The safety of medicines can be difficult to investigate and often the only possibility is to conduct observational studies to see whether people receiving a drug are more likely to experience a side effect than people who do not receive the drug. Recently more and more of these studies have been done and frequently their results have been of major public health importance (e.g. the lack of association between MMR vaccine and autism). It is important that the results of these studies are reliable, and yet their underpinning methodologies are subject to many possible biases. I plan to investigate some of these methodologies and explore their biases, with the objective of increasing the reliability of drug safety studies. This will be achieved through at least four studies using an electronic database of anonymised patient medical records. The studies will investigate possible links between SSRI antidepressants and cerebral haemorrhage, tiotropium (a treatment for chronic obstructive pulmonary disease) and stroke, ezetimibe (a lipid lowering agent) and cancer and glitazone antidiabetics and myocardial infarction. This programme of research will therefore answer questions directly related to the safety of four drugs/drug classes and will improve methodological approaches to drug safety studies in the future.