Structured Additive Regression (STAR) Modelling of Women´s Age of Menarche and Years of Fertility in Central Portugal / Biomarkers Evaluation: Study Designs and Statistical Methods (Seminários no âmbito do Mestrado Bioestatística)


  • Prof. Doutor Bruno de Sousa – Instituto de Higiene e Medicina Tropical – UNL / Prof. Doutora Ana Luísa Papoila – Faculdade de Ciências Médicas da UNL, Ceaul
  • FCUL (DEIO) – Campo Grande – Bloco C6 – Piso 4 – Sala: 6.4.30 – 16:00 horas
  • Segunda-feira, 28 de Novembro de 2011

Abstract 1

The Portuguese Cancer League (LPCC) is a private non-profit organization dealing with multiple issues related to oncology, including the National Breast Cancer Screening Program. In this study, we analyze approximately 260,000 data records of women that entered the Screening Program for the first time in the central region of Portugal. It is believed that the period of time between the age of menarche and the age of menopause has been increasing over time. Therefore, a new variable called Window was defined as the difference between the age of menopause and the age of menarche, which represents a woman’s years of fertility. In this study, the evolution in time and space of the variables Window and the Age of Menarche were analyzed through STAR models, exploring the possible associations with other variables, such as Hormone Replacement Therapy, Pregnancy Status, Nursing Status, and Contraceptive Pills.

Keywords: Structured Additive Regression (STAR), Breast Cancer, Screening program, Biostatistics.

Abstract 2

Biomarkers are defined as “biological characteristics that are objectively measured and evaluated as indicators of normal biological processes, pathogenic processes or pharmacologic response to therapeutic interventions”. They are of great importance and their development is rapidly increasing due to the benefits they provide to both patient and clinical communities. In fact, their relevance has justified the delineation of new standards for study design and guidelines for reporting study, in order to ensure the good quality that is needed for this kind of medical tools. Among the different applications of biomarkers, classification and prediction are highly appreciated and used by physicians in their daily activities. In this talk, an overview of the study designs and methods that are most commonly used in the development and evaluation of a biomarker will be presented. An application of the reviewed methodologies will be applied to a real medical dataset.