Document Type: Research Articles
Department of Epidemiology and Silesia Cancer Registry, Cancer Center and Institute of Oncology, ul. AK 15, 44-101 Gliwice, Poland
Department of Epidemiology and Biostatistics, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran.
2Department of Radiotherapy, Cancer Center & Institute of Oncology, ul. AK 15, 44-101 Gliwice, Poland
Department of Radiotherapy, Regional Clinical Hospital, ul. Zyty 26, 65-001 Zielona Góra, Poland
Background: Clinical datasets for epithelial ovarian cancer brain metastatic patients are usually small in size. When adequate case numbers are lacking, resulting estimates of regression coefficients may demonstrate bias. One of the direct approaches to reduce such sparse-data bias is based on penalized estimation. Methods: A re- analysis of formerly reported hazard ratios in diagnosed patients was performed using penalized Cox regression with a popular SAS package providing additional software codes for a statistical computational procedure. Results: It was found that the penalized approach can readily diminish sparse data artefacts and radically reduce the magnitude of estimated regression coefficients. Conclusions: It was confirmed that classical statistical approaches may exaggerate regression estimates or distort study interpretations and conclusions. The results support the thesis that penalization via weak informative priors and data augmentation are the safest approaches to shrink sparse data artefacts frequently occurring in epidemiological research.