BACKGROUND: Pancreatic cancer is becoming an increasingly important health problem in Spain. This study aimed to analyze the hospital management of this process and the factors associated with mortality by using an administrative data base. MATERIAL AND METHOD: We performed a descriptive study. Socio-demographic, clinical, diagnostic, and therapeutic variables of episodes registered in the national Hospital Discharge Minimum Data Set for 2004 were gathered. Comorbidity was assessed with the Charlson index. A logistic regression model was built to explain the individual influence of variables on in-hospital mortality. Mortality in the 17 autonomous regions of Spain was analyzed by using standardized mortality rates, through predicted mortality obtained from the multivariate model. RESULTS: The mean age was 68 years and men represented 56%. Readmissions accounted for 80% of the cases. The most frequent localization was in the pancreatic head and the most frequently employed procedures were computed tomography and surgery. In-hospital mortality was 25%, was higher in men, and increased with age. Mortality was higher in new admissions than in readmissions. Factors associated with higher mortality in the multivariate analysis were male sex, age, unspecified location or location in the tail, emergency admission, hospital stay, and comorbidity. Observed mortality was higher than expected in the Canary Islands and Madrid and was lower than expected in Catalonia and the Valencian Community. DISCUSSION: Future studies with more detailed information should be performed to allow the factors associated with in-hospital mortality from pancreatic cancer to be confirmed and to clarify the reasons for the geographical differences identified.
BACKGROUND:Pancreatic cancer is becoming an increasingly important health problem in Spain. This study aimed to analyze the hospital management of this process and the factors associated with mortality by using an administrative data base. MATERIAL AND METHOD: We performed a descriptive study. Socio-demographic, clinical, diagnostic, and therapeutic variables of episodes registered in the national Hospital Discharge Minimum Data Set for 2004 were gathered. Comorbidity was assessed with the Charlson index. A logistic regression model was built to explain the individual influence of variables on in-hospital mortality. Mortality in the 17 autonomous regions of Spain was analyzed by using standardized mortality rates, through predicted mortality obtained from the multivariate model. RESULTS: The mean age was 68 years and men represented 56%. Readmissions accounted for 80% of the cases. The most frequent localization was in the pancreatic head and the most frequently employed procedures were computed tomography and surgery. In-hospital mortality was 25%, was higher in men, and increased with age. Mortality was higher in new admissions than in readmissions. Factors associated with higher mortality in the multivariate analysis were male sex, age, unspecified location or location in the tail, emergency admission, hospital stay, and comorbidity. Observed mortality was higher than expected in the Canary Islands and Madrid and was lower than expected in Catalonia and the Valencian Community. DISCUSSION: Future studies with more detailed information should be performed to allow the factors associated with in-hospital mortality from pancreatic cancer to be confirmed and to clarify the reasons for the geographical differences identified.