Faculty Bibliography

BACKGROUND:The extent to which a surgeon's risk aversion influences their clinical decisions remains unknown. We assessed whether a surgeon's attitude toward risk ("risk aversion") influences their surgical decisions and whether the relationship can be explained by differences in surgeons' perception of treatment risks and benefits. MATERIALS AND METHODS/METHODS:We presented a series of detailed clinical vignettes to a national sample of surgeons (n = 1,769; 13.4% adjusted response rate) and asked them to complete an instrument that measured how risk averse they are within their clinical practice (scale 6-36; higher number indicates greater risk aversion). For each vignette, participants rated their likelihood of recommending an operation and judged the likelihood of complications or full recovery. We examined whether differences in perceived likelihood of complications versus recovery could explain why risk-averse surgeons may be less likely to recommend an operation. RESULTS:Surgeons varied in their self-reported risk aversion score (median = 25, interquartile range[22,28]). Scores did not differ by level of surgeon experience or gender. Risk-averse surgeons were significantly less likely to recommend an operation for patients with exactly the same condition (65.5% for surgeons in highest quartile of risk aversion versus 62.3% for lowest quartile; P = 0.02). However, after controlling for surgeons' perception of the likelihood of complications versus recovery, there was no longer a significant association between surgeons' risk aversion and the decision to recommend an operation (64.7% versus 64.8%; P = 0.96). CONCLUSIONS:Surgeons vary widely in their self-reported risk aversion. Risk-averse surgeons were significantly less likely to recommend an operation, a finding that was explained by a higher perceived probability of post-operative complications than their colleagues.

BACKGROUND:Hospital benchmarking is essential to quality improvement, but its usefulness depends on the ability of statistical models to adequately control for inter-hospital differences in patient mix. We explored whether the addition of injury-specific clinical variables to the current American College of Surgeons-Trauma Quality Improvement Program (TQIP) algorithm would improve model fit. METHODS:We analyzed a prospective registry containing all adult patients who presented to a regional consortium of 14 trauma centers between 2010 and 2011 with severe traumatic brain injury (TBI). We used hierarchical logistic regression and stepwise forward selection to develop two novel risk-adjustment models. We then tested our novel models against the current TQIP model and ranked hospitals by their risk-adjusted mortality rates under each model to determine how model selection affects quality benchmarking. RESULTS:Seven hundred thirty-four patients met inclusion criteria. Stepwise selection resulted in two distinct models: one that added three TBI-specific variables (pupil reactivity, cerebral edema, loss of basal cisterns) to the model specification currently used by TQIP and another that combined two TBI-specific variables (pupil reactivity, cerebral edema) with a three-variable subset of TQIP (age, Abbreviated Injury Scale score for the head region, Glasgow Coma Scale motor score). Both novel models outperformed TQIP. Although rankings remained largely unchanged across model configurations, several hospitals moved across quality terciles. CONCLUSION:The inclusion of injury-specific variables improves risk adjustment for patients with severe TBI. Trauma Quality Improvement Program should consider replacing several of its general patient characteristics with injury-specific clinical predictors to increase efficiency, reduce the risk of overfitting, and improve the accuracy of hospital benchmarking. LEVEL OF EVIDENCE:Prognostic and epidemiological, level II.

BACKGROUND:For many surgical operations, there is a well-established relationship between surgical volume and outcome. We investigated whether this relationship exists for infants with gastroschisis. METHODS:Using the Kids' Inpatient Database for years 2003, 2006, 2009, and 2012, we identified all patients undergoing gastroschisis repair. Controlling for patient characteristics and complexity of disease (comorbid intestinal atresia/perforation, necrotizing enterocolitis, and respiratory distress syndrome), we compared surgical outcomes (mortality, length of stay, and incidence of TPN cholestasis) by hospital volume based on quartile for gastroschisis cases treated per year. RESULTS:We identified 7769 patients treated at 743 hospitals. The majority of hospitals were low-volume (n=445), while only 49 were high-volume. The overall mortality rate was 4.3%, and the median length of stay was 34days. Adjusting for clinical and demographic characteristics, patients treated at high-volume hospitals had similar rates of TPN cholestasis and similar mortality rates, but a higher chance for a prolonged length of stay compared to those treated at low-volume hospitals. CONCLUSIONS:Using national data, we found that gastroschisis patients treated at high-volume hospitals did not have improved outcomes. The benefits of high-volume hospitals, which seem to be important for complex pediatric surgery, may not apply to treatment of gastroschisis. LEVEL OF EVIDENCE/METHODS:Level III Retrospective Study.

BACKGROUND:Several studies have demonstrated the safety and short-term success of nonoperative management in children with acute, uncomplicated appendicitis. Nonoperative management spares the patients and their family the upfront cost and discomfort of surgery, but also risks recurrent appendicitis. METHODS:Using decision-tree software, we evaluated the cost-effectiveness of nonoperative management versus routine laparoscopic appendectomy. Model variables were abstracted from a review of the literature, Healthcare Cost and Utilization Project, and Medicare Physician Fee schedule. Model uncertainty was assessed using both one-way and probabilistic sensitivity analyses. We used a $100,000 per quality adjusted life year (QALY) threshold for cost-effectiveness. RESULTS:Operative management cost $11,119 and yielded 23.56 quality-adjusted life months (QALMs). Nonoperative management cost $2277 less than operative management, but yielded 0.03 fewer QALMs. The incremental cost-to-effectiveness ratio of routine laparoscopic appendectomy was $910,800 per QALY gained. This greatly exceeds the $100,000/QALY threshold and was not cost-effective. One-way sensitivity analysis found that operative management would become cost-effective if the 1-year recurrence rate of acute appendicitis exceeded 39.8%. Probabilistic sensitivity analysis indicated that nonoperative management was cost-effective in 92% of simulations. CONCLUSIONS:Based on our model, nonoperative management is more cost-effective than routine laparoscopic appendectomy for children with acute, uncomplicated appendicitis. LEVEL OF EVIDENCE/METHODS:Cost-Effectiveness Study: Level II.

