The Costs of Complications and Unplanned Readmissions after Pancreatoduodenectomy for Pancreatic and Periampullary Tumors: Results from a Single Academic Center

Simple Summary Complications lead to unplanned readmissions (UR) and are reported to be associated with a two- to threefold increase in hospital admission costs. Since healthcare costs are increasing worldwide, cost containment is the major challenge for future healthcare. In the literature, there are only a few studies that analysed hospital costs after pancreatoduodenectomy (PD). In this study, we aimed to create an understanding of the costs of complications and UR in patients who underwent a PD. Abstract Background/Objectives: Complications after pancreatoduodenectomy (PD) lead to unplanned readmissions (UR), with a two- to threefold increase in admission costs. In this study, we aimed to create an understanding of the costs of complications and UR in this patient group. Furthermore, we aimed to generate a detailed cost overview that can be used to build a theoretical model to calculate the cost efficacy for prehabilitation. Methods: A retrospective cohort analysis was performed using the Dutch Pancreatic Cancer Audit (DPCA) database of patients who underwent a PD at our institute between 2013 and 2017. The total costs of the index hospital admission and UR related to the PD were collected. Results: Of the 160 patients; 35 patients (22%) had an uncomplicated course; 87 patients (54%) had minor complications, and 38 patients (24%) had severe complications. Median costs for an uncomplicated course were EUR 25.682, and for a complicated course, EUR 32.958 (p = 0.001). The median costs for minor complications were EUR 30.316, and for major complications, EUR 42.664 (p = 0.001). Costs were related to the Comprehensive Complication Index (CCI). The median costs of patients with one or more UR were EUR 41.199. Conclusions: Complications after PD led to a EUR 4.634–EUR 16.982 (18–66%) increase in hospital costs. A UR led to a cost increase of EUR 12.567 (44%). Since hospital costs are directly related to the CCI, reduction in complications will lead to cost-effectiveness.


Introduction
A pancreatoduodenectomy (PD) is a complex procedure usually performed in cases of malignant tumours of the pancreatic head or surrounding structures such as the duodenum, the ampulla of Vater, or the distal bile ducts. Over the past decades, the mortality of pancreatic resections has been reduced to less than 5% by centralization of care. However, the overall morbidity of pancreatic resections is still high, ranging from 30% to 60% [1][2][3].

Materials and Methods
Since July 2013, all patients in the Netherlands who underwent a pancreatic resection have been registered prospectively in the Dutch Pancreatic Cancer Audit (DPCA). This study is a retrospective cohort analysis using the DPCA database of patients who underwent a pancreatoduodenectomy at our tertiary referral center, the University Medical Centre Groningen (UMCG), between 2013 and 2017. In total, 160 patients were included. Pre-, intra-and postoperative variables were obtained. All consecutive adult patients who underwent PD (classic Whipple procedure/pylorus-resecting pancreatoduodenectomy (PRPD) or pylorus-preserving pancreatoduodenectomy (PPPD)) between 2013 and 2017 at our institute were included in the study.
All treatment-related incurred healthcare costs of the index hospital admission and unplanned readmission (UR) related to the PD, from one day before PD until the date of discharge, were obtained from the financial department. These healthcare costs included all components of the surgical procedure, postoperative in-hospital care, postoperative hospital visits within 90 days after discharge, and in-hospital rehabilitation program, and concerned the actual individual patient-related costs that our institute incurred to treat the specific patient for PD. Therefore, a prolonged stay at the ICU or a longer surgical procedure resulted directly in higher costs. The costs for preoperative work-up were not included. For further analysis, the total costs were subdivided into seven cost domains: general diagnostics (e.g., laboratory, pathology investigation, and microbiology), imaging (e.g., CT or MRI scans), outpatient clinic (e.g., visits to the emergency department or outpatient clinic within 90 days after discharge), clinical care (ward care), surgical (operating room and surgical supplies), ICU (critical care on ICU department), other costs (e.g., percutaneous drainage or similar procedures and bloodtransfusion). Analysis was carried out using 2017 costs in euros (EUR). Complicated postoperative course and complications were graded by the Clavien-Dindo classification [12,13]. The Comprehensive Complication Index was calculated with the online calculator from www.assessurgery.com (accessed on 3 October 2021) [14]. The primary outcome measure was in-hospital costs calculated in euros (EUR). POPF, delayed gastric emptying (DGE), bile leakage (BL), postoperative bleeding, and chylous leakage were defined as stated by the definitions of International Study Group for Pancreatic Surgery (ISGPS).

