PERSPECTIVES

Improving the use of confidential enquiries into maternal deaths in developing countries

Julia Hussein

IMMPACT, University of Aberdeen, Health Sciences Building, Foresterhill, Aberdeen AB25 2ZD, Scotland (email: j.hussein@abdn.ac.uk)

Confidential enquiries into maternal deaths aim to identify areas where clinical practice is deficient. This approach thi can improve practice and prevent maternal deaths.¹ Despite the general acceptance of the utility of confidential enquieries, achieving a high level of collective interest in their practical applicability as an evidence-based approach in developing countries will require further study. Although much has already been done in the promotion of confidential enquiries and related approaches, more action-based research, publication of systematic field experiences and conduct of studies comparing different approaches and adaptations will be of value. This paper discusses the reasons why confidential enquiries are not more widespread and describes the hypothesis behind ideas on how traditional confidential enquiry approaches could be adapted to make their application more user-friendly in developing countries.

Confidential enquiries related to maternal deaths are systematic, multidisciplinary and anonymous observational studies of adverse events. To conduct an enquiry, a committee is established, cases for inclusion are identified, assessments are carried out and a report with collated findings and recommendations for action at the policy level is prepared.¹ Relatively simple systems are required to set up an enquiry and well-organized health systems and reporting procedures facilitate the enquiry. Most experiences with conducting confidential enquiries are in developed countries and those where maternal mortality ratios have already fallen rapidly, including Egypt, Jamaica, Malaysia and South Africa. Paradoxically, where confidential enquiries are needed the most, i.e. in the lowest-income developing countries with the highest maternal mortality ratios, they are less commonly conducted.

Why confidential enquiries?

Maternal death reviews and clinical audits appear to be more widespread and have been conducted in many countries including Ghana,² Indonesia,¹ Nepal and Yemen. Many audits and reviews are not published because they are conducted as part of ongoing clinical practice, and so information on adaptations and difficulties in implementation are not easily obtainable. Just as with confidential enquiries, audits and reviews aim to identify the events leading up to poor maternal outcome and to address mismanagement and causes of avoidable death. The differences between audits, reviews and confidential enquiries are sometimes not obvious. Clinical audits measure against standards of good practice or criteria and usually involve an internal feedback loop, with health workers reviewing their own work. Maternal death reviews are also specific for a certain area or health facility.¹ In audits and reviews, the internal nature and lack of linkage to policy can result in apathy and poor follow-up of recommendations. Death reviews can be considered routine and a chore, so the quality of analysis is poor and confidentiality hard to preserve. Health personnel feel that fault-finding is the objective, and thus do not participate fully at meetings. Confidential enquiries, however, have a requirement of commitment from the bodies involved in high-level planning and policy. They do not necessarily assess all maternal deaths, and can draw from representative samples of deaths, so that resulting recommendations can be generalized for a specified area. Thus, confidential enquiries have the potential to make the greatest impact on maternal mortality due to their linkage to policy, as well as perceived prestige and representativeness.¹

Despite efforts to introduce confidential enquiries in developing countries for over 30 years, their acceptance remains poor. The reasons are likely to be complex and multifaceted, and include poor-quality documentation, lack of time, interest or accountability and poor organization of the health system. The capacity of health professionals in many developing countries may be overstretched, resulting in a reluctance to add to their already heavy workloads. Fears of litigation, exposure of ignorance and feelings of futility may exist. Moreover, as many maternal deaths occurring in the community either never present to a facility or arrive too late for timely clinical intervention, tracking of deaths and documentation of events become difficult. To overcome these barriers, several adaptations, other than clinical death audits and maternal death reviews, have been suggested. These include conducting the confidential enquiry process through time-limited studies; assessing only a sample of deaths, including "near-miss" events;³ and implementing subnational rather than national enquiries.¹

Promoting user-friendliness

Given the reasons for poor uptake, two possible modifications to the traditional enquiry process are proposed: addressing the negative image of confidential enquiries as critical and intimidating, and gathering additional information on community events prior to maternal death.

Building on strengths

There are high expectations on health workers, especially those who practice outside of health facilities or in rural areas. They usually work alone, have little supervision or mentorship and are expected to deal with a wide range of diseases in the community. Thus, it is not surprising that when they are asked to attend a maternal death review, they can feel challenged and unappreciated, even when the process is supposed to be non-punitive. If a death is reviewed in a confidential enquiry, the anonymity ensures that the risk of attribution of events to individuals is entirely removed. However, confidential enquiries focus on adverse factors and can be a particularly critical and negative process. In situations where resources are limited and conditions difficult, this targeting of negative aspects may be discouraging and overwhelming. Even clinicians involved as participants on enquiry panels in the United Kingdom have indicated dissatisfaction with the focus on adverse circumstances.⁴ Identifying favourable factors may help to redress this imbalance and alleviate the anxiety and defensiveness felt by health workers and health authorities. For every maternal death, it is likely that a complex series of circumstances occur which builds up to an adverse outcome. However, all circumstances leading to a death may not necessarily be adverse in nature. Deaths may occur despite a series of positive events and interventions. Conducting enquiries that identify both adverse and favourable events occurring during the course of a maternal death may allow existing strengths in the provision of care to be noted, thereby boosting morale and acknowledging good care, while also identifying areas for improvement.

