Stroke is the second most common cause of death worldwide. However stroke is not invariably fatal and survivors may experience major physical, social and psychological problems. The United Kingdom government identified the improvement of the quality of life of stroke survivors as a key objective in the recent "Health of the Nation" consultative document. The concept of health related quality of life has developed over the past few years. Although there is no one universally agreed definition, there are several instruments which claim to measure at least some aspects of health related quality of life. These instruments have not been extensively tested in stroke patients, so the hypothesis which I shall test in this thesis is as follows: that a simple instrument can prove a feasible, valid, reliable, and clinically useful measure of health related quality of life in stroke survivors. I selected a simple measure of health related quality of life (the EuroQol questionnaire) and evaluated its validity in a sample of patients with stroke. A small, but important proportion of patients were unable to complete EuroQol questionnaires either by themselves or by interview, so I investigated whether a proxy (e. g. a spouse or carer) could assess the patient's health status after stroke accurately and without bias. Previous studies comparing one or more different health status instruments did not involve strictly random allocation, so could not provide reliable information on the "best" measure of quality of life to use in stroke patients. I therefore performed a study in a sample of survivors of stroke which directly compared the EuroQol and SF- 36 by using a strict random allocation of questionnaires. It was not possible to compare quantitatively the reliability and validity of the EuroQol and SF-36; however, a qualitative comparison suggested their reliability was similar and they appeared to be sampling broadly the same areas of health. I finally investigated patients' perception of their own quality of life after stroke. The data suggested that many disabled stroke survivors might not view survival in a dependent state as badly as one might expect. This somewhat surprising finding will inform decisions about whether to accept the high risks associated with certain treatments (e. g. thrombolysis) in order to reduce the chances of survival in a dependent state. Assessments of health related quality of life may therefore provide a more comprehensive and relevant view of the patients' outcome than simple measures of disability or impairment.