Statistics

Statistics

What is statistics?

Statistics is a branch of mathematics dealing with the collection, analysis, interpretation, presentation and organisation of data. In applying statistics to a research problem, it is conventional to begin with a statistical population or a statistical model process to be studies.

Why you may want to include statistics support

  • to obtain an accurate power and sample size calculation; as this is integral to the trial being able to fulfill its objectives, funders of research expect that this is performed in a robust, justified fashion and the details of this calculation will be peer reviewed at many stages of the submission and approval process. As the study numbers impact directly on the cost of running the study, an accurate calculation is imperative in ensuring that costs are also accurate;
  • the results and conclusions from the study are dependent on a robust analysis performed using the correct statistical tests. Unless your particular study is extremely straightforward a qualified statistician would be required to fulfil this role; again this is expected by most, if not all, funding bodies;
  • to ensure that the quality and integrity of the data is maintained in addition to assisting with securing regulatory and ethical approvals and being involved in the running of any required DMEC.

What expertise can a statistics advisor offer?

  • advice on a wide variety of trial designs in addition to being in a position to provide expertise in non-trial work;
  • work with you to formulate a design appropriate for both your project and the targeted funding stream.

What to think about before meeting with a statistics advisor

  • for a sample size calculation it is dependent on the specific design under consideration but at the most basic level the following pieces of information could be required:
    • primary outcome measure & comparison of interest (defined in terms of a measurable variable at a specific timepoint)
    • typical values (e.g. mean in relevant population/existing data) of this outcome measure
    • what a clinically important difference would be on this measure
    • variability of this measure (e.g. SD estimate from relevant population/existing data)
  • details of proposed secondary outcome measures and any further comparisons that are envisaged. This information should be as detailed as possible in order to ensure that costs and time requirements are accurately estimated and sufficient for the proposed work;
  • Statistician time is generally costed into a bid on a co-applicant basis although for more basic work a merely advisory involvement may be possible; we will be able to advise on the appropriate arrangements given sufficient information.

Useful Resources

Piantadosi, S., 2013. Clinical trials: a methodologic perspective. John Wiley & Sons.

Kutner, M.H., Nachtsheim, C. and Neter, J., 2004. Applied linear regression models. McGraw-Hill/Irwin.

Molenberghs, G. and Verbeke, G., 2000. Linear mixed models for longitudinal data. Springer.

Klein, J.P. and Moeschberger, M.L., 2005. Survival analysis: techniques for censored and truncated data. Springer Science & Business Media.