Textbook outcome as a novel composite quality measure in pancreatic surgery: A nationwide analysis

Abstract

Background: Quality assurance programs are becoming increasingly popular in surgery but require objective assessment of surgical outcome. Textbook Outcome (TO) is a multidimensional measure, reflecting the ‘ideal’ surgical outcome but has never been used in pancreatic surgery. Methods: Patients who underwent pancreatoduodenectomy (PD) or distal pancreatectomy (DP) for all indications between 2014-2017 were evaluated. Data were obtained from the Dutch Pancreatic Cancer Audit (DPCA), a mandatory nationwide registry. An international survey (24 experts, 10 countries, 4 continents) was conducted to reach consensus on the definition of TO in pancreatic surgery. Univariable and multivariable logistic regression was performed to identify predictors of TO. Between-hospital variation in TO rates were compared using observed-versus-expected rates, based on casemix-adjustment for baseline characteristics. Results: Overall, 3341 patients were included, of whom 2633 (79%) underwent PD and 708 (21%) underwent DP. Based on the survey (92% response rate), TO was defined by the absence of postoperative pancreatic fistula, bile leak, postpancreatectomy hemorrhage (all ISPGS grade B/C), unplanned intensive care admission, severe complications (Clavien-Dindo grade III or higher), readmission and in-hospital mortality. The overall proportion of patients that achieved TO was 60.3%, with 58.3% for PD and 67.4% for DP. On multivariable analysis, only class ASA 3 and 4 predicted a decreased TO rate after PD (OR 0.66 [0.48-0.90] and OR 0.18 [0.03–0.99]), whereas a dilated pancreatic duct (>3mm) was associated with an increased TO rate (OR 2.59 [2.15–3.11]). For DP, none of the parameters predicted TO. When comparing institutions, the observed-versus-expected rate for achieving TO varied from 0.70 to 1.50 per hospital after adjustment for casemix. Conclusion: Textbook Outcome is a novel quality measure in pancreatic surgery. The rate of TO varied considerably between pancreatic centers in the Netherlands, supporting the need for quality assurance programs.