficits in operative trauma experiences, such as specific rotations for senior residents to build autonomy, feedback tools such as SIMPL, and fellowship training.8,17,18 For trauma specifically, initial published experiences with the American College of Surgeons Advanced Surgical Skills Exposure for Trauma course have shown improvements for learners after completing the course. 19 If the trends in operative trauma outlined in our study continue, simulation may be necessary, either as part of surgical residency or during fellowship training.
Our study does have several limitations. This study is a retrospective data review, and is subject to biases in data collection and reporting. Specifically, some studies have noted that ACGME case logs, which are generally completed by the resident, may not accurately reflect the case being performed. 20 Residents may forget to enter cases, log operations differently to meet ABS minimum case requirements, or even not log cases once minimums for board eligibility are met. 20 Thus, ACGME case logs may not accurately depict the true GSR operative experience. More specifically, the numbers of trauma and non-trauma cases may differ based on whether the case was specified as involving trauma. Additionally, reporting of certain operative categories was variable over the study period. We cannot make any determination, based on case log data, whether the observed trends in GSR experience had any impact on patient outcomes. We designed the study period based on when duty hours requirements were mandated by the ACGME; some programs may have adopted duty hours rules in advance, whereas some may have had issues with compliance, and thus these data may not be completely free of the impact of duty hours. Finally, as other authors have noted, residents can have widely disparate operative experiences when logging a given case, both in terms of autonomy and complexity. Consequently, case numbers may also not necessarily reflect procedural competence.
Despite these limitations, we conclude that the operative experiences for GSR in SOI have continued to decrease, but these deficits may be partially offset by increased operative volumes in non-traumatic liver and pancreas cases. As the number of GSR continues to increase, residency programs must explore other avenues to teach this vital practical knowledge. Regardless of whether residents pursue careers in trauma surgery, the decision making and technical skills learned from these experiences can be broadly applicable. It may be time to include trauma skills simulation in resident education, similar to the successful incorporation of Fundamentals of Laparoscopic Surgery and Fundamentals of Endoscopic Surgery. In addition, future work should link quantified operative experience with competence to ensure that surgical residents are adequately trained to begin their careers with confidence and competence.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.