Statistics
Data were collected in Microsoft Office Excel 2010 and later transferred to SPSS Statistics 24 (IBM, Somer, NY, USA) for statistical analysis. Cost analysis was performed using R 3.6.1 (R Foundation for Statistical Computing, Vienna, Austria). Descriptive statistics were calculated as percentages, median, or mean, as appropriate. For the cost calculations, the interquartile range was used. A Mann-Whitney test was used for univariate analysis of continuous variables if data were not normally distributed. In case the data were normally distributed, an unpaired t-test was used. A Kruskal-Wallis test was used if data was not normally distributed to analyse continuous variables for multiple groups. In case of nominal/categorical variables, a Chi-squared test or a Fisher's exact test was used; p-values < 0.05 were considered significant. Analyses were performed in SPSS v.24 (IBM, Somer, NY, USA)

Ethics
All data were retrieved from an existing database from the Dutch Pancreatic Cancer Audit (DPCA). Since we used the existing data register from the DPCA from the patients of our own hospital, no formal approval from the medical ethics committee was needed. Our local ethics committee waived the need for informed consent.

Baseline Characteristics
A total of 160 patients (53% men, mean age 65 years, median age 67 years) who underwent a PD between January 2012 and November 2017 were included for analysis. Table 1 presents an overview of the baseline characteristics of the population. An uncomplicated course after PD was observed in 35 patients (21.9%). A total of 87 patients (54.4%) had minor complications, whereas 38 patients (23.8%) had severe (Clavien-Dindo IIIA or more) complications. UR occurred in 40 patients (25%), 18 (45%) of whom had minor, and 19 (47.5%) of whom had major complications during index admission. A total of 125 (78.1%) of the patients underwent a PPPD, and 35 (21.9%) patients underwent a classic Whipple procedure. A vascular resection was performed in 26 patients (16.3%). In total, 97 patients (62.9%) underwent preoperative biliary stenting. In 80 patients (50%), the tumor originated from the pancreas. In 120 patients (70%), the pathology showed an adenocarcinoma. In total, 12 patients (7.5%) developed a clinically relevant (grade B/C) postoperative pancreatic fistula (CR-POPF) (see Tables 1 and 2 for baseline characteristics).     For patients with an uncomplicated postoperative course, the clinical care costs were EUR 5.845, while this amount increased to EUR 8.713 and EUR 16.276 for patients suffering minor and major complications, respectively (p = 0.001). Surgical costs seemed to be more or less the same for patients suffering no complications, minor complications, or major complications, and ranged from EUR 10.930 to EUR 12.210 (p = 0.642) (Tables 3 and A1).
For patients with Clavien-Dindo grade IV complications, the median costs in every cost domain seemed to be higher compared to the group of patients with an uncomplicated course or less severe complications (p = 0.001-0.023). ICU costs, general diagnostic costs, imaging costs, and other costs seemed to relatively rise the most in case of Clavien-Dindo grade IV complications. (Table A2).
The mean and median days of hospital stay were 11.7 days and 11 days, respectively, for patients with an uncomplicated course after PD. For patients with a complicated course (in general) after PD, the mean and median days of hospital days were 23.3 days and 17 days, respectively (p = 0.001). The mean and median days of hospital stay for patients with minor complications were 17.1 days and 15 days, respectively. For patients with major complications, the mean and median days of hospital days were 37.6 days and 31 days, respectively (p = 0.001) (Tables 3 and A1).

