Rapid responses and the Child Death Overview Panel - time for reflection and learning
Written by Simon Lenton
We are writing this short article based on our experience as designated doctors both involved in the process of death investigation including rapid response visits, chairing local child death review meetings and attending Child Death Overview Panel (CDOP) meetings, in order to promote discussion on how best to improve practice and outcomes. Although based on the experience in England, the observations should be relevant to both Scotland and Northern Ireland as they are starting to implement a child death review process.
Excess childhood mortality is a good example of a "public health issue" where there are widespread public and professional concerns, modifiable factors, affordable interventions and political will to intervene.
The size of the problem is well described in "Why Children Die" (http://www.rcpch.ac.uk/improving-child-health/child-mortality/child-mortality) which states that in the UK in 2012:
over 3,000 babies died before age one, mortality being strongly influenced by pre-term delivery and low birth weight; with risk factors including maternal age, smoking and disadvantaged circumstances,
over 2,000 children and young people died between the ages of one and nineteen years. Over three quarters of deaths were due to injury in 10-18 year olds,
suicide remains a leading cause of death in young people in the UK, and the number of deaths due to intentional injuries and self-harm have not declined in 30 years
Indeed this list could form the "agenda for action" to address avoidable mortality.
Looking back the death investigation process was introduced to improve multi-agency practice following high-profile concerns about miscarriages of justice after repeated sudden unexpected infant deaths and to improve the quality of support families received after a child dies.
Today the Local Safeguarding Children Board (LSCB) is responsible for:
ensuring there is a review of each death in the LSCB area,
determining whether the death was deemed preventable,
making recommendations to the LSCB or other relevant bodies so action can be taken to prevent such deaths in the future at either local or national levels.
To do this the CDOP considers the modifiable factors in parenting capacity, local environment or within service provision that constitute a public health or individual safety concern arising from a particular death or from a pattern of deaths. The aggregated findings from all child deaths should then inform local strategic planning on how best to safeguard and promote the welfare of children in the area.
So how well has the process of the rapid response and child death overview panels met the challenge of reducing overall mortality?
The rapid response visit
Which children require a rapid response? Certainly those who die suddenly, unexpectedly and where there are concerns about the cause of death. But should this include road traffic accidents, children with life limiting conditions or the infant who dies after major cardiac surgery in intensive care? Deaths of children known to safeguarding teams will be investigated with a serious case review if there are safeguarding concerns or significant service failures.
How competent is the rapid response team? Death investigation requires a significant input from social care, the police and health services and our experience has been that every death involves a different set of individuals who often have never worked together before. Is this unavoidable given the infrequent nature of sudden unexpected deaths or could the system be organised differently? In an ideal world experienced people from each agency should be involved in order to make the most appropriate decisions to prevent scenarios such as police investigating meningococcal septicaemia on a paediatric intensive care unit. Perhaps the best option would be for the individuals involved to at least meet and train together, before working together.
How should the health component be provided? Some places have a dedicated consultant paediatrician rota available 24/7 whereas others have a health visitor rota working office hours only. Our experience is that the most deaths occur outside office hours so the obvious suggestion would be to combine the child death rota with the child protection on call rota. Further studies of the different options to determine what works best are required.
Has the quality and outcome of investigation for Sudden Unexpected Death in Infancy improved? Our experience would suggest that a process of standardised investigations following sudden unexpected infant deaths, followed by a paediatric post-mortem has certainly improved the process, although there are still concerns about the timeliness of some post-mortem reports. The challenge remains to implement more effective strategies to reduce co-sleeping, smoking and substance misuse to further reduce deaths in this group.
Has support following death improved? Most places do not have a formal "bereavement support service" and the responsibility for family support may rest with police, social care or health depending on circumstances. It is important that families have a "named contact" that is knowledgeable about death, grief and local procedures. Generally families find feedback following the local case review and the opportunity to ask questions a helpful process.
What about support for rapid responders? Bring involved with the families of children and young people who die is often stressful, particularly for professionals who are less experienced or those who have their own children of a similar age. Currently little support is on offer for those involved and this is a potential area for improvement.
The Child Death Overview Panel (CDOP)
Which deaths should be reviewed? Currently all deaths are reviewed by the CDOP panel but is this necessary? Approximately 50% of the deaths are either in the perinatal period or related to poor perinatal outcomes. In theory, all of these deaths should have been reviewed within the neonatal network morbidity and mortality meetings, with lessons being learned on how to improve practice implemented within the network. The "added value" of a second review by the CDOP should be minimal if the neonatal network process is working well. Likewise the specialist networks such as cardiac networks or cancer networks should be reviewing mortality as part of their internal quality improvement programmes.
It is often difficult given the skill set around the CDOP table to know whether the right decisions were made by highly specialised teams are example, was the surgical intervention appropriate for clinical findings? Were the right decisions made in intensive care at the right time? Were the right drugs and the right dose used? Only clinical colleagues and their peers can competently discuss these issues. Our experience is that the "themed reviews" on a regional basis may have some value particularly for issues like suicide, drowning or road traffic accidents, but detailed clinical discussions should be reserved for meetings within specialist networks.
