Wednesday 13th October 2010, 2.00pm - 5.00pm at Manchester Dental Education Centre (MANDEC), Higher Cambridge Street, Manchester, M15 6FH (tea will be served about mid-afternoon)
(building 41, entrance on corner facing building 35) 

Joint meeting with Manchester University's Biostats Group

Theme: Statistics in Primary Care

Tim Peters (School of Clinical Sciences, University of Bristol)
Treatment of depression in primary care: trials, tribulations and the occasional triumph?

This seminar will use practical examples of the design and analysis of three rather different randomised controlled trials of primary care-based treatments for depression, to illustrate and evaluate the impact of applying various simple and complex trials methods.

As well as conventional challenges of standard sample size calculations and practical issues of recruitment in community-based settings, the more challenging aspects include: designing trials to investigate subgroup effects; multiple follow-up times; allowing for practice- and therapist-based clustering effects; missing data; and explanatory (causal) analyses. The approaches and the findings from the three example trials will be presented, along with general conclusions from the perspective of a trialist on the value and role of the various analytical methods employed for these studies.

David Reeves (University of Manchester)
Some alternatives to the Randomised Controlled Trial in Primary Care Research

For evaluating a new intervention, the randomised controlled trial is acknowledged to be the gold standard research design. However, for practical reasons an RCT is often not a feasible or even - on ethical grounds - acceptable method for researching an intervention in primary care. This talk presents some applications of other kinds of research designs in studies undertaken within the Primary Care Research Group at Manchester University.

Applications of four types of design are illustrated: the Interrupted Time Series; wait-list control trial, non-equivalent case-control study, and the factorial experiment. The benefits and limitations of each design as used in practice are discussed, together with the methods used to reduce risks to the validity of the research findings.

Sandra Eldridge (Queen Mary's, University of London)
Addressing community health issues in primary care: identification tools, pilots and cluster randomised trials

In this talk I will describe some of the statistical challenges that arise in research trying to address community health issues, using three very different examples: falls in older people, domestic violence, and the growth in TB in inner cities.

For all three, I will describe how relevant individuals in the community were identified (in two cases this involved designing specific tools), how interventions and trials were designed to address the issues and the findings and challenges of the eventual trials. Topics covered will include construction of tools, maximising the use of pilot data, and specific issues in the design, analysis, and data collection in cluster randomised trials.

If you would like to attend this free seminar please contact Wendy Lamb (+44 (0)161 275 5764):