MATH463 Project: Study Design and the Choice of Control Group in Clinical Trials

Dr Thomas Burnett

t.burnett1@lancaster.ac.uk

Department of Mathematics and Statistics - Lancaster University

You should submit an electronic copy to Moodle by 10am Monday 3rd February 2020.

  1. 1.

    Previous studies of Vinpocetine (an extract from the periwinkle plant) suggest that it improves the memory by increasing the blood flow in the brain. A neurologist suspects that the increased blood flow might also affect the IQ (as measured by the Wechsler Adult Intelligence Scale) by fostering cortical re-mapping. You have been approached to help design a Phase III clinical study and to prepare a trial protocol. The working hypothesis is that 10mg of Vinpocetine daily increases the IQ by at least 10 points: a difference of 10 points is clinically relevant.

    1. (a)

      Describe in detail how the study should be undertaken with clear motivation for: conducting the trial, the design choices made and the analysis planned. (18 Marks)

    2. (b)

      Derive the required sample size for your study ensuring that this derivation links to your planned analysis. You may assume that IQ measurements are normally distributed with a standard deviation of 15. The mean IQ without intervention is 100, a two-sided type-I-error of 10% should be used and the power should be at least 80%. (18 Marks)

    NOTES: If you need any further information from the neurologist clearly point out what information you require and make an assumption about it for the sample size calculations. You may also specify questions that you wish to ask, more generally.

    Parts 1a) and 1b) should be prepared as a trial protocol using ICH E9 and the lecture notes etc. The protocol should not exceed 4 pages of text using size 11 font and excluding references. You may also have an appendix tables and diagrams to illustrate your design plans.

  2. 2.

    A systematic review was undertaken to determine the efficacy of non-invasive pressure ventilation (NPPV) to treat respiratory failure resulting from exacerbations of chronic obstructive pulmonary (lung) disease (COPD) compared to standard medical treatment. The literature review yielded the following five clinical trial results and a ‘meta analysis’ of the primary response measure, which is mortality (death) following respiratory failure is to be conducted.

    Study number No. of patients No. of deaths
    NPPV Standard NPPV Standard
    1 29 29 7 12
    2 14 10 4 1
    3 30 30 5 15
    4 43 42 12 33
    5 118 118 22 35
    (a)

    What is ‘meta analysis’ and what is the motivation for conducting a meta analysis? (2 Marks)

    (b)

    Describe the two key methodological issues surrounding a meta analysis? (4 Marks)

    (c)

    Perform a test of heterogeneity clearly stating your hypotheses. (5 Marks)

    (d)

    Using the ‘inverse variance weighted’ method perform a meta analysis to provide an ‘overall’ estimate of the log-odds ratio together with a 95% confidence interval. (4 Marks)

    (e)

    Based upon your analyses interpret your findings in context of the mortality odds. (5 Marks)