CCI Score
The median costs for patients with a CCI score of 0-30 ranged from EUR 27.350 to EUR 33.160. The median costs for patients with a CCI score between 30 and 50 ranged from EUR 40.246 to EUR 44.552. In cases of a CCI score of greater than 50, the median costs rose to EUR 93.311 (p = 0.001). The median costs for every different category rose when the CCI score increased. With clinical care costs, ICU costs and imaging costs increased the most in cases in which complications added up (Tables 4 and A3). Overall, the median costs seemed to be higher in cases of an unplanned readmission, except for patients with a CCI score of >50, but the number of patients in this group was small (n = 6) ( Figure A1). Figure 1A,B show the median total hospital costs by Clavien-Dindo classification and by the CCI score. Figure 1C shows a strong correlation (Pearson correlation coefficient = 0.77) between the CCI score and the log of the total costs. Therefore, hospital costs were directly related to the CCI score, with every single complication adding up in total hospital costs.  The mean and median days of hospital stay for patients with a CCI score of 0-30 ranged from 12.4 to 19 days and 12 to 18 days, respectively. The mean and median days of hospital stay for patients with a CCI score between 30 and 50 ranged from 28.4 to 34.6 days and 26 to 31 days, respectively. In cases of a CCI score of greater than 50, the mean and median days of hospital stay were 58.7 and 53 days, respectively. Hospital stays seemed to be longer when the CCI score increased (p = 0.001) ( Table 4 and Table A3).

Unplanned Readmissions
The median total hospital costs of patients with one or more UR were EUR 41.199 (IQR EUR 33.030-EUR 45.834, p = 0.001). For patients with UR and minor complications, the median costs were EUR 40.875 (IQR EUR 30.473-EUR 43.260). For patients with UR and major complications, these costs were EUR 41.739 (IQR EUR 36.211-EUR 52.347, p = 0.076). A UR was associated with an increase of EUR 12.567 (44%) in total hospital costs compared to the group of patients without readmissions (p = 0.001). This increase was mainly explained by an increase in general diagnostics costs, imaging costs, clinical care costs, and other costs, as these costs were higher in cases of a UR (p = 0.001-0.002) (Tables 5 and A4, Figure A2). The median hospital costs in patients without UR versus patients with UR categorized by complication type (no complications, minor complications and major complications) can be found in Table 6 and Table A5. The mean and median days of hospital stay for patients with a CCI score of 0-30 ranged from 12.4 to 19 days and 12 to 18 days, respectively. The mean and median days of hospital stay for patients with a CCI score between 30 and 50 ranged from 28.4 to 34.6 days and 26 to 31 days, respectively. In cases of a CCI score of greater than 50, the mean and median days of hospital stay were 58.7 and 53 days, respectively. Hospital stays seemed to be longer when the CCI score increased (p = 0.001) (Tables 4 and A3).

Unplanned Readmissions
The median total hospital costs of patients with one or more UR were EUR 41.199 (IQR EUR 33.030-EUR 45.834, p = 0.001). For patients with UR and minor complications, the median costs were EUR 40.875 (IQR EUR 30.473-EUR 43.260). For patients with UR and major complications, these costs were EUR 41.739 (IQR EUR 36.211-EUR 52.347, p = 0.076). A UR was associated with an increase of EUR 12.567 (44%) in total hospital costs compared to the group of patients without readmissions (p = 0.001). This increase was mainly explained by an increase in general diagnostics costs, imaging costs, clinical care costs, and other costs, as these costs were higher in cases of a UR (p = 0.001-0.002) (Tables 5 and A4, Figure A2). The median hospital costs in patients without UR versus patients with UR categorized by complication type (no complications, minor complications and major complications) can be found in Tables 6 and A5. The mean and median days of hospital stay were 17.1 days and 13 days, respectively, for patients with no readmissions. For patients with readmissions, the mean and median days of hospital days were 31.7 days and 26 days, respectively (p = 0.001) (Tables 5 and A4).

