First speaker
Matthew Sperrin, Department of Mathematics and Statistics, Lancaster University.
Title
Modelling the effect of interventions on onset and progression of chronic disease.
Abstract
The present economic austerity is putting pressure on healthcare commissioners to target resources to local needs. However, health systems are faced with national policies that may be difficult to translate to local needs and contraints. We will present a policy decision-support model that can be used to explore locally relevant scenarios of healthcare intervention options and their simulated impacts on public health.
The model generates incident cases of specific long term conditions among a locally representative general population. We assume that an individual is born with a certain stock of health (SOH). On an annual basis, an individual's health stock depreciates. The rate of that depreciation depends on underlying fixed and variable risk factors and demographic factors, as well as random chance. Once an individual's health stock reaches a critical point they present with symptomatic disease or die from the condition before there is an opportunity for treatment.
The wider public health objective is to delay the overall population onset of the condition in question by slowing the decline in SOH, reducing incidence and ultimately increasing expectation of healthy life. Two strategies for intervention are: i) population interventions, in which the entire distribution of a risk factor is changed; and ii) targeted interventions, in which individuals with especially high risks of developing the disease of interest are targeted with preventive measures.
The SOH approach to modelling the onset and development of chronic disease in defined populations enables policy makers, planners and modellers to explore “what-if” scenarios of primary prevention. We will illustrate the approach with coronary heart disease intervention scenarios in a defined population.
Second speaker
Thomas Jaki, Department of Mathematics and Statistics, Lancaster University.
Title
Designing multi-arm multi-stage clinical studies.
Abstract
We consider the problem of multiple comparisons with control in a clinical trial with treatment selection at the interim(s). We introduce different, so called, "pre-planned" adaptive approaches which differ mainly in how treatments are selected and discuss how these approaches can be made fully flexible using the conditional error principle. We compare these approaches to fully adaptive ideas based on p-value combination and discuss advantages and disadvantages of each approach.
Organizers
This double seminar is co-organized with the Research Group in Statistics and Biostatistics at the Department of Mathematics.