Factors Associated With Outcomes and Costs After Pediatric Laparoscopic Cholecystectomy

Key Points

Question  What factors are associated with the morbidity and cost of laparoscopic cholecystectomies in Canadian children?

Findings  Among 3519 laparoscopic cholecystectomies in this study, population data, operative indication, patient comorbidities, and total surgeon volume were found to be associated with both the morbidity and cost of pediatric laparoscopic cholecystectomy.

Meaning  Surgeon volume should be an important consideration when Canadian children are referred for laparoscopic cholecystectomy from both a patient safety and economic perspective.

Importance  The prevalence of pediatric cholelithiasis is increasing with the epidemic of childhood obesity. With this rise, the outcomes and costs of pediatric laparoscopic cholecystectomy become an important public health and economic concern.

Objective  To assess patient and health system factors associated with the outcomes and costs after laparoscopic cholecystectomy among Canadian children.

Design, Setting, and Participants  This was a retrospective, population-based study of children 17 years and younger undergoing laparoscopic cholecystectomy from April 1, 2008, until March 31, 2015. The data source was the Canadian Institute for Health Information. The Canadian Institute for Health Information Discharge Abstract Database includes data from all Canadian hospitals. The analysis was limited to inpatient cholecystectomies. All Canadian children undergoing laparoscopic cholecystectomy were included.

Exposure  The exposure in this study was laparoscopic cholecystectomy.

Main Outcomes and Measures  The primary outcome was all-cause morbidity, a composite outcome of any complication that prolonged length of stay by 24 hours or required a second, unplanned procedure. The cost of the index admission was also calculated as a secondary outcome. These outcomes of interest were determined before data analysis. Odds ratios and 95% CIs were estimated using multilevel logistic regression models.

Results  During the study period, 3519 laparoscopic cholecystectomies were performed; of these, 79.1% (n = 2785) were in girls, and 98.0% (n = 3450) were for gallstone disease. The overall morbidity rate was 3.9% (n = 137). After adjustment, patients with comorbidities were more susceptible to morbidity (odds ratio, 2.68; 95% CI, 1.78-3.86; P < .001). Operations for gallstones were less morbid. High-volume general surgeons had lower morbidity rates compared with low-volume pediatric surgeons (odds ratio, 0.32; 95% CI, 0.12-0.69; P = .005) independent of pediatric volumes. The mean (SD) unadjusted cost of a laparoscopic cholecystectomy was $4115 ($7273). Operative indication, complications, comorbidities, emergency admission, and surgeon volume were associated with cost.

Conclusions and Relevance  The high-volume nature of adult general surgery translated to lower morbidity and cost after pediatric laparoscopic cholecystectomy, suggesting that adult volume is associated with pediatric outcomes. As the rate of pediatric gallstone disease increases, surgeon volume, rather than specialty training, should be considered when pursuing operative management.

The landscape of pediatric gallbladder disease has changed over the past 3 decades.^1-3 While previously rare and largely attributable to hemolytic disorders, the prevalence of pediatric cholelithiasis has increased with the epidemic of childhood obesity.^4,5 Several estimates report the prevalence of cholelithiasis among children to be as high as 1.9% to 4%,^1,3,6 with the increase in prevalence being proportional to the rising rate of childhood obesity.⁷ One-third of American children are overweight or obese,⁸ and similar rates have been reported in Canada.⁹ At an approximate cost of $2721 per laparoscopic cholecystectomy,¹⁰ this increase in pediatric gallstone disease represents an important public health and economic concern.

While numerous studies have reported the increasing incidence of cholelithiasis in children, there is a paucity of data in the literature regarding outcomes of pediatric laparoscopic cholecystectomy. A report by the Canadian Pediatric Adverse Events Study¹¹ investigators indicated that surgical services account for 35% of all pediatric adverse events. Identifying the morbidity of common surgical procedures is paramount to successfully reducing the risk of such adverse events. This has been a principal aim of the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) and its pediatric equivalent (NSQIP-P).¹² Furthermore, specific to laparoscopic cholecystectomy, there is a disparity in surgical volume for the procedure between adult and pediatric surgeons, and, to date, the implications of this have not been previously explored.