Discussion
This single academic center study revealed that complications after PD led to a EUR 4.634-EUR 16.982 (18-66%) increase in hospital admission costs. These costs were mainly due to an increase in clinical care costs. In cases of Clavien-Dindo grade IV complications, the median costs in every cost domain were higher compared to an uncomplicated course or less-severe complications (Clavien-Dindo I-III). In particular, ICU costs, general diagnostic costs, imaging costs, and other costs seemed to relatively rise the most compared to lesssevere complications (Clavien-Dindo I-III). A UR was associated with a EUR 12.567 (44%) increase in hospital costs. Log hospital costs seemed to be directly related to the CCI score, with a Pearson correlation coefficient of 0.77. Clinical care costs, ICU costs, and imaging costs especially seemed to rise the most in case complications add up.
With healthcare costs increasing annually worldwide, hospital cost containment will be one of the biggest challenges. One of the reasons for the increase in healthcare costs is the ageing population, resulting in patients with multimorbidity, and more often a complicated course after surgery. Postoperative complications have a high impact on quality of life and healthcare costs for society. A PD is a complex procedure often performed in specialized pancreatic centres. Although the mortality rate of pancreatic resections has been reduced over the past decades to less than 5% by means of centralization, leading to higher volumes and lower failure to rescue rates, the overall morbidity remains around 30% to 60% [6][7][8] Well-known complications after pancreatic resection include PF, DGE, haemorrhage, and deep and superficial surgical site infections [6][7][8]. The number of general nonsurgical complications such as pulmonary and cardiovascular complications rose in general with increasing age. A complicated postoperative course may lead to readmissions, and both were associated with high hospital costs. In the literature, there were only a few studies that analysed hospital costs after PD. Therefore, we aimed to gain an understanding in the costs of complications and unplanned readmissions in patients who underwent a PD. Previous studies have reported that a complicated course after PD increased hospital costs substantially, and may even lead to double or triple in-hospital costs, depending on the severity of complications [1,3,9]. Our study showed comparable results, but the effect was less pronounced. In our study, the median hospital costs for patients without complications were EUR 25.682. For patients with severe complications, the median costs rose to EUR 42.664. An UR was associated with a EUR 12.567 increase in hospital costs. In a study by Santema et al., the median hospital costs for patients without complications were EUR 17.482. For patients with grade I complications (defined as CD I-IIIa) the median hospital costs were EUR 28.380, and for grade II complications (defined as CD > IIIb), the median hospital costs were EUR 57.060 [1]. An American study by Enestvedt et al. showed comparable results to the study of Santema at al., with median hospital costs for patients with severe complications being significantly higher than for those without (USD 56.224 vs. USD 29.038) [3]. A study by Staiger at al. showed a strong correlation (Pearson correlation coefficient = 0.70) between log hospital costs and the CCI score [10]. Our study showed even a better correlation, with a Pearson correlation coefficient of 0.77. As complications are associated with severely increased hospital costs, continuous efforts are needed to reduce complications aiming to lower hospital costs for PD. Since our study and previous studies have shown that reinterventions under full anaesthesia or ICU admission (CD IIIb-IV) were associated with up to a threefold increase in hospital costs, early detection and management of complications is important to reduce both the length of hospital stays and hospital costs [1,3].
The following three strategies could reduce costs. Early detection and treatment of complications might prevent worsening of severity of complications. In 2018, the DPCG started the PORSCH trial, a nationwide stepped-wedge cluster randomized trial [15].
The aim of this trial was to evaluate if implementation of a best-practice algorithm for postoperative care focusing on early detection and step-up management of POPF could result in a lower rate of major complications and mortality after pancreatic resection, as compared to the current practice. Since hospital costs seemed to increase substantially after a complicated course after surgery, and every single complication added up in costs, as shown by the relationship between hospital costs and CCI, efforts to reduce complications and to minimize risk factors for complications will most likely be cost-effective soon. Additionally, insight into hospital costs could be a surrogate marker for quality control of surgical outcome.
