How can we use statistical methods to determine whether a new treatment is effective, and which patients might benefit most?

Abstract

To work out if a new treatment is effective for a particular disease, it is common to run a clinical trial. This involves recruiting patients with the disease, and randomly allocating half of them to the new treatment and the other half to an alternative, which could be the current standard treatment for the disease (if there is one) or a placebo. An important outcome is measured for every patient at a certain timepoint (e.g., mortality at 28 days) and then we use statistical methods to compare how many patients have this outcome in the two treatment groups, thereby determining whether the new treatment is effective or not. Often, multiple, similar clinical trials are carried out across the world. It is important to collate all the evidence together when making clinical decisions, and so we can use another statistical technique – meta-analysis – that quantitively synthesises the results from each of the similar trials together to estimate an average effect of the treatment. In my current role, I am involved in the design, conduct and analysis of clinical trials and meta-analysis for patients with prostate cancer and those with COVID-19.

In this seminar, I will describe how my Mathematics degree led me down this path to clinical trials and meta-analysis. Following this, I will talk about clinical trials and meta-analysis in more detail and discuss how we properly estimate the effect of treatment and describe how we can use these methods to work out which groups of patients might benefit most. I will finish with a recent example of a meta-analysis¹ that included patients hospitalised with COVID-19 and demonstrate how applied statisticians can impact how we treat patients during the pandemic.

References

[1] Shankar-Hari M, Vale CL, Godolphin PJ, Fisher DJ, Higgins JPT, et al. Association Between Administration of IL-6 Antagonists and Mortality Among Patients Hospitalized for COVID-19: A Meta-analysis. JAMA. 2021