Optimizing patient outcomes is a cornerstone of surgical care and health services research. Considering the paucity of research and the increasing incidence of pediatric gallstone disease, there is a necessity to better study the surgical care of gallstone disease and, more specifically, laparoscopic cholecystectomy in the pediatric population. Accordingly, the primary objective of this study was to determine the factors associated with all-cause morbidity and cost after laparoscopic cholecystectomy among Canadian pediatric patients.

This study was a retrospective, population-based analysis of administrative data. Patients 17 years and younger undergoing laparoscopic cholecystectomy in Canada from April 1, 2008, until March 31, 2015, were included, excluding the province of Quebec. Quebec was excluded because its data are not accessible directly from the Canadian Institute for Health Information (CIHI), but only through the Quebec Ministry of Health and Social Services. All cholecystectomies (inpatient and outpatient) were included to ensure accurate estimates of surgeons’ volumes. The analysis of the outcomes and costs was limited to inpatient cholecystectomies, which accounted for approximately 90% of all cholecystectomies. The data source was the CIHI Discharge Abstract Database. The CIHI database includes patient demographics, comorbidities, surgical procedure, and complications. This study received ethics approval from the Hamilton Integrated Research Ethics Board (HIREB). Informed consent from individual patients was not required because CIHI data are collected from a central agency with strict ethical oversight. Furthermore, data received by investigators from the CIHI are anonymized as per privacy protocols.

The primary outcome was all-cause morbidity. This was a composite outcome determined externally by the CIHI and included any documented complication during the index admission that extended length of stay by at least 24 hours or required a second, unplanned procedure. This outcome has been previously validated and is the most comprehensive clinically relevant outcome within the CIHI Discharge Abstract Database data set.¹³ A secondary outcome was the cost of the index admission. These outcomes of interest were determined before data analysis. Odds ratios and 95% CIs were estimated using multilevel logistic regression models. The CIHI uses a standardized costing method in which a resource intensity weight is assigned to each patient based on resource use. This is then multiplied by the mean cost per weight unit, which is controlled for the health region and year in which the admission occurred. The cost analysis was adjusted for inflation, and costs are reported in 2015 Canadian dollars.

Prognostic factors of interest included patient, hospital, and surgeon factors. Patient factors were sex, age, operative indication (gallstone disease vs other indication), patient comorbidities, emergency admission type, and complications. Hospital factors included volume (presented as effect per 20 pediatric cases). This was highly correlated with being a pediatric center; therefore, a prognostic factor for pediatric centers was not included. Surgeon factors were volume and specialty (pediatric surgeon vs general surgeon). Two different volume prognostic factors were used in this study. First, the total number of laparoscopic cholecystectomies completed, regardless of age, during the study period was used as total volume. This was to explore the association of adult cholecystectomy volume with outcomes after pediatric cholecystectomies by general surgeons. For total volume, high-volume surgeons were compared with low-volume surgeons. A high-volume surgeon was defined as one who had completed 50 or more laparoscopic cholecystectomies over the 7-year study period. This cutoff was empirically derived through a univariate exploration of the data. In addition, total volume and specialty were combined to explore the association of high-volume pediatric surgeons compared with low-volume pediatric surgeons and to avoid colinearity (because most pediatric surgeons were low volume). The second volume outcome was pediatric laparoscopic cholecystectomy volume, which was used to determine whether pediatric volume was singularly important for pediatric outcomes.

Descriptive statistics (means, medians, SDs, and interquartile range) were used to characterize the patient population where applicable. The χ² statistic was used to assess the association between all-cause morbidity and categorical variables, and analysis of variance (ANOVA) or t test was used for continuous variables where applicable. Risk-adjusted hierarchical regression models were used to determine predictors of morbidity and cost. To better delineate for unmeasured confounding at the hospital and clinician level and to account for correlation, surgeon and hospital identifiers were used as independent random effects for both models. Fixed effects included sex, age, operative indication, comorbidity status, admission type, surgeon total volume, surgeon specialty, annual surgeon volume (pediatric), annual hospital volume (pediatric), and complications (for cost analysis). Morbidity was modeled using a logistic hierarchical regression, whereas cost was modeled using the standard linear regression. Monte Carlo Markov chain estimation was used for all models, with 100 000 iterations after a 5000-iteration burn-in. All chains were examined for convergence. Statistical significance was set at 2-sided P < .05. Data were analyzed using Stata (version 12.1; StataCorp LP) and MLwiN (version 2.26; Centre for Multilevel Modeling, University of Bristol).