Improving assessment of community events

It is usually recommended that confidential enquiries include events that occur in health facilities and the community.¹ However, the main source of information for many confidential enquiries is from clinical case notes, which do not generally document community events in detail; this is an acknowledged limitation of confidential enquiries.¹ Documented examples of how community factors can be ascertained are few, although in Egypt, testimonies of surviving family members were used during enquiry.⁵ Socioeconomics, access, resources and cultural factors in developing countries have led to the posting of health workers outside health facilities to improve the provision of delivery care for pregnant women, such as the multipurpose health worker in Nepal, the village-based health professional midwife in Indonesia and the auxiliary nurse–midwife in India. The assessment of the quality and effect of care provided in this manner has been problematic not only because community practitioners do not usually keep good documentation of cases, but also because poor documentation by trained health workers in both facility and community level care is not acceptable. In such situations, conducting a confidential enquiry can act as an intervention or catalyst to improve documentation (and practice), as well as being a means of eliciting information.

Social science and anthropological approaches have the potential to provide information for confidential enquiries. In-depth, unstructured interviews conducted with key informants (which include family members, traditional birth attendants and health personnel, such as trained health workers of various grades) may provide additional information on events preceding admission to a facility. The confidential enquiry panel can assess transcripts of the interviews in the usual manner. However, this approach is not without difficulty. Experience with verbal autopsy techniques, which determine the cause of death through interviews with family members, suggests that the information obtained through such means may not be dependable.⁶ Cognizance is needed regarding the varying quality of information from different key informants, the means to deal with contradictory perspectives from different people, panel members' level of experience and knowledge of community care, the skills of the interviewers who collect data, as well as the resources necessary for the additional effort of collecting community-based information.

Resource implications

One of the possible reasons for not conducting confidential enquiries in developing countries may be related to constraints of time and resources. Most developing countries are acutely short of senior experienced clinicians. Partial or non-attendance at panel meetings, incomplete assessment of case notes or unwillingness to take up further commitments due to busy work schedules may be very real limitations. There are currently few data on the time and costs required to conduct a panel enquiry, which is an important information gap to be filled.

Testing the adaptations

The utility of using the proposed adaptations will be tested as part of two large-scale evaluations being conducted in Ghana and Indonesia during 2006. In Ghana, the traditional confidential enquiry process will be adapted to encourage identification of favourable factors from clinical case notes. In Indonesia, a confidential enquiry will be conducted to identify favourable and adverse factors, assessing the care provided by professional midwives working in villages and to obtain further information on socio-economic, cultural and community factors. The research will ascertain whether the adapted approaches are workable, including issues such as acceptability, utility, time, monetary costs and intensity of effort.

Acknowledgements

I wish to thank the many team members of the Initiative for Maternal Mortality Programme Assessment (IMMPACT).

Funding: This study is funded by the Bill & Melinda Gates Foundation, the Department for International Development, the European Commission and United States Agency for International Development.

Conflict of interest: The author is a full-time employee of the IMMPACT initiative, through which this hypothesis will be tested.

References

1. Beyond the numbers: reviewing maternal deaths and complications to make pregnancy safer. Geneva: World Health Organization; 2004.

2. Wagaarachchi PT, Graham WJ, Penney GC, McCaw-Binns A, Yeboah Antwi K, Hall MH. Holding up a mirror: changing obstetric practice through criterion-based clinical audit in developing countries. Int J Gynaecol Obstet 2001;74:119-30.

3. Pattinson RC, Hall M. Near misses: a useful adjunct to maternal death enquiries. BMJ 2003;67:231-43.

4. Rankin J, Bush J, Bell R, Cresswell P, Renwick M. Impacts of participating in confidential enquiry panels: a qualitative study. BJOG 2006;113:387-92.

5. el Kady AA, Saleh S, Gadalla S, Fortney J, Bayoumi H. Obstetric deaths in Menoufia Governorate, Egypt. Br J Obstet Gynaecol 1989;96:9-14.

6. Sloan NL, Langer A, Hernandez B, Romero M, Winikoff B. The etiology of maternal mortality in developing countries: what do verbal autopsies tell us? Bull World Health Organ 2001;79:805-10.

(Submitted: 30 May 2006 – Final revised version received: 11 August 2006 – Accepted: 22 August 2006)