Reducing maternal mortality: reaudit of recommendations in reports of confidential inquiries into maternal deaths.

increase in the hospital specialties. This cohort was surveyed for the first time several years into the qualifiers' medical careers. We are cautious about interpreting recalled career choices because they may be influenced by the passage of time. None the less, the decline in choice of general practice as a first career between 1983 and 1988 and between 1988 and 1993 suggests that interest in general practice among newly qualified doctors has fallen gradually. We showed that, by 1995, the number of 1988 qualifiers who had left the NHS was no higher than that of 1983 qualifiers by 1990. However, there was a small but significant increase in the percentage of male doctors who were not in medical practice. We thank all the doctors who participated. We thank Janet Justice and Alison Stockford for coding data, Karen Hollick for help with tables, and Jean Davidson for assisting with programming. Contributors: Both authors designed the survey. TWL coordinated the conduct of the study, analysed the data, wrote the first draft of the manuscript, and contributed to further drafts. MJG suggested the use of capture-recapture methods and contributed to further drafts of the manuscript. Both authors will act as guarantors for the paper. Appendix Capture-recapture estimates of numbers of 1988 qualifiers working in NHS in Great Britain in 1995 were calculated, as shown in the table (above right), from numbers of doctors known to the Medical Careers Research Group and to the Department of Health. Confidence intervals for the number of doctors in the NHS were obtained by calculating the standard error of d using the formula SE = √ bc (a+b+1)(a+c+1) ```````` (a+1) 2 (a+2) and multiplying by 1.96 to give the half width of a 95% confidence interval. For example, the half width of the confidence inter-The value of capture-recapture methods even for apparent exhaustive surveys. Effect of variation in probability of ascertainment by sources (" variable catchability ") upon " capture-recapture " estimates of prevalence. The development of clinical audit over the past 10 years has led to questioning of the role of the triennial reports of the confidential inquiries into maternal deaths. Recently, the maternal death rate has been 6-7 per 100 000 maternities, with the proportion of deaths attributed to substandard care remaining around 40%. To investigate the uptake of the recommendations of the confidential inquiries into maternal deaths Hibbard and Milner audited the facilities in consultant …