Spatial Cox Model with Applications on Multiple Sclerosis Patients

• Activity Number: 427 - SPEED: Bayesian Methods, Part 2
• Type: Contributed
• Date/Time: Tuesday, July 30, 2019 : 3:05 PM to 3:50 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): HSIUCHING CHANG* and Hyokoung Grace Hong and Yu Yue
• Companies: IQVIA and Michigan State Universtiy and The City University of New York
• Keywords: Multiple sclerosis; EDSS; Cox model; ICAR ; time-varying; transformed Bernstein polynomial

Abstract

Multiple sclerosis (MS) is the most common cause of neurological disability in working-age adults and has a very high economic burden on the healthcare system. In US, the incidence of MS is estimated to be 2 cases per 100,000 person-years in men and 3.6 in women. To better understand how MS patients progress, we used Expanded Disability Status Scale (EDSS) symptoms to measure the level of MS disability (Mild, Moderate and Severe) over time. The focus of this study is to model the duration for a newly identified MS patient from the first time he/she was diagnosed to the time he/she experienced progression to severe disability level. As the incidence of MS diseases is closely related to a specific geographic distribution, the model also accounts for areal data with intrinsic conditional autoregressive (ICAR) prior. Besides, the disease modifying drug treatment and relapse frequency are included as time-varying variables in the Cox model where the baseline survival function is modelled as a transformed Bernstein polynomial to provide a better picture of patients’ transitory patterns.