Another strategy to reduce complication and thereby lower healthcare costs might be the introduction of surgical care bundles (SCBs) and enhanced recovery programs (ERPs) in pancreatic surgery. Over the past decade, the introduction of SCBs in colorectal and liver surgery have proven to significantly reduce surgical site infections (SSI) [16,17]. Various components of an SCB usually include hair removal prior to surgery, perioperative antibiotic prophylaxis, and perioperative normoglycemia and normothermia [17,18]. For pancreatic surgery, an overall SSI rate of around 20-30% was described in the literature [19,20]. In a study by Lawrence at al., an SCB for pancreatic surgery was implemented, and was found to decrease the SSI rate from 22% to 11% [20]. Furthermore, ERP implementation seems to be associated with a 50% reduction in risk for postoperative complications, as well as a reduction in the length of hospital stay by up to 2.4 days [18,[21][22][23]. Between 2013 and 2017, ERPs and SCBs were not yet implemented in our hospital. The combination of implementing SCBs and ERPs in pancreatic surgery might lead to a reduction in complications and thereby healthcare costs, but future studies are needed to confirm this hypothesis.
A third strategy to reduce complication and prevent worsening of severity of complications and thereby lower healthcare costs might be prehabilitation. Prehabilitation is a way to preoperatively optimize modifiable risk factors such as poor physical fitness, malnutrition, low mental resilience, iron deficiency anaemia, and smoking and/or drinking alcohol [23][24][25]. This can be achieved by physical exercise/training with a physiotherapist, optimizing a patient's diet with a dietician, and psychological support and/or coaching prior to surgery. In addition, optimizing hemoglobin levels prior to surgery by iron infusion is effective in reducing length of hospital stay and number of red cell transfusions [24,25]. Prehabilitation has shown in several studies to contribute to reduction in postoperative complications, faster physical recovery after surgery, and a shorter hospital stay, and is thought to lower healthcare costs in the end [24][25][26]. Previous studies have shown that prehabilitation reduces the overall and pulmonary morbidity after surgery [27]. Future studies to prove the cost-effectiveness of prehabilitation are needed, but since our study, as well as previous studies, showed that hospital were directly related to any increase in CCI, it is plausible that efforts to prehabilitate patients before PD or surgery in general might soon be cost-effective.
Our study had potential limitations. Only healthcare costs incurred in our institute were available for analysis. Since our hospital is a tertiary referral center for pancreatic surgery, some patients might have been readmitted in a local hospital nearby instead of the index hospital. This means that some costs might not be included in our study, meaning the overall costs of a complications and UR after PD might be even higher than previously stated. Nevertheless, we expect that the number of patients who did not go to our hospital for their complication was low, since patients were advised to go to the index hospital in case problems occurred. Another limitation was that we only examined hospital costs. Costs from an eventual admission in a nursing home or rehabilitation center after hospital discharge were not included. This means that the overall costs of a complications and UR after PD might be even higher than previously stated.
The strength of our study was that we used a larger study population (160 patients) and a more detailed hospital costs analysis compared to previous studies. Furthermore, our study and detailed cost overview served as a substantiation for a business model for a multimodal prehabilitation program in our hospital [11].

Conclusions
Our study showed that complications after PD led to a EUR 4.634-EUR 16.982 (18-66%) increase in hospital admission costs, mainly due to an increase in costs of clinical care. A UR was associated with a EUR 12.567 (44%) increase in hospital costs. Hospital costs seemed to be directly related to the CCI score. Therefore, every reduction in postoperative complications will contribute to an increased cost-effectiveness. Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Institutional Review Board Statement: All data were retrieved from an existing database from the Dutch Pancreatic Cancer Audit (DPCA). Since we used the existing data register from the DPCA from the patients of our own hospital, no formal approval from the medical ethics committee was needed. Our local ethics committee waived the need for informed consent.

Informed Consent Statement:
Our local ethics committee waived the need for informed consent.
Data Availability Statement: An anonymized version of our database can be provided upon request.

Conflicts of Interest:
The authors declare no conflict of interest.