The feasibility of using a postal survey method to assess the health and development of 7 year old children of different birth weight

Abstract

Study objective. To test the feasibility of using a postal survey to assess health and development in 7 year old children of different birth weights and to examine the response and any resulting response bias. Design. A multi-stage postal survey, using a sample stratified by birth weight. Setting. The four counties of Oxfordshire, Buckinghamshire, Berkshire, and Northamptonshire, which make up the former Oxford NHS Region. Sample. All children born in 1985 to the residents of the former Oxford region who weighed either under 1500 g or over 5000 g at birth and all those with unstated birth weight were included. Samples of approximately 130 live births were randomly selected from each 500 g birth weight band between 1500 g and 5000 g. The total sample size was 1319. Methods. The children in the sample were identified from birth registration and traced through the National Health Service Central Register (NHSCR). Self administered questionnaires were sent through the relevant Family Health Service Authority (FHSA) and then via the child's general practitioner (GP) to the child's parent(s). When parents' permission was given, self administered questionnaires were also sent to the child's GP and teacher. Main results. Of the 1319 children, 1169 were alive at the age of 7 and were successfully traced. Questionnaires were forwarded by GPs to parents of 1071 children and completed questionnaires were obtained from parents of 805 children. Parents of 753 children gave permission to approach the child's teacher and replies were received for 695. Parents of 770 children gave permission to send a questionnaire to the child's GP and 724 were returned completed. Response bias was assessed for the parents' questionnaire using data recorded about all births at birth registration. Response rates varied by sex of child, social class, and country of birth of the father. In addition, parents with a child with a serious motor or sensory deficit appeared to be more likely to respond. Conclusions. Response rates at each stage of the survey were good. The bias in response to the parents' questionnaire needs to be taken into consideration in future work. The method was found to be feasible and could be applied more widely than in monitoring child health.