Migraine is a debilitating headache disorder, ranked by the World Health Organization as one of the leading causes of disability worldwide. Headache disorders are one of the commonest reasons a person will see a GP. Most patients seeing their GP for headache will have migraine and there is a significant unmet medical need for migraine treatments. Currently, migraine can be treated with acute analgesics, including NSAIDs (non-steroidal anti-inflammatory drug), triptans or with migraine preventatives, such as topiramate. However, many migraine sufferers are prone to develop intolerable side effects or in some cases, the migraine does not respond to treatment. In this study, we will examine records of drugs prescribed, not currently used in migraine treatment, to assess whether this treatment is associated with the resolution or worsening of headaches.
We will examine drugs, not used in the treatment of migraine, that have been prescribed for another indication to assess whether this treatment is associated with remission or relapse of headache by undergoing a cohort and two nested case-control studies. We will use: (1) Cox proportional hazards regression where the outcomes will be time to remission and time to relapse before moving onto multivariate Cox regression with time dependent covariates to explore potential nonlinear time trends. (2) Interrupted time-series analysis to measure trends in headache clinical and therapeutic events before and after a drug interruption. (3) Artificial neural network self-organising maps to build patient feature-maps (features would include: the drug of interest, covariates such as prescription records and comorbidities, time to remission and relapse). The map will reveal clusters of feature-similar patients, linking headache remission/relapse features to drug of interest.
Nested case-control studies will be used to explore the association between headache remission, persistence and relapse with prescription of drugs not used in migraine. Conditional logistic regression will be used to calculate odds ratio of a headache moving into remission or a patient moving into relapse. Critical variables will be controlled for and confounding factors will be taken into account in both cohort and case-control analysis.
Primary outcome:
1) Cessation of headache-related primary healthcare utilisation (a composite of GP consultation or analgesic prescriptions for headache and migraine)
Secondary outcomes:
1) Cessation of headache-related primary health GP consultations.
2) Cessation of headache-related analgesic prescriptions.
3) Cessation of headache-related neurology referrals.
Zameel Cader - Chief Investigator - University of Oxford
Anthony Nash - Corresponding Applicant - University of Oxford
Caleb Webber - Collaborator - Cardiff University
Clare Bankhead - Collaborator - University of Oxford
Richard Hobbs - Collaborator - University of Oxford
Patient Level Index of Multiple Deprivation;Practice Level Index of Multiple Deprivation