Location PSQ Devonport LT, University of Plymouth

Practical issues in medical statistics

Speaker: Gavin Koh (GSK)

Statistical methods provide a variety of options for designing a clinical trial: for example, superiority, equivalence and non-inferiority. Selecting a design depends not just on statistical considerations, but contradictory clinical, regulatory and ethical considerations. After completing a study, secondary multivariate analyses are often performed and further choices need to be made: Which exposure group should be chosen as comparator? Or should you fit a no-intercept model? How should the independent variables be parameterised?

Tafenoquine is a new chemical entity being developed as a novel treatment for malaria. I will illustrate these questions using data from a recently completed phase III trial of tafenoquine in malaria. This will be an interactive seminar with the audience asked to suggest and justify design solutions.

Dietary intake in the early years and its relationship to BMI in a bi-ethnic group: the Born in Bradford 1000 study

Speaker: Samuel Mahoney (Covance)

The number of infants, toddlers and children who were overweight increased from 32 million globally in 1990 to 42 million in 2013. This figure is predicted to rise to 70 million by 2025. In the UK it is estimated by 2020 that 20% of all boys and 33% of all girls will be obese.

Using data from the Born in Bradford 1000 cohort this study aimed to assess relationships between dietary intake at age 12, 18 and 36 months and BMI Z-scores at age 36 months in a bi-ethnic group.

Results indicate that dietary intake at 18 and 36 months was somewhat related to BMI Z-score at age 36 months and suggest the importance of early interventions aimed at establishing healthy eating behaviours.