The National Child Death Review Case Reporting System (NCDR-CRS) was developed in the USA to provide child death review teams with a simple method for capturing, analysing, and reporting on the full set of information shared at a child death or serious injury review. The NCDR-CRS is a web based system currently being used by 35 of the 50 US states. This article describes the purpose, features, limitations, and strengths of the system. It describes current and planned efforts for the dissemination of the data to inform and catalyse local, state, and national efforts to keep children safe, healthy, and alive.
Statistics from Altmetric.com
A comprehensive review of a child's death requires the sharing of case records from multiple sources on the wide ranging set of circumstances leading up to and causing a child's death. An effective review requires using this information to improve systems and prevent deaths. Capturing all of the information from review using reports from multiple sources and in a format useful for analysis and prevention is the purpose of the National Child Death Review Case Reporting System (NCDR-CRS). This is a passive epidemiologic surveillance system. It allows for the ‘ongoing systematic collection, analysis, and interpretation of data essential to the planning, implementation, and evaluation of public health practice closely integrated with the timely dissemination of these data to those who need to know’.1 Most importantly, the system can help to identify the aetiologic or causal factors in deaths of children so that communities can reduce or eliminate exposure to those factors as the basis for prevention.2
Development of the NCDR-CRS
When the National Center for Child Death Review (NCCDR), based at the Michigan Public Health Institute (MPHI), was funded in 2002 by the US governmenti, a major project objective was to explore the feasibility of building a standardised reporting tool for local and state child death review (CDR) teams. NCCDR found that 44 of 50 states had a case reporting tool for CDR; however, there was little consistency in the type of information that was being collected and analysed. Thirty CDR leaders from 19 states volunteered to design and test a case reporting system. NCCDR managed the system design and software development. It was originally proposed that the system would be a minimal dataset, capturing only the final outcomes of a case review. The 30 volunteer designers argued instead for a system that would capture the whole story of a child's death or serious injury, such that the version in use today contains over 1700 data elements.3
Thirty-five states are now enrolled in this web based system and have entered more than 84 000 reviewed child deaths. The database primarily reflects a period of review between 2005 and 2009. Table 1 provides a summary of the types of cases entered as of December 2010.
Purpose and objectives of the NCDR-CRS
The purpose of the system is to provide CDR teams with a simple method for capturing, analysing, and reporting on the full set of information shared at a child death or serious injury review, so that the information can be used at the local, state, and national levels to inform improvements in child safety and prevent deaths.
The objectives of the system are to:
Permit local and state CDR teams to systematically collect comprehensive information on every child death or serious injury reviewed including:
Child, family, supervisor, and perpetrator
Incident place, events, and emergency response
Risk and protective factors by cause of death
Further detail on acts of omission or commission contributing to the deaths, on sleep related infant deaths and on consumer product related deaths
Services needed, provided or referred
Recommendations for and actions taken to prevent deaths
Factors affecting the quality of the case review
Enable local and state CDR teams to easily analyse and report on their CDR findings
Enable child health and safety advocates to access aggregated state and national CDR findings to inform child health and safety prevention policies and practices.
NCDR-CRS is a web based reporting structure, built using MS-ASP.net. Data