BACKGROUND:Small studies done during the past decade have demonstrated same-day discharge after appendectomy as an option for non-perforated appendicitis. Here we have examined a large cohort to confirm that same-day discharge in acute non-perforated appendicitis is a safe option. STUDY DESIGN:This was a retrospective study of patients from 14 Southern California Region Kaiser Permanente medical centers. All patients older than 18 years of age with acute, non-perforated appendicitis who underwent a laparoscopic appendectomy between 2010 and 2014 were included. We compared patients discharged on the day of surgery with patients hospitalized for 1 night. We examined readmission rates, complication rates, postoperative emergency department visits, postoperative diagnostic or therapeutic radiology visits, reoperations, and cost of treatment. RESULTS:The cohort was composed of 12,703 patients; 6,710 patients were in the same-day discharge group and 5,993 patients were in the hospitalized group. Patients in the same-day discharge group had a lower rate of readmission within 30 days when compared with the hospitalized group (2.2% vs 3.1%; p < 0.005). In both groups, postoperative rates of visits to emergency or radiology department for diagnostic or therapeutic imaging studies were statistically similar. Postoperative general surgery department visits were slightly higher in the hospitalized group (85% vs 81%; p < 0.001). CONCLUSIONS:Adult patients with acute, non-perforated appendicitis can be discharged safely on the day of surgery without higher rates of postoperative complication or readmission rates compared with those hospitalized after surgery. In addition, same-day discharge in this patient group is cost-effective.

OBJECTIVE:To determine how surgeons' perceptions of treatment risks and benefits influence their decisions to operate. BACKGROUND:Little is known about what makes one surgeon choose to operate on a patient and another chooses not to operate. METHODS:Using an online study, we presented a national sample of surgeons (N = 767) with four detailed clinical vignettes (mesenteric ischemia, gastrointestinal bleed, bowel obstruction, appendicitis) where the best treatment option was uncertain and asked them to: (1) judge the risks (probability of serious complications) and benefits (probability of recovery) for operative and nonoperative management and (2) decide whether or not they would recommend an operation. RESULTS:Across all clinical vignettes, surgeons varied markedly in both their assessments of the risks and benefits of operative and nonoperative management (narrowest range 4%-100% for all four predictions across vignettes) and in their decisions to operate (49%-85%). Surgeons were less likely to operate as their perceptions of operative risk increased [absolute difference (AD) = -29.6% from 1.0 standard deviation below to 1.0 standard deviation above mean (95% confidence interval, CI: -31.6, -23.8)] and their perceptions of nonoperative benefit increased [AD = -32.6% (95% CI: -32.8,--28.9)]. Surgeons were more likely to operate as their perceptions of operative benefit increased [AD = 18.7% (95% CI: 12.6, 21.5)] and their perceptions of nonoperative risk increased [AD = 32.7% (95% CI: 28.7, 34.0)]. Differences in risk/benefit perceptions explained 39% of the observed variation in decisions to operate across the four vignettes. CONCLUSIONS:Given the same clinical scenarios, surgeons' perceptions of treatment risks and benefits vary and are highly predictive of their decisions to operate.

OBJECTIVE:The aim of this study was to determine whether exposure to data from a risk calculator influences surgeons' assessments of risk and in turn, their decisions to operate. BACKGROUND:Little is known about how risk calculators inform clinical judgment and decision-making. METHODS:We asked a national sample of surgeons to assess the risks (probability of serious complications or death) and benefits (recovery) of operative and nonoperative management and to rate their likelihood of recommending an operation (5-point scale) for 4 detailed clinical vignettes wherein the best treatment strategy was uncertain. Surgeons were randomized to the clinical vignettes alone (control group; n = 384) or supplemented by data from a risk calculator (risk calculator group; n = 395). We compared surgeons' judgments and decisions between the groups. RESULTS:Surgeons exposed to the risk calculator judged levels of operative risk that more closely approximated the risk calculator value (RCV) compared with surgeons in the control group [mesenteric ischemia: 43.7% vs 64.6%, P < 0.001 (RCV = 25%); gastrointestinal bleed: 47.7% vs 53.4%, P < 0.001 (RCV = 38%); small bowel obstruction: 13.6% vs 17.5%, P < 0.001 (RCV = 14%); appendicitis: 13.4% vs 24.4%, P < 0.001 (RCV = 5%)]. Surgeons exposed to the risk calculator also varied less in their assessment of operative risk (standard deviations: mesenteric ischemia 20.2% vs 23.2%, P = 0.01; gastrointestinal bleed 17.4% vs 24.1%, P < 0.001; small bowel obstruction 10.6% vs 14.9%, P < 0.001; appendicitis 15.2% vs 21.8%, P < 0.001). However, averaged across the 4 vignettes, the 2 groups did not differ in their reported likelihood of recommending an operation (mean 3.7 vs 3.7, P = 0.76). CONCLUSIONS:Exposure to risk calculator data leads to less varied and more accurate judgments of operative risk among surgeons, and thus may help inform discussions of treatment options between surgeons and patients. Interestingly, it did not alter their reported likelihood of recommending an operation.