|11. What is the Evidence for the Role of Audits to Improve the Quality of Obstetric Care|
In general, all three broad dimensions of health care (structure, process and outcome) can be audited. Structure refers to the question: what facilities, equipment, staff etc. were there, process implies what was done to the patient and outcome questions what was the result for the patient in terms of mortality, morbidity or patient satisfaction. The number of potential topics for audit is virtually unlimited and the choice will largely depend on the local relevance of and the extent to which the problem can be successfully tackled. Clinical concern is undoubtedly the most important factor as it is unlikely that the audit group will develop sufficient enthusiasm for topics that they do not perceive as clinically relevant (Crombie et al . 1997).
Adverse outcomes, particularly mortality, have been the focus of obstetric audit for many years. One of the best known examples of audit, the Confidential Enquiries into Maternal Deaths, was primarily concerned with avoidable maternal mortality (Department of Health and Social Security 1982). The Confidential Enquiries, introduced in England in 1952, involve a systematic review of all maternal deaths and are aimed at identifying instances of unsatisfactory management in order to make recommendations for improvements in clinical care and service provision. The aggregate analysis is published at regular intervals (every three years in the United Kingdom) and includes an independent assessment of the nature of and the frequency with which substandard care was present. National or regional enquiries into maternal deaths have now been put into place in many countries, as are national or regional enquiries into stillbirths, neonatal or perinatal deaths (De Reu et al. 2000, Richardus et al. 1997, Bouvier-Colle et al. 1995, Mancey-Jones 1997). Because perinatal deaths are much more common than maternal deaths, they are also seen as a useful topic for obstetric audit at facility level (Mancey-Jones 1997, Ward et al. 1995).
In recent years, cases of severe acute maternal morbidity have emerged as a promising alternative to the investigation of maternal deaths. In particular, cases at the very severe end of the morbidity spectrum, the so-called near misses, are seen as a useful outcome measure for the evaluation and improvement of maternal health services (Stones et al. 1991, Filippi et al. 1996, Baskett & Sternadel 1998, Mantel et al. 1998). Severe obstetric complications have the advantage over maternal deaths in that they are more common and possibly less threatening to providers than deaths, and since the woman survives she can be interviewed about the care she received. The latter is important as it reveals aspects of quality of care such as patient satisfaction, that may otherwise be overlooked. However, unlike maternal deaths, severe obstetric morbidity is not so easy to define. Obstetricians may not agree on common criteria of severity and in most countries, severe cases are defined on the basis of management rather than clinical criteria. For obstetric morbidity to become a useful topic for audit, much more work needs to be done in the search for precise and reliable criteria of severity (Filippi et al. 1996, Ronsmans & Filippi 2000).
Other than mortality and morbidity, the most common type of audit is that of process, examining what was done to the patient in terms of investigations, diagnosis and treatment. The numbers of examples of process audits are endless, and only a few are listed here. One of the most prominent examples of process audit in obstetric care is that of caesarean sections. The widely observed increasing trends in caesarean section rates has been a cause for concern, and multiple investigations have compared the clinical indications for caesarean sections across regions or hospitals (Barrett et al. 1990, Rosenberg et al. 1982, Opit & Selwood 1979). Audit might then be used, for example, to set a desirable rate of caesarean sections in a particular type of health facility or to reduce the number of caesarean sections for specific indications, such as in women with a previous caesarean section (Joffe et al. 1994, Lomas et al. 1991). Other examples of processes of care that have been the subject of audit include the mode of delivery in breech presentations (Healey et al. 1997, Biswas & Johnstone 1993), the use of magnesium sulphate for the treatment of eclampsia (Taylor et al. 1998), and the use of a prostaglandin vaginal gel for induction of labour (Somerset et al. 1995). Clearly, all these processes represent clinically important problems. Whether or not they are a useful topic for audit largely depends on the willingness and enthusiasm of those involved in the audit to tackle the deficiencies that may emerge as a result of the review.