On Friday 25th May Dr David McLernon from the University of Aberdeen will present his work at the University of Liverpool. All are welcome to attend this presentation at 1pm in Math 211 (Whittaker Room) within the Maths Department. There is no need to register in advance.
DYNAMIC PREDICTION MODELLING USING LANDMARKING: AN OVERVIEW WITH AN EXAMPLE IN INFERTILITY
Standard clinical prediction models attempt to estimate the probability of an outcome from a single point in time. However, in a clinical area such as infertility, their clinical utility has been limited by their inability to update the chances of this outcome (e.g. conception) at different time points. As couples with better prognoses become pregnant, those remaining represent an increasingly biased group with poorer prognoses. Reapplication of a model, developed at an earlier point when couples have better prognosis, to this group of patients, has a tendency to overestimate predicted success. Dynamic prediction using a landmarking approach allows the updating of predictions in one model whilst taking account of the fact that patients leave the risk set over time. Through application to record-linked fertility and pregnancy databases, David will show how the chances of conception with and without treatment diminish over a fixed time horizon.
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|