An Exposition to Integrated Modeling of Patients' Random Early Safety Signals for Effective Detections (Impressed): Cytogenetic, Molecular, and Hematologic Disorders

• Activity Number: 501
• Type: Contributed
• Date/Time: Wednesday, August 6, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Sandip Acharya and D. Das Purkayastha*+
• Companies: Novartis and Novartis
• Keywords: Integrated hierarchical modeling ; Effective detection ; Contour matrix ; Probability distribution.

Abstract

Effective and early detection of likelihood or pattern of incidence of adverse events is an important contemporary issue for safety analyses of clinical data. Identification of a disorder or a pattern of its incidence constitutes a signal of increased risk of a specific adverse event. But routine analyses are performed in a segmented way within a stand-alone body system rather than in an integrated set-up. Generally interdependent pattern of adverse events and lab values are ignored in traditional identification process of adverse events and lab assessments. In order to have effective detection of safety signals for a disease, this paper provides an integrated hierarchical model with sequential clustering of safety signals for multiple body systems. A bivariate pattern of safety signals is developed using a contour matrix system. Estimated conditional probabilities distribution function obtained from each individual system is used in the integrated model. Thus, this paper provides a statistical methodological approach with an application of the model primarily for effective detection, independence and relationships among cytogenetic, molecular, and hematologic disorders.