The influence of institutional pancreaticoduodenectomy volume on short-term outcomes in the Brazilian public health system: 2008-2021

ABSTRACT Introduction: pancreaticoduodenectomy is a complex surgical procedure that can result in high rates of complications and morbimortality. Due to its complexity, the establishment of referral centers has increased in recent decades. This study aims to evaluate the influence of the institutional volume of pancreaticoduodenectomy for periampullary cancer on short-term outcomes in the Brazilian public health system. Methods: this study used a population-based approach and investigated the number of pancreaticoduodenectomies performed by institutions within Brazil’s public health system between 2008 and 2021. High-volume institutions were defined as those that performed more than two standard deviations above the mean number of procedures per year. Specifically, if a center performed eight or more pancreaticoduodenectomies annually, it was considered a high-volume institution. Results: in Brazil, 283 public hospitals performed pancreaticoduodenectomy for cancer between 2008 and 2021. Only ten hospitals performed at least eight pancreaticoduodenectomies per year, accounting for approximately 3.5% of the institutions. High-volume institutions had a significantly lower in-hospital mortality rate than low-volume institutions (8 vs. 17%). No significant differences between groups were observed for length of stay, hospitalizations using the ICU, and ICU length of stay. The linear regression model showed that the number of hospital admissions for pancreaticoduodenectomy and age were significantly associated with hospital mortality. Conclusion: institutional pancreaticoduodenectomy volume implies a lowering of in-hospital mortality. The findings of this nationwide study can affect how the public health system manages pancreaticoduodenectomy care.


Szor
The influence of institutional pancreaticoduodenectomy volume on short-term outcomes in the Brazilian public health system: 2008-2021

MATERIALS AND METHODS
This study, which employed a population-based approach, encompassed the inclusion of both available codes on the platform to include all cases submitted to duodenopancreatectomy through the Brazilian Public Health System (SUS).The data utilized in the analysis were extracted from DATASUS, the informatics branch of Brazil's public health system.Specifically, we relied on the SUS Procedures, Medicines, and OPM Table Management System (SIGTAP) codes '04.16.04.012-8'for 'duodeno pancreatectomy' and '04.07.03.020-4' for 'pancreato duodenectomy'.
However, it is important to acknowledge that this study has a limitation due to the reliance on manual inclusion of cases in the platform, which can result in underreporting or sub-notification of surgeries.The potential for incomplete data entry poses a constraint on the study's findings and must be taken into consideration when interpreting the results.
The data on the number of hospitalizations The study was approved by the local Ethics Committee (SGPP 5338-22) and did not require informed consent forms to be obtained.The raw data used in the study are publicly accessible.The STROBE statements were followed 13 .

RESULTS
In Brazil, 283 public hospitals performed 4,763 PDs for cancer between 2008 and 2021.Most procedures were performed in the Brazilian Southeast region, which accounted for 57% of total procedures (Figure 1).There was a trend in increasing the annual number of hospitalizations for PD in the country, but the in-hospital mortality was similar over the years (Figures 2 and 3).for LOS, hospitalizations using the ICU, and ICU length of stay.The mean age of patients treated in high-volume institutions was slightly higher than those treated in lowvolume institutions (59.32 vs. 57.74years old).Table 1 summarizes the comparisons between high-and lowvolume institutions.
In univariate analysis, the linear regression model showed that the number of hospital admissions for PD was inversely associated with hospital mortality (Coef: -10.667; p=0.009).In-hospital mortality and patient age were selected for multivariate analysis, and both were considered significant in the final model (Table 2 and Figure 4).

DISCUSSION
This large public health system study showed that institutional experience with PD impacts postoperative outcomes.This study's findings might    The Brazilian healthcare system probably lacks more significant control and centralization of PD care, which is a highly complex procedure and demands institutional expertise and expensive resources.Despite SUS stratifying healthcare into four levels of complexities, in which the most complex conditions would be treated in the third and fourth levels, maybe highly complex procedures, including PD, should be centered in only a few specialized institutions, following the strategy previously determined by the Brazilian healthcare system for liver transplantation services 24 .The third and fourth levels of complexity hospitals are usually overwhelmed, with high occupancy rates, and complex conditions end up being managed in hospitals with no expertise in oncology.Botega et al. 25 analyzed the spatial organization of the Brazilian health system and highlighted the need to reallocate resources to potentialize hospital utilization of inpatient care without increasing access inequities.
Concentrating PD in centers of excellence is an attractive solution for managers, improving postoperative outcomes and potentially reducing treatment-related costs.This strategy might be even more relevant in large public health systems, where demands for complex procedures are high while resources are scant.This strategy is also suggested for other conditions that demand complex care.Anacleto et al. 26  Namely, in this database, patients who developed severe complications and eventually died were not excluded from the analysis.This may be misrepresented as a short stay in low-volume institutions, which had higher mortality.
The mean age of patients in high-volume institutions was significantly higher than that of patients