Is the CDOP process for identifying modifiable factors robust? The critical process for the CDOP is to identify modifiable factors in order to plan interventions to prevent future deaths. However, this is something of a "Catch-22" situation, because to know whether something is "modifiable" requires knowledge of what interventions are effective. The terms modifiable, controllable, preventable, amenable (to health service interventions) and contributory factors are all used inconsistently within and between CDOP's. Take for example, a 15 year old, who smokes, delivers a 26 week preterm infant at home, without prior antenatal care; then the ambulance is delayed arriving and takes the baby to the emergency department rather than the neonatal unit and the baby dies. What factors are modifiable and what recommendations should the CDOP make? How should the CDOP decision-making and coding system develop to ensure consistency on a national basis? Our experience would suggest that this whole issue requires resolution if national statistics are to be interpreted appropriately.
So what happens to the data and recommendations for action from CDOP's? Data from CDOP's is collected nationally but not systematically analysed and therefore not linked to public health interventions or national strategies to reduce the deaths of children.
Are the lines of reporting and accountability for CDOP's correct? Child Death Overview Panels report to Local Children Safeguarding Boards and while LSCBs have a responsibility for health and well-being, most of the time they focus on safeguarding so recommendations from CDOPs to address wider issues may be more difficult for LSCBs to take forward. An obvious example would be improving the training of first contact practitioners in the recognition of sick children.
Could the value of CDOPs be improved? There has not been a step change in mortality trends following the introduction of the child death investigation process. There are certainly potential duplications between neonatal and specialist network mortality reviews and the CDOP. Trends in infrequent causes of death such as blind cord strangulation or gunshot injuries are not currently identified, as national statistics are not available.
What is the right approach? The whole child death review process was initially driven by the need to improve investigation of sudden unexpected infant deaths, but then expanded to cover all deaths assuming the process would be similar, however, investigation for road traffic injuries and suicide require a very different approach.
There are many types of investigation/analysis and it is not always clear how to choose the right public health tool for the job. It would be helpful to review investigation methodologies and their ability to correctly identify modifiable factors relevant to different scenarios.
One example is the Haddon matrix which was designed for road traffic injuries and now reproduced below. Originally designed by William Haddon in 1970 it creates a matrix to aid the identification of modifiable factors and create a framework to help develop preventative strategies.
The Haddon matrix for a road traffic accident
|Human (victim)||Agent (Hazard)||Environment|
child behavior e.g. ADHD
e.g. braking capacity
visibility e.g. parked cars
resistance to injury
e.g. sharp edges
|street design e.g. lighting, road service|
e.g. access to first aid
Rapidity of energy absorption
e.g. bumper design
|emergency response e.g. ambulance times|
The same basic principles can be applied to infectious diseases.
The Haddon matrix for an infectious disease
|Human (victim)||Agent (Hazard)||Environment|
exposure to pathogens
awareness of signs
|event||Recognition of illness||
control of vectors
Access to primary care
Access to emergency care
|Emergency care responses|
This matrix could be modified, depending on circumstances, for example, the human elements can be expanded to include past and present vulnerabilities, the agent could be physical, chemical, disease or human (as in child protection) and the environment might include physical and social elements as well as access to services.
From these matrices a number of interventions can be explored, each roughly relating to a single cell.
The classical public health approach would then be to identify evidence-based interventions to address the underlying modifiable factor, determine the numbers of children and young people who could potentially benefit and then decide on the overall cost benefit of implementation. Obvious examples include the use of cycle helmets, reducing traffic speed in residential areas which have a high proportion of young children, separating pedestrians and cyclists from vehicles wherever possible, increasing access to safe play spaces and safe routes to school.
The current move of Public Health Departments into Local Authorities may have the potential to embed this type of approach into Local Authority business plans.
We think the current situation could be summarised with the words "less talking and more action". We believe that the responsibilities for investigating morbidity and mortality arising from the newborn period should rest within neonatal networks. Likewise children who die from well-established conditions such as cancer or cardiac problems should again be reviewed within their own networks as part of a rigorous quality improvement programme. Where these networks identify wider system based problems then they can refer to CDOP.
The number of deaths then coming to CDOP would be substantially reduced so then the number of CDOP's could be reduced and so they can develop greater expertise studying those deaths were lessons can be learned.
A national review should be undertaken to identify the most appropriate investigation models for different types of death in order to better identify modifiable factors or factors amenable to health service interventions. More reliable data from CDOPs should then be analysed on a national basis to better inform the introduction of prevention programs, identify emerging hazards such as blind cords and a national forum established to share best practice from the child death investigation process.
National strategies to reduce low birth weight, suicides, decrease morbidity and mortality from injuries and improve the recognition and management of sick children remain urgent national priorities.
The role of CDOPs then evolve to become a place where paediatrics and public health come together, with other agencies, to address the long-standing issue of excess mortality in the UK by implementing and evaluating locally based prevention programs.
Simon Lenton Consultant Paediatrician (retired) and BACAPH co-chair
Fiona Finlay, Consultant Paediatrician