Please find below details of the Merseyside RSS Local Meeting to be held on Wednesday 22nd March at 3.00. =========================================================================================================================== Royal Statistical Society Merseyside Local Group Wednesday 22nd March 2006 3.00pm-4.30pm (Tea & Coffee @ 4.30 in Common room, 3rd Floor) Venue : Penthouse, Maths & Oceanography building University of Liverpool Epidemiology and Risk Please contact Ashley Jones ([log in to unmask]) to register a place ========================================================================================================================= New approaches to prognostic modelling with reference to a breast cancer benchmark Paulo Lisboa (Liverpool John Moores University) Prognostic modeling in cancer has relied heavily on two well-established methodologies, empirical modeling estimation using Kaplan-Meier methods and parametric estimation with the proportional hazards model. Recent developments in non-linear methods now extend the arsenal of modelling tools to more powerful estimators without prior assumptions regarding non-linearities and pre-set models of time dependence. This supports detailed analysis of covariate dependencies of hazard ratios, opening the door to valuable new insights and more accurate predictions for individual patients. The talk will use two studies of breast cancer - on a single risk, mortality, and on competing risks of intra-breast recurrence and distant metastases - to illustrate the capabilities of the new analytical methods, which have been developed by a collaborative effort with the Istituto dei Tumori in Milan. The role of statistics in health & safety policy making Simon Webster (Health and Safety Executive) The role of the statistician in the Health and Safety Executive (HSE) has changed considerably over the last 2-3 years and a high-priority focus of HSE's current work is based on the government's Public Service Agreement (PSA) targets, which were agreed in 2000. The targets include a pledge to reduce incidence of work-related ill health by 6% and reduce incidence rate of fatal and major injuries by 3% by 2007/08. Several 'programme teams' came into being, focussing their attentions on particular areas of concern including stress, musculoskeletal disorders and slips and trips. Given the obvious emphasis on statistics to gauge the programmes' progress towards the targets, statisticians have been involved to a great extent in planning and monitoring the programmes. This talk attempts to paint a broad picture of the kind of work this involves and the variety of challenges we face. These include planning and designing research, assisting the programme teams with setting targets for their own performance and in general providing an analyst's viewpoint on policy interventions and decisions.