Spatial gender-age-period-cohort analysis of pancreatic cancer mortality in Spain (1990-2013)

• Published in: PloS one Vol. 12; no. 2; p. e0169751
• Main Authors: Etxeberria, Jaione, Goicoa, Tomás, López-Abente, Gonzalo, Riebler, Andrea, Ugarte, María Dolores
• Format: Journal Article
• Language: English
• Published: United States Public Library of Science 15.02.2017; Public Library of Science (PLoS)
• Subjects: Adenomatous polyposis coli; Adult; Age; Age Factors; Aged; Aged, 80 and over; Analysis; Bayesian analysis; Biology and Life Sciences; Cancer; Cohort analysis; Constraint modelling; Demographic aspects; Disease; Female; Females; Gender; Humans; Lethality; Male; Males; Mathematical models; Medical research; Medicine and Health Sciences; Middle Aged; Models, Biological; Mortality; Multivariate analysis; Operations research; Pancreatic cancer; Pancreatic Neoplasms - mortality; Patient outcomes; People and Places; Population; Public health; Research and Analysis Methods; Retrospective Studies; Sensitivity analysis; Sex Characteristics; Sex differences; Sex Factors; Spacetime; Spain - epidemiology; Spatial analysis; Studies; Trends; Spain
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0169751

Recently, the interest in studying pancreatic cancer mortality has increased due to its high lethality. In this work a detailed analysis of pancreatic cancer mortality in Spanish provinces was performed using recent data. A set of multivariate spatial gender-age-period-cohort models was considered to look for potential candidates to analyze pancreatic cancer mortality rates. The selected model combines features of APC (age-period-cohort) models with disease mapping approaches. To ensure model identifiability sum-to-zero constraints were applied. A fully Bayesian approach based on integrated nested Laplace approximations (INLA) was considered for model fitting and inference. Sensitivity analyses were also conducted. In general, estimated average rates by age, cohort, and period are higher in males than in females. The higher differences according to age between males and females correspond to the age groups [65, 70), [70, 75), and [75, 80). Regarding the cohort, the greatest difference between men and women is observed for those born between the forties and the sixties. From there on, the younger the birth cohort is, the smaller the difference becomes. Some cohort differences are also identified by regions and age-groups. The spatial pattern indicates a North-South gradient of pancreatic cancer mortality in Spain, the provinces in the North being the ones with the highest effects on mortality during the studied period. Finally, the space-time evolution shows that the space pattern has changed little over time.