![]() |
|
March 24th, 2006
Location: Zeus Grill and Seafood
6:00 pm Registration and Happy Hour (cash bar)
A latent structural model for postural instability process in Parkinson′s disease
The onset of postural instability (PI) is a key hallmark of Parkinson's disease (PD) progression. Since the true PI status is not clinically observable, it is assessed through a surrogate binary outcome. The biological feature of the PI process is that the process is inexorably decline, i.e., it is difficult to reverse by current known medications although a temperate relieve of symptom could be observed. However, the large inter rater variability leads to inconsistent results that contradict the irreversible nature of the PI process, making it difficult to determine whether a patient has lost postural reflexes. A novel Brownian motion based latent structure model is proposed that not only incorporates patient's multiple onset times, but also captures the specific variation structure of the PI process as well. The proposed model does not require even spaced visiting schedules. Two Bayesian diagnostic and outlier detection methods are proposed. The methodology is demonstrated through a PD clinical trial data.
Contact Matteo Bottai for more information.
|
[South Carolina Chapter Homepage] Copyright ©2005 American Statistical Association |
Modified February 2, 2005 |