Is enumeration district level an improvement on ward level analysis in studies of deprivation and health?

AIM: To investigate whether enumeration district (ED) level data reflect the aggregate characteristics of people living in that district better than ward level analysis. DESIGN AND SETTING: Reanalysis of the fourth national morbidity survey in general practice (MSGP4). Socioeconomic data on patients who had consulted 60 practices over one year were linked via postcode to ED and thence to small area statistics data for that ED and to the corresponding electoral ward. RESULTS AND CONCLUSIONS: There was no evidence that patients were likely to be more representative samples of the population of an ED than of a ward.