How hard do general practitioners work?

assessment of the so called "avoidable factors" as in the Confidential Inquiry into Maternal Deaths.'4 This raises two problems. Firstly, the concept of avoidability, like that of causality, is highly subjective. Secondly, no one can know whether the absence of the "avoidable" factor would have prevented the death; indeed, the Northern region working party abandoned its search for "avoidable" factors. Other studies, including those in Exeter'2 and Leicestershire,'3 have avoided this dilemma by doing a case-control study.'5 The spate of perinatal mortality surveys which took place in NHS areas and districts in the late 1970s seems now to be subsiding. In some cases local surveys acted as pilot studies for regional surveys which have now superseded them. In others, the funding ran out, or the key person moved to another job. So what is the future? In Scotland, the perinatal mortality survey begun in 197716 has now developed so that it has been linked to the SMR2 maternity discharge data system and become part of that country's routine data collection system. 17 The regions of England and Wales might profitably follow this example when they too have a maternity data system-based either on the Standard Maternity Information System or the system proposed by the Steering Group on Health Services Information.8 Basing a survey on a system containing data about all births both reduces duplication of effort and increases the potential for selecting controls for case-control studies. Attention also needs to be given to the question of an appropriate cut off point for perinatal mortality surveys. Unlike perinatal mortality rates, late neonatal and post-neonatal mortality rates are not falling,4 and the principal factors behind many late neonatal and postneonatal deaths are determined at or around the time of birth (p 151 1).19 Some surveys cover only the perinatal period, while others cover the late neonatal period as well, but even these miss some relevant deaths. Calls have been made from time to time for a single national inquiry to be mounted in England and Wales as is done for maternal deaths. The consensus seems to be, however, that local and regional surveys are more flexible and allow attention to be directed to local problems. At the same time the local surveys would be more powerful if they had a common core of data.2021 At present the Scottish and Northern region survey teams are pooling their experience and making recommendations about how …