Description of the condition
'Maternal mortality' is defined as the death of a woman during pregnancy or within 42 days of delivery, from any cause related to or aggravated by the pregnancy or its management but not from accidental or incidental causes (WHO 2004). 'Perinatal mortality' is defined as the stillbirth or death of a newborn baby within the first seven days of life (WHO 2006). 'Child mortality' is defined as the death of a child under the age of five years (UNICEF 2015). The maternal mortality ratio and child mortality rate are expressed per 100,000 live births (UNICEF 2015; WHO 2014). The perinatal mortality rate is expressed per 1000 total births (WHO 2006).
The United Nations' Sustainable Development Goals include reducing the global maternal mortality ratio to less than 70 per 100,000 live births and ending preventable deaths of newborns and children under five years of age, with all countries aiming to reduce neonatal mortality to at least as low as 12 per 1000 live births and mortality in children aged less than five years to at least as low as 25 per 1000 live births, by 2030 (UN 2017). Although progress is being made towards achieving these goals, it is not fast enough, especially in low-income countries (Wang 2014; WHO 2014). The absolute number of maternal, child and perinatal deaths, and the corresponding death rates, are higher in Africa than in any other region. In 2015, there were an estimated 303,000 maternal deaths globally, 99% of which were in low- and middle-income countries, and 66% in sub-Saharan Africa alone (WHO 2015). In 2016, there were an estimated 5,642,000 child deaths globally, more than half of which occurred in sub-Saharan Africa (UNICEF 2017).
Description of the intervention
'Death audit and review' is a broad term intended to include every different method of reviewing deaths, which we define as not only identifying the medical cause of death, but also identifying avoidable factors and making recommendations for avoiding such deaths in the future. The principle methods used are community-based audit (verbal and social autopsy), facility-based audits (significant event analysis (SEA)) and a combination of both (confidential enquiry).
In low-income countries without comprehensive death registration, deaths in the community are often investigated using verbal autopsy. The family of the deceased is interviewed according to a standard questionnaire (developed by WHO 2007), and the information is then interpreted by physicians or by a computer to ascertain the most likely medical cause of death (Waiswa 2010). However, there is usually no attempt to identify avoidable factors as it is assumed that it is already known which interventions are needed to tackle each principle disease. Verbal autopsy has been incorporated into wider health and demographic surveillance strategies (Adazu 2005), although its accuracy has been questioned due to the non-specific nature of signs and symptoms that may not be easily observed or remembered at interview (Butler 2010; Sloan 2001; Waiswa 2010). Social autopsy was designed as an add-on to verbal autopsy, and indeed the two are sometimes combined as a 'verbal and social autopsy' (VASA) (Kalter 2011). The aim is to make a 'social diagnosis', identifying avoidable factors prior to death in the home and community, within health facilities and at different stages of the patient pathway. In India this has been used in a participatory manner, which has been termed social audit for community action (SACA) (Nandan 2005). In this method, the community is asked to identify causes of death and avoidable factors. In this review we will not include studies investigating stand-alone verbal autopsy (whether conducted by a physician or a computer) with the sole purpose of identifying the medical cause of death.
Death audits in health facilities are usually based on SEA. This is an important part of revalidation for doctors in the UK, and the Royal College of General Practitioners (RCGP 2014) recommends that “SEA team discussions should be a routine part of your practice’s quality improvement and clinical governance”. Cases are usually discussed in a multidisciplinary team meeting (Hussein 2007). After discussing the details of the case, health workers identify avoidable factors and learning needs, and propose actions to be taken and changes to be made. The process does not intend to place blame, but the names of staff involved are not kept confidential. Indeed it is argued that “non-confidential straightforwardness and open-mindedness” are vital for a successful strategy (Supratikto 2002). A similar process occurs in 'mortality meetings', 'root cause analysis' meetings and, indeed, 'serious case reviews' (in child protection cases). Most mortality meetings take place at secondary healthcare facilities, drawing upon medical records to identify the diagnosis and key management interventions. Severe morbidity or near-miss reviews are also used to learn lessons; these review cases in which an individual almost died.
