Relationship between volumes and outcomes in colon cancer.

Background there is a wide debate in the literature about whether concentrating surgical procedure for colon cancer in larger hospitals could improve quality of care and outcomes. The aim of our study is to assess if there is an association between volume and surgical outcomes such as: reintervention in the index hospitalization, 30 day readmission and 30 day inhospital mortality in patients with a diagnosis of colon cancer in the surgical wards in Emilia-Romagna Region, Italy. Methods 14,200 colon cancer interventions performed from January 1st 2005 to December 31st 2010 were considered for analysis. Gender, age, comorbidities, presence of metastases, admission status and type of intervention were collected from hospital discharge records and considered as potential risk factors. We categorized ward volume as low (< 50 interventions per year) and high (50 interventions per year). We then performed a random intercept multilevel logistic regression with three levels (patients, surgical wards and hospitals). In the multilevel logistic regression model only the factors selected by using a stepwise procedure (p-entry = 0.05 and p-stay = 0.05) were included. Results preliminary results in the final model showed a significant association between volume and each of the three outcomes. The adjusted ORs of reintervention and 30-day readmission in high compared to low volume wards were 0.64 (95% CI = 0.450.93) and 0.83 (95% CI = 0.720.96), respectively. The adjusted ORs of 30 day in-hospital mortality in high compared to low volume wards were 0.45 (95% CI = 0.270.72) for elective patients and 0.82 (95% CI = 0.55–1.24) for urgent patients.