Implementation of Best Practices in Pancreatic Cancer Care in the Netherlands

This study aims to improve the nationwide implementation of guideline-based best practices in pancreatic cancer care and assess the impact on survival.


eMethods 2. Deviations to the protocol
The study protocol described the inclusion of length and weight in the baseline characteristics, due to 80% missing values these are only described in the supplementary material.Additionally, descriptive statistics were planned to use to evaluate the adherence of the best practices.To determine the independent effect of the intervention, mixed-effect linear regression models using a random intercept for hospital and a random slope on intervention effect for hospital and adjusted for (calendar) time were performed instead.Consequently, adjusted percentages (considering the random intercept, -slope, and adjustment for calendar time) were reported.A sensitivity analysis before and after the implementation of the European Society of Gastrointestinal Endoscopy guideline on stenting was planned.This could not be performed, as this guideline was not published within our study period.Post-hoc analysis was performed assessing the number of patients who did not receive tumor targeted therapy, and the OS rates of the patients who received the best practice treatments (perioperative / palliative chemotherapy, and PERT).

eFigure 1. Schematic representation of individual hospitals within the 17 pancreatic cancer networks in the Netherlands
One colour reflects on regional network.The large dots represent the 17 centers for pancreatic surgery.The small dots represent the referring hospitals per network.

eTable 2. Median overall survival in months in subgroups receiving the best practices treatments Subgroups N (%) # Median OS
Outcomes are based on the non-imputed set, no missing data in survival outcomes.Median follow-up of patients alive for current practice: 35.9 months, best practice: 24.7 months.LAPC: locally advanced pancreatic cancer.# Percentage of patients receiving a specific therapy compared to the total (sub)group of patients: i.e. all patients, (borderline) resectable, locally advanced pancreatic cancer, and metastasized.eTable 3.

Outcomes QLQ-C30 eTable 3A. Global Health Score: area under the curve (0 to 12 months)
Area under the curve.LAPC: Locally Advanced Pancreatic Cancer.AUC: Area under the curve.CI: confidence interval.

QLQ-C30 QOL subscales: difference in current and best practices of delta (change between baseline and time point)
B: beta.CI: confidence interval.Bold numbers indicate statistical significance.eTable4.

QLQ-PAN26 subscales: difference in current and best practices of delta (change between baseline and time point)
Users of smartphone application per week e 2 (IQR: 1-5) Bold numbers indicate statistical significance.OR: Odds ratio.CI: confidence interval.WHO: World Health Organisation.Locally Advanced Pancreatic Cancer.MDT: Muti-Disciplinary Team.NA: Not Applicable.LAPC: a Number depicts number of patients that received chemotherapy, last date chemotherapy received only evaluated in 2018, attributing to the missing values.b Number depicts number of patients that received chemotherapy, 55% missing values on whether there were any complications, as this was only registered for patients within the PACAP PROMs.c Evaluated in patients in which a complication was registered during the study period c Evaluated in surgical patients registered in Dutch Pancreatic Cancer Audit that received either a plastic or metal stent.e Evaluated between 6-08-2018 until 17-11-2019, smartphone application was introduced to aid the physicians in the use of the best pratices.*Only evaluated the trials: PREOPANC (Netherlands Trial Register: 3709), PELICAN (ClinicalTrials.govIdentifier: NCT03690323), IMPALA (Netherlands Trial Register: 4079/4230), HALO, CROSS FIRE (ClinicalTrials.govIdentifier: NCT02791503).*Random intercept removed, to mitigate singularity errors of the random effects $ Evaluated in the Dutch Pancreatic Cancer Audit (surgical patients only).