Confidential enquiry is the most comprehensive method by which to investigate deaths, because it considers not only the diagnosis and treatment in health facilities, but also the entire course of an illness and treatment-seeking pathway, to identify avoidable factors and to recommend changes at every level of the health and social care system and beyond, in order to prevent future deaths. This is particularly important in low-income countries where the majority of child deaths occur outside of any health facility (Breman 2001). A key feature of such enquiries is that the names of the individuals and any health workers involved are kept confidential, so that blame is avoided. These enquiries were pioneered in high-income countries, based entirely on written (usually medical) records examined by a multidisciplinary panel of experts, which includes not only health workers but also other professionals such as social services and the police (Lewis 2011; Pearson 2008). Such confidential enquiries have been useful for evaluating gaps in healthcare in the UK (Pearson 2008; WHO 2004), but are not yet widely used in low-income countries (Hussein 2007). The expert analysis involves both quantitative and qualitative elements. In the UK, all the included deaths were analysed quantitatively for basic information such as age, sex, socioeconomic status, location of death, time (seasonality) of death, and cause of death. Further detailed investigations were carried out on all maternal deaths and a subset of child deaths. A multidisciplinary panel reviewed each of these cases and identified avoidable factors. These were analysed thematically, illustrated by cases, and were used to generate recommendations as to how deaths might be avoided in future (Pearson 2008a).
How the intervention might work
Participation by communities in death audits is a strong basis for collective action to reduce mortality. In health facilities, significant event audit is a potentially powerful intervention to enable staff to learn from their mistakes and to institute important changes to procedures within their institution; the key mechanism is believed to be recommendation, then implementation of the proposed solutions (Pattinson 2009). The confidential enquiry approach is designed to identify avoidable factors at every step of the treatment-seeking pathway, and to make recommendations not only to improve the health system, but also to address avoidable factors outside of health facilities. Case review meetings, followed by the dissemination of recommendations to health workers, communities, or both, are essentially aiming to change clinician and patient behaviour. There are many theories of behaviour change, but these have been synthesised and integrated into the theoretical domains framework, which consists of 14 domains (Cane 2012; Michie 2014). Many of these domains are addressed by death audit and review. Those participating in the death review meetings gain knowledge about avoidable factors. The recommendations often set goals, and progress towards these can be audited. Repetition of similar recommendations may help clinicians to better remember guidelines, whereas social pressure may encourage them to better follow these guidelines. Death review meetings may also change health workers' beliefs about the consequences of their actions: the knowledge that deaths will be investigated and reviewed may motivate them to avoid poor practice. Discussing deaths, especially of mothers and children, often evokes an emotional response, which usually motivates health workers and parents to do all they can to prevent such deaths.
Death reviews may conceivably have some adverse effects. First, there is a cost (time and financial) to conducting death reviews. In the community, field workers need to be employed to investigate cases. In health facilities, staff are taken away from frontline duties to review cases, which may have an adverse impact on the delivery of care. It has been argued that these resources should instead be spent directly on implementing interventions that are known to be effective (Koblinsky 2017). Second, if death reviews are not handled sensitively, they may lead to blaming, humiliation and demotivation of staff, which may in turn lead to poorer quality of care. Third, focus on only one level of care (such as a district hospital) may lead to the diversion of resources away from other levels of care (such as primary care facilities). Fourth, there is the potential for inaccuracy – reviews based on indirect information (especially at the community level) may be incomplete or inadequate at diagnosing the likely cause of death.
Why it is important to do this review
The World Health Organization recommends that health facilities should conduct maternal and perinatal death reviews (WHO 2013; WHO 2016). In general, there is an underlying assumption that death reviews are useful and will impact on mortality but there is little robust evidence to support this (Pattinson 2005). It would be useful for policy-makers to understand which type of death review has the greatest impact on maternal, perinatal and child death rates, and what the essential features of an effective death review process are. Although confidential enquiry seems to be the most comprehensive method for addressing the whole range of avoidable factors, and hence has the potential to have the greatest impact, it is unclear whether it could be adapted, whether it would be feasible or whether it would be effective in reducing mortality in low-income countries.There is no comprehensive systematic review in the literature examining the impact of the aforementioned methods of investigating deaths.