P1-336 Validation of NHS hospital admission records for ischaemic heart disease in the million women study

Introduction The Million Women Study is a large longitudinal UK study of women's health. Between 1996 and 2001, 1.3 million women mostly aged 50–64 years were recruited through NHS breast screening centres in England and Scotland. Their current average age is 68 years. For all participants, the study has linkage to hospital admission records (Hospital Episode Statistics [HES] for England, Scottish Morbidity Records) as one source of study outcomes. Ischaemic heart disease (IHD) is one of the leading causes of disability and death in women in their 60s. However, the accuracy of capturing IHD in women using hospital admission records unclear. Methods The aim of this study is to validate the occurrence or absence of IHD in HES records in Million Women Study participants using general practice records. A random sample of 1000 women with a hospital record including an IHD diagnosis (ICD I20–I25) and 1000 women with no such admission (controls) were selected. We wrote to each woman's registered general practitioner (GP) asking for confirmation of an IHD diagnosis or not. Each GP was asked to complete a study form, supply relevant hospital documents and return these in a prepaid reply envelope. For each admission, HES data will be compared with GP data. Results Response rates were 83% for IHD and 88% for controls. Results will include analysis of the agreement between general practice medical records and HES admissions data for both cases and controls. The implications of using administrative datasets for epidemiological research outcomes will be discussed.