Over the study period, 3519 laparoscopic cholecystectomies were performed in Canadian children, not including those living in Quebec. Of these, 79.1% (n = 2785) were in girls, and the median age was 15 years. Few cholecystectomies were performed in children younger than 4 years. Further demographic information is listed in Table 1. Ninety-eight percent (n = 3450) of all cholecystectomies were performed for gallstone disease, and 31.7% (n = 1114) were conducted in an emergency setting. Hereditary spherocytosis accounted for the largest proportion of non–gallstone-related cholecystectomies (n = 16). The most common comorbidities were consequences of gallstone disease and included cholecystitis, choledocholithiasis, and biliary pancreatitis. Non–gallstone-related comorbidities included sickle cell anemia (n = 29), umbilical hernia (n = 21), and hereditary spherocytosis (n = 12). There was no association between comorbidity status and whether a patient received care by a pediatric or general surgeon. More than half of cholecystectomies in children were performed by high-volume general surgeons. The 2 most common complications were nausea or vomiting and pain.

Table 2 lists the univariate association between patient and health system factors and all-cause morbidity after cholecystectomy. Of the 3519 patients undergoing a laparoscopic cholecystectomy, 137 (3.9%) had some form of morbidity. The morbidity rate in boys was 6.3% (46 of 734) and in girls was 3.3% (91 of 2785) (P < .001). Patients with comorbidities had a morbidity rate of 9.4% (65 of 693) compared with 2.5% (72 of 2826) in their healthy counterparts (P < .001). Surgeon volume was significantly associated with morbidity rate, with the highest rate observed in low-volume pediatric surgeons. This finding also held true in the subset analysis for children younger than 12 years. Finally, hospitals with greater pediatric volumes incurred more morbidity than hospitals with low pediatric volumes. In the unadjusted analysis, emergency admission was not related to increased morbidity.

Unadjusted cost differences are listed in Table 3. The mean (SD) unadjusted cost of a laparoscopic cholecystectomy was $4115 ($7273). The most significant increase in cost occurred in cases completed for indications other than gallstone disease, with an excess of $15 393 (95% CI, $4980-$25 807; P = .004). Laparoscopic cholecystectomies were more expensive in cases with complications by $9153 (95% CI, $3808-$14 498; P = .001) and in children with comorbidities by $3490 (95% CI, $2343-$4636; P < .001). High-volume general surgeons were able to perform laparoscopic cholecystectomies for the lowest cost, with a decrease of $2615 (95% CI, −$3369 to −$1860; P < .001) from the mean. Before adjustment, annual hospital volume and clinician pediatric volume were also associated with an increased cost.

The adjusted odds of all-cause morbidity are listed in Table 4. The factors that significantly affected morbidity were operative indication (gallstone disease vs other), comorbidities, surgeon volume, and surgeon specialty. The morbidity rate of laparoscopic cholecystectomy in patients with indications other than gallstone disease was 8.07 (95% CI, 3.83-14.80; P < .001) times that of patients with gallstone disease. The morbidity rate of patients with comorbidities was 2.68 (95% CI, 1.78-3.86; P < .001) times that of their healthy counterparts. The morbidity rate for pediatric laparoscopic cholecystectomy was 68% lower when the operation was performed by a high-volume adult general surgeon compared with a low-volume pediatric surgeon (OR, 0.32; 95% CI, 0.12-0.69; P = .005). Within the CIHI database and using the definition for high-volume surgeons given in the Prognostic Factors subsection of the Methods section, only 7 high-volume pediatric surgeons were identified in Canada. After adjustment for other confounding factors, annual hospital volume and clinician pediatric volume were no longer significantly associated with morbidity. Emergency admission was again unrelated to morbidity. Owing to the few outcomes in the subset analysis for children younger than 12 years, this group could not be incorporated into the regression analysis.

