BACKGROUND: In pancreatic cancer, surgical resection with neoadjuvant therapy improves survival, but survival relies significantly on the margin status of the resected tissue. This study aimed to develop a model that predicts margin positivity, and then to identify facility-specific factors that influence the observed-to-expected (O/E) ratio for positive margins among facilities. METHODS: This retrospective review analyzed patients in the National Cancer Database (2004-2016) with pancreatic head adenocarcinoma [tumor-node-metastasis (TNM) stage 1 or 2] who received neoadjuvant therapy for a pancreaticoduodenectomy. Logistic regression was used to develop a model that predicts margin positivity. This model then was used to identify outlier facilities with regard to the O/E ratio. Hospital volume was defined as the total number of pancreaticoduodenectomies per year. RESULTS: The study enrolled 4085 patients, and 16.8% of these patients had positive margins. Most of the patients (64%) had a tumor size of 2 to 4 cm, and approximately 51% of the patients did not have positive lymph nodes at resection. A logistic regression model showed that the predictors of positive margins after resection with neoadjuvant therapy were male sex, larger tumor size, and positive lymph nodes. This model was validated to yield a bootstrap-corrected concordance index of 0.632. The study calculated O/E ratios with the model, identifying 12 low- and 17 high O/E-ratio outlier facilities among 401 studied hospitals. The outlier hospitals did not differ in facility type (i.e., academic vs integrated network), but did differ significantly in terms of yearly hospital volume (low outlier of 20.6 vs high outlier of 10.7; p = 0.008). CONCLUSIONS: An association of lower-volume facilities with higher than expected rates of positive margins was found to indicate a disparity in care. This disparity was identified via an O/E ratio as a quality indicator for facilities. Facilities can gauge the efficiency of their own practices by referencing their O/E ratios, and they also can improve their practices by analyzing the framework of low O/E-ratio facilities.
BACKGROUND: In pancreatic cancer, surgical resection with neoadjuvant therapy improves survival, but survival relies significantly on the margin status of the resected tissue. This study aimed to develop a model that predicts margin positivity, and then to identify facility-specific factors that influence the observed-to-expected (O/E) ratio for positive margins among facilities. METHODS: This retrospective review analyzed patients in the National Cancer Database (2004-2016) with pancreatic head adenocarcinoma [tumor-node-metastasis (TNM) stage 1 or 2] who received neoadjuvant therapy for a pancreaticoduodenectomy. Logistic regression was used to develop a model that predicts margin positivity. This model then was used to identify outlier facilities with regard to the O/E ratio. Hospital volume was defined as the total number of pancreaticoduodenectomies per year. RESULTS: The study enrolled 4085 patients, and 16.8% of these patients had positive margins. Most of the patients (64%) had a tumor size of 2 to 4 cm, and approximately 51% of the patients did not have positive lymph nodes at resection. A logistic regression model showed that the predictors of positive margins after resection with neoadjuvant therapy were male sex, larger tumor size, and positive lymph nodes. This model was validated to yield a bootstrap-corrected concordance index of 0.632. The study calculated O/E ratios with the model, identifying 12 low- and 17 high O/E-ratio outlier facilities among 401 studied hospitals. The outlier hospitals did not differ in facility type (i.e., academic vs integrated network), but did differ significantly in terms of yearly hospital volume (low outlier of 20.6 vs high outlier of 10.7; p = 0.008). CONCLUSIONS: An association of lower-volume facilities with higher than expected rates of positive margins was found to indicate a disparity in care. This disparity was identified via an O/E ratio as a quality indicator for facilities. Facilities can gauge the efficiency of their own practices by referencing their O/E ratios, and they also can improve their practices by analyzing the framework of low O/E-ratio facilities.
Authors: Linda M Youngwirth; Daniel P Nussbaum; Samantha Thomas; Mohamed A Adam; Dan G Blazer; Sanziana A Roman; Julie A Sosa Journal: J Surg Oncol Date: 2017-04-13 Impact factor: 3.454
Authors: Ryan P Merkow; Karl Y Bilimoria; David J Bentrem; Henry A Pitt; David P Winchester; Mitchell C Posner; Clifford Y Ko; Timothy M Pawlik Journal: Ann Surg Oncol Date: 2013-12-13 Impact factor: 5.344
Authors: Piotr Knapik; Daniel Cieśla; Wojciech Saucha; Małgorzata Knapik; Michał O Zembala; Piotr Przybyłowski; Bogusław Kapelak; Mariusz Kuśmierczyk; Marek Jasiński; Zdzisław Tobota; Bohdan J Maruszewski; Marian Zembala Journal: J Cardiothorac Vasc Anesth Date: 2019-05-02 Impact factor: 2.628
Authors: Lauren E Colbert; William A Hall; Dana Nickleach; Jeffrey Switchenko; David A Kooby; Yuan Liu; Theresa Gillespie; Joseph Lipscomb; John Kauh; Jerome C Landry Journal: Cancer Date: 2014-01-03 Impact factor: 6.860
Authors: Ali A Mokdad; Rebecca M Minter; Hong Zhu; Mathew M Augustine; Matthew R Porembka; Sam C Wang; Adam C Yopp; John C Mansour; Michael A Choti; Patricio M Polanco Journal: J Clin Oncol Date: 2016-09-30 Impact factor: 44.544
Authors: Karl Y Bilimoria; Mark S Talamonti; Stephen F Sener; Malcolm M Bilimoria; Andrew K Stewart; David P Winchester; Clifford Y Ko; David J Bentrem Journal: J Am Coll Surg Date: 2008-06-30 Impact factor: 6.113