[Pediatric cancer: analysis of hospital records].

OBJECTIVE To analyze pediatric cancer hospital records regarding the occurrence of new cases; to verify the relation between sex, age, race, origin and the disease clinical extension; to describe the mortality of children with cancer; to explore the association of new cases with proposed demographic variables, disease clinical extension and vital status. METHODS Observational, descriptive, longitudinal study in children up to 14 years old, from Santa Catarina. Three hundred and seventy-one new cases of cancer were treated in a regional reference outpatient clinic (1994-98) and recorded according to the International Classification of Childhood Cancer. Multiple Correspondence Analysis was applied to evaluate the categorical variables. RESULTS Three-hundred and seventy-one new cases of cancer were recorded with constant distribution in each year. Cancer affected preschoolers (41.5%) and males (55.8%). Leukemia was observed in 36.6% of the cases. Non-localized stage on the diagnosis occurred in 55.7%. The death occurred in 55.7% of the cases with non-localized disease and 16.3% with localized disease. At the end of the study 58.8% of the children were alive. CONCLUSIONS We observed more cases of cancer at pre-school age and among male subjects. Leukemia is the most frequent type of cancer. Non-localized disease predominates on the diagnosis. The death frequency is higher in the group with non-localized disease. There is direct association with non-localized disease, living vital status, school age, adolescent age, female, and for non-localized disease with death vital status, infant age, pre-school age and male.