Table 5 lists the estimated cost difference by patient and health system factors after adjustment for potential confounders. The most significant association with cost was seen for operative indication, with indications other than gallstone disease resulting in an increase of $12 706 (95% CI, $11 099-$14 302; P < .001). Complications significantly increased cost, with an excess of $6571 (95% CI, $5400-$7725; P < .001) as did comorbidities, with an excess of $2161 (95% CI, $1595-$2727; P < .001). Emergency admissions had an excess cost of $2069 (95% CI, $1540-$2591; P < .001). Compared with low-volume pediatric surgeons, high-volume general surgeons had a −$1970 (95% CI, −$3034 to −$866; P < .001) decrease in cost. There was no significant difference in cost for low-volume general surgeons and high-volume pediatric surgeons. Patient sex and age and hospital pediatric volume were not associated with a cost difference after adjustment.

In this large retrospective, population-based study among Canadian children, we found that indication for surgery, comorbidities, and surgeon volume were all associated with morbidity. Most important, this is one of the first studies, to our knowledge, to demonstrate that adult surgical volumes are associated with pediatric outcomes because high-volume general surgeons had an odds of all-cause morbidity that was 68% lower than that of low-volume pediatric surgeons. This study is also the first to examine the cost of pediatric laparoscopic cholecystectomy in Canada, and we found complications and emergency admissions, in addition to the 3 aforementioned variables at the beginning of this section, to be implicated in the cost differences in pediatric laparoscopic cholecystectomy.

The overall morbidity rate after pediatric laparoscopic cholecystectomy was 3.9% in our analysis. A similar study¹⁴ conducted in the United States in 2011 found an overall complication rate of 15% after pediatric laparoscopic cholecystectomy. The most common complication in that study was gastrointestinal, with a rate of 6.6%, compared with a rate of 0.6% for nausea or vomiting in our analysis. Another study from the United States, by Sacco Casamassima et al,¹⁵ used NSQIP-P data and found an overall complication rate of 13.6% for all pediatric laparoscopic cholecystectomies. Their study aimed to identify the safety of outpatient cholecystectomy in children and concluded that outpatient laparoscopic cholecystectomy was safe in properly selected patients. Their 30-day readmission rate was 1.4% and their 30-day reoperation rate was 0.7%.

Our study also examined the cost of pediatric laparoscopic cholecystectomy in Canada. We found that the mean unadjusted cost of a laparoscopic cholecystectomy in Canadian children was $4115. Cholecystectomies for gallstone disease cost less than those performed for other indications. Patient comorbidities, emergency admission, and complications all increased cost. Previous studies in the adult population have demonstrated higher costs for laparoscopic cholecystectomies performed in the emergency setting.¹⁶ Patient comorbidities and postoperative complications have also been shown to increase the costs associated with surgical procedures.¹⁷ Furthermore, our study confirmed a previously reported finding that high-volume general surgeons have a lower cost compared with low-volume pediatric surgeons. This cost difference was reported in 2012 by Chen et al,¹⁴ who demonstrated a cost savings when laparoscopic cholecystectomy was performed by high-volume surgeons.

Perhaps most notably, adult surgeon volume was found to be associated with all-cause morbidity independent of pediatric volumes. Low-volume pediatric surgeons had almost 3 times the morbidity of high-volume adult general surgeons. Within the CIHI database, there were only 7 high-volume pediatric surgeons, which did not allow for an adequate comparison of high-volume pediatric surgeons vs general surgeons. These findings are consistent with the 2012 study by Chen et al,¹⁴ which found surgeon volume, rather than specialty training, to be a better predictor of postoperative outcomes after pediatric laparoscopic cholecystectomy. However, their study did not explore the association of adult volumes with outcomes. Previous studies examining surgeon specialty and volume have found that pediatric surgeons have better outcomes after pediatric inguinal hernia repair¹⁸ and pyloromyotomy¹⁹ and have lower negative appendectomy rates compared with adult general surgeons.²⁰ However, in the case of laparoscopic cholecystectomies, it has been suggested that general surgeons have improved outcomes because they are able to “borrow” from their adult cholecystectomy experience, especially in older children with similar anatomy.¹⁴ Moreover, unlike the aforementioned procedures (inguinal hernia repair, pyloromyotomy, and appendectomy), the volume differential in laparoscopic cholecystectomy favors the general surgeons. The lower morbidity rate demonstrated by adult general surgeons was also seen herein in the subset analysis of children younger than 12 years, suggesting that the difference at the population level is not simply attributable to pediatric surgeons operating on younger patients. Finally, in our study, hospital volume was not associated with a difference in morbidity, suggesting that it is truly the clinician’s experience, rather than disparities in hospital infrastructure and resources, that accounts for the difference in outcomes. Our findings are likely more robust than those of previous studies because we explored both the adult volume and pediatric volume of general surgeons.

