Medicines are tested in clinical trials prior to being licenced for use. These trials often exclude vulnerable patients and are unable to detect rare side-effects or those that occur a long time after starting the medicine. Therefore it is important to continue monitoring a medicine's safety once it starts to be used routinely. The UK Medicines and Healthcare products Regulatory Agency operates the Yellow Card scheme which gathers reports of potential side-effects from patients and healthcare professionals. However, just because a patient experiences a particular event while taking a medicine it does not mean that the event was caused by it. Therefore reports are carefully examined before a decision is made as to whether the medicine is likely to have caused the event. If this is considered to be the case then actions can be taken to inform people of the risk and prevent it happening. To inform this decision it is important to understand the characteristics of patients prescribed that medicine and the risk they would be at of experiencing the event even without taking it. This study will explore to see if the CPRD can provide data for this purpose and if it can inform the decisions made.
Despite use of signal detection algorithms the assessment of causality using spontaneous adverse event reports is complicated by the limitations of such schemes, which are subject to under-reporting and which lack comparator data in unexposed patients meaning that the impact of any biases in reporting is unclear. There is likely to be a benefit in placing potential safety signals arising from such reports into the context of the size and characteristics of the treated population at an early stage to assist assessment and that electronic healthcare record data could enable this. This is done routinely in vaccine pharmacovigilance through the use of observed vs. expected analyses and methodological and logistical approaches have been developed facilitating wider adoption of this approach for all medicines. This study will explore when and how descriptive analyses of CPRD data can more routinely provide useful insights that will increase efficiency and/or scientific robustness of regulatory assessment regarding safety signals earlier in the pharmacovigilance cycle. Analyses will be conducted at a drug substance and Read code level for drug-event pairs raised as signals and will include descriptive analyses of patients initiating treatment and/or experiencing an adverse event including exploration of any temporal relationship using chronographs.
The analysis will be applied to all adverse events of interest using Read codes at a pre-specified level within the code hierarchy so that the outputs are accessible if needed. However, the decision about whether the analyses will be specifically accessed and reviewed for a specific product will be made on the basis of signals arising from other sources (e.g. Yellow cards) that are discussed at an expert advisory group. Sub-Read codes will be combined with their parent code. The three pre-identified case studies will focus on salivary hypersecretion, alopecia, and cognitive disorder. Custom groups for combining Read codes to better identify a number of adverse events commonly associated with drug exposure will be included in the analyses and, if such events feature in signals raised, will be accessed and reviewed.
Katherine Donegan - Chief Investigator - MHRA
Katherine Donegan - Corresponding Applicant - MHRA
Alexander Smith - Collaborator - MHRA
Brian Burch - Collaborator - MHRA
Helena Bird - Collaborator - MHRA
Philip Tregunno - Collaborator - MHRA
Rebecca Owen - Collaborator - MHRA