Statistical issues for the analysis of add-on epilepsy trials

Between 2% and 3% of the population will be diagnosed with epilepsy at some time during their lives and the majority of these patients will experience seizure remission. However, up to 30% will develop refractory epilepsy despite treatment with combinations of antiepileptic drugs (AEDs). For the management of these patients, the primary goal of treatment then becomes seizure reduction rather than remission. The Cochrane library currently includes 26 systematic reviews of randomised controlled trials (RCTs) of AEDs taken with other antiepileptic drugs, suggesting that many add-on therapies can reduce seizures in people with difficult to treat epilepsy. However, there are notable limitations with this evidence base. In particular, each review provides an overall comparison of clinical outcomes of individual AEDs against placebo whereas clinicians and patients want to know how AEDs compare against each other. Furthermore, the outcomes examined in previous trials and reviews may not necessarily be most relevant to patients and the approach to analysis may not be the most statistically efficient. Finally, few studies have explored how individual patient characteristics could be used to determine future prognosis and identify patients with refractory epilepsy that may respond better to particular AEDs. For these questions to be addressed reliably the Individual Participant Data (IPD) from each relevant RCT will be required.

In this PhD, IPD will be collected from eligible RCTs of add-on treatment in patients with refractory epilepsy. Statistical issues for the analysis of these trials will be explored and a network meta-analysis undertaken using the most efficient and statistically reliable methods. It is anticipated that the characteristics of trials and their participants will have changed over time due to the changing global nature of epilepsy research, the widening population of available AEDs, and changes in the classification and management of epilepsy. Patients within, and across, trials will also differ in the adjunctive AEDs taken alongside the experimental add-on AED or placebo. Such differences are likely to impact upon the underlying assumptions made in a network meta-analysis and the value of having IPD to thoroughly explore such issues will be fully exploited. The research would be completed with support from the Cochrane Epilepsy Group.

You should have, or expect to hold a BSc in Statistics or relevant discipline. An MSc in Statistics or relevant discipline is desirable.

*Please note the English Language Requirement for EU Students is an IELTS score of 6.5 with no band score lower than 5.5.

To apply please email your CV and cover letter to Rachael Kelly via [Email Address Removed].

For application enquires please contact Professor Catrin Tudur Smith [Email Address Removed].