CONCLUSION
Institutional high PD volume implies lower inhospital mortality.The findings of this nationwide study can affect how public health manages PD care and its resources.

R E S U M O R E S U M O
per patient, as identified by the Hospitalization Authorization (AIH), were obtained from TabNet/ DATASUS.Information on the length of stay (LOS), ICU stay, number of in-hospital deaths, and hospital costs for each hospitalization was also extracted.The population sizes for each Brazilian federative unit were calculated using data from the Brazilian Institute of Geography and Statistics (IBGE) from 2015.The number of hospitalizations related to PD for each Brazilian federative unit was expressed as a rate per 100,000 inhabitants using population size.The incidence graph was created using TabNet/DATASUS and Microsoft 365 (Office) Excel spreadsheet software.High-volume and low-volume institutions were compared.High-volume institutions were defined as those that performed an average of at least eight PDs per year from 2008 to 2022.Student's t-test was used to determine differences between groups for continuous variables, considering unequal variances.A 0.05 significance level was adopted.Linear correlation analysis was used to examine the associations between surgical volume and length of hospital stay, ICU stay, the proportion of patients demanding postoperative ICU care, and in-hospital mortality and was expressed as coefficient and robust standard errors (SE).Covariates with a p-value <0.05 in univariate analysis were selected for multivariate analysis.Scatter plots with the corresponding fit lines are also presented.We chose a cutoff of 8 annual PD based on evaluating which institution performed more than two standard deviations above the mean number of procedures per year, corresponding to Brazil's top 3.5% volume hospitals.Statistical analyses were performed with STATA 16.1 Software (StataCorp 4905 Lakeway Drive College Station, Texas 7).

Figure 1 .
Figure 1.Number of hospitalizations for pancreaticoduodenectomy according to the Brazilian regions.PD: Pancreaticoduodenectomy.

Figure 3 .
Figure 3. In-hospital mortality after pancreaticoduodenectomy in Brazil over the years.

Figure 2 .
Figure 2. Hospitalization for pancreaticoduodenectomy in Brazil over the years.
impact the management of administrative capacities, resources, and organizational structures for optimizing the treatment of periampullary cancers.PD is a technically demanding surgery that requires specialized skills from the surgeon.The critical location of periampullary neoplasms means that cancer resection comprises gastric, biliary, duodenal, and potentially vascular resections and reconstructions.Accordingly, periampullary neoplasm management demands the institutional availability of high-quality hepatobiliary and vascular surgeons, oncologists, endoscopists, and intensivists, among others.Each of these specialties must overcome a particular learning curve to achieve excellence.The institutional surgical volume translates the sum of the experience and learning acquired over the years of each specialty within the institution.

Figure 4 .
Figure 4. Scatter plot evaluating the number of hospitalizations for pancreaticoduodenectomy and the outcomes a) In-hospital mortality; b) Length of hospital stay; c) percent of patients demanding ICU care; and d) ICU length of stay.
evaluated thoracoabdominal aortic aneurysm repair in the Brazilian Public Health System and suggested the creation of specialized referral centers.Our data show that PD in-hospital mortality in Brazilian centers decreases from 16% to 8% when the patient is managed in centers of excellence.Due to the population-based nature of this study, some data regarding the cause of mortality and surgical complications are lacking.Although our findings do not demonstrate differences in LOS and ICU stay, the DATASUS database sums the length of stay for every patient, including those who die during hospitalization.

Table 2 -
Linear regression evaluating the outcome in-hospital mortality.PD: pancreatoduodenectomy; LOS: length of stay; ICU: intensive care unit; SzorThe influence of institutional pancreaticoduodenectomy volume on short-term outcomes in the Brazilian public health system:2008-2021