SUMMARY Obstetric and perinatal records have been assembled on 250 infant deaths and an equal number of live controls including 55 deaths associated with congenital anomalies. The information was used to construct a scoring system to identify high-risk infants at birth. Parents of 115 of the cases and their controls were also interviewed and all hospital, general practitioner, and health service records abstracted. Cases and controls were compared item by item in respect of all information available up to the age of one month and a scoring system constructed for use at one month.The 'at birth' and combined scoring systems are presented. The chance of death by age attained is presented for various risk groups. In a small prospective test, the multistage scoring system was nearly 50 % more effective than the birth score alone.In 1973 (Protestos et al.) we reported a retrospective study of the obstetric and perinatal states of children who presented later as unexpected deaths in infancy and compared the findings with a control group. From these data we developed a scoring system designed to identify high-risk infants at the time of birth.The system was evaluated in a 2-year prospective study in Sheffield. The first-year results indicated that the scoring system was highly effective (Carpenter and Emery, 1974, 1975;Emery and Carpenter, 1975). The findings at the end of the study are essentially similar. The high-risk control group had a relative risk of death 6 8 times that of the low-risk infants. It is estimated that 56% of the unexpected home deaths up to the age of 20 weeks occur in the high-risk group, which comprises 15 *7 % of all births, i.e. the sensitivity of the scoring system is 56 % and the specificity 84 -3 %.This scoring at birth was thus partially successful but there were still 44% of infants to be identified. In order to extend the system further we have attempted two things. First, to use data available at birth from a larger number of infants and controls to refine the birth score; and second, we have used recalled and recorded information about the first postnatal month from the parents of infants who died, and from controls, and have used that information to constitute a second scoring system at one Received 3 January 1977 month. We have also supplemented this information with data on hospital admissions.We report the information obtained from these three groups of data and the effect of combining them. Material and methodsOur earlier study was based on obstetric and perinatal records of 135 cases of unexpected infant death and 135 controls. Details of the method of selecting controls and the data abstracted from the medical records are given by Protestos et al. (1973). Since that study, data have been assembled on a further 115 cases of unexpected infant death that have since occurred in Sheffield or Rotherham, and on 115 controls. 39 of the new cases were excluded because death was associated with congenital anomalies, as were 16 in the earlier study, leaving 76 comparable new cases. As be...

No abstract

No abstract

No abstract

No abstract