This study evaluated the outcomes and costs after pediatric laparoscopic cholecystectomy in a national cohort and has several important findings. Our study demonstrates the implications of distinct practice patterns for laparoscopic cholecystectomies between adult and pediatric surgeons. Whereas laparoscopic cholecystectomy is a common, high-volume procedure for adult general surgeons, this may not be the case for pediatric surgeons. In Canada, only 7 pediatric surgeons performed more than 50 laparoscopic cholecystectomies during the 7-year study period, and those who did not meet this volume threshold demonstrated substantially higher odds of all-cause morbidity. This is notable because, while volume-outcome relationships and cholecystectomies are not a major area of study for adult general surgery, it is perhaps more important in pediatric surgery owing to lower volumes. This finding is adjusted for pediatric surgical volume, which indicates that the volume and experience in the adult population is applicable to the pediatric population. In our study, the association with volume also affected costs independent of complication rates because cholecystectomies performed by high-volume general surgeons were almost $2000 less costly than those performed by low-volume pediatric surgeons. The reasons for this difference are beyond the scope of this study; however, it may be related to increased familiarity with the procedure and shorter lengths of stay. The public health implications of this study indicate that laparoscopic cholecystectomies, especially in older children, can be safely performed by adult general surgeons.

A strength of our study is that, to our knowledge, it is the first to examine outcomes after pediatric laparoscopic cholecystectomy in Canadian children, as well as the first to conduct a cost analysis. We used the CIHI Abstract Database, which is a highly accurate data source and allowed for examination of national data. Nevertheless, our study was susceptible to the limitations inherent in any large administrative database, including coding and recording error. Furthermore, given the few high-volume pediatric surgeons, our study was not sufficiently powered to detect a difference between high-volume pediatric surgeons and high-volume general surgeons; therefore, we can only draw conclusions based on surgeon volume and not specialty training. In addition, information regarding anesthesiologists was unavailable in the CIHI database; therefore, we were unable to track whether certain complications were attributable to the anesthesia technique. While there is some evidence suggesting that pediatric anesthesia training reduces the risk of major cardiorespiratory complications,^21,22 there is little evidence of its influence on complications like nausea or vomiting and pain. Finally, conclusions drawn from this study may not be applicable to pediatric patients treated on an outpatient basis; however, the outpatient population accounts for only 11.0% (435 of 3954) of the total number of cholecystectomies performed, which is a small portion of the population. Our study accounted for outpatient procedures when calculating surgeon volumes, but the outcomes and costs could not be tracked in this population.

In summary, morbidity after pediatric laparoscopic cholecystectomy is associated with the indication for surgery, patient comorbidities, and surgeon volume. These factors are also implicated in the economics of this operation. While the operative indication and patient comorbidities cannot be modified, surgeon volume, rather than specialty training, needs to be an important consideration from both a patient safety and economic perspective when Canadian children are referred for laparoscopic cholecystectomy.

Accepted for Publication: October 8, 2017.

Corresponding Author: Dennis Hong, MD, MSc, Division of General Surgery, Department of Surgery, St Joseph’s Healthcare, 50 Charlton Ave E, Room G836, Hamilton, ON L8N 4A6, Canada (perinoa@mcmaster.ca).

Published Online: January 17, 2018. doi:10.1001/jamasurg.2017.5461

Author Contributions: Drs Akhtar-Danesh and Doumouras had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Akhtar-Danesh, Doumouras, Hong.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Akhtar-Danesh, Doumouras, Hong.

Critical revision of the manuscript for important intellectual content: Doumouras, Bos, Flageole, Hong.

Statistical analysis: Akhtar-Danesh, Doumouras, Hong.

Obtained funding: Doumouras, Hong.

Administrative, technical, or material support: Bos.

Study supervision: Flageole, Hong.

Conflict of Interest Disclosures: None reported.