Who is implicated in the care of digestive cancers? A population-based study over a 25-year period.

Objectives: The aim of this study was to study health care delivery patterns during the initial phase of the management of digestive cancers in the Côte-d'Or area between 1976 and 2000 and to learn about the impact of patient- and disease-related features on trends in care patterns.

Method: Health care patterns were divided into 3 categories depending on whether care was delivered by public or private health care providers or a combination of the two. A polytomous logistic regression model was used to search for factors associated with the choice of health care delivery.

Results: In the Côte-d'Or area, 12 055 new digestive cancer cases were diagnosed between January 1, 1976 and December 31, 2000, including 6 304 colorectal cancers. Public care structures managed 39% of patients, private care structures 42% and both (combined pattern) 19%. Women, young patients and patients with early-stage cancer were more often managed by private than by public facilities. Private health care institutions exclusively managed more colorectal cancers than public institutions (49.7% vs 34.2%) and fewer liver cancers (23.3% vs 62.8%) or pancreatic cancers (34.8% vs 47.5%). There was an increasing trend toward management by private instead of public institutions over the period. The proportion of exclusive private care rose from 35.0% (1976-1980) to 44.6% (1996-2000). Cancers diagnosed following symptoms concerned 85% of cases and were more often managed in private clinics. Cancers diagnosed in an emergency setting were more often managed in public institutions (48.3% vs 42.0%).

Conclusion: This study highlights the disparity of the characteristics of patients with digestive cancers depending on the type of institution delivering health care, with a trend toward an increasing role for private health care centres. The present findings suggest that data collected from a specific care pattern cannot be extrapolated to the general population.