BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Modelling health scores with the multivariate skew normal - Jane H
 utton\, Department of Statistics\, University of Warwick
DTSTART:20100622T133000Z
DTEND:20100622T143000Z
UID:TALK23489@talks.cam.ac.uk
CONTACT:Michael Sweeting
DESCRIPTION:Health care interventions which use quality of life or health 
 scores often\nprovide data which are skewed and bounded. The scores are ty
 pically formed\nby adding up responses to a number of questions. Different
  questions might\nhave different weights\, but the score will be bounded\,
  and might be scaled\nto the range 0 to 100. If improvement in health over
  time is measured\,\nscores will tend to cluster near the 'healthy' or 'go
 od' boundary as time\nprogresses\, leading to a skew distribution.  Furthe
 r\, some patients will\ndrop out as time progresses\, so the scores reflec
 t a selected population.\n\nWe fit multivariate skew normal distributions 
 to data from a randomised\ncontrolled trial of four treatments for spraine
 d ankles\, in which scores\nwere recorded at baseline and 1\, 3 and 9 mont
 hs. In these data\, the scores\nat 3 and 9 months have skew marginal distr
 ibutions\, but the variance is\nsimilar across the four times points. We c
 onsider the extent to which\nvariance and skewness can be explained by cov
 ariates including treatment\nand age. In order to address the effect of cl
 ustering at the boundary\, we\nconsider censored multivariate normal and s
 kew model. The extended skew\nnormal is used to model the selection due to
  drop-out.
LOCATION:Large Seminar Room\, 1st Floor\, Institute of Public Health\, Uni
 versity Forvie Site\, Robinson Way\, Cambridge
END:VEVENT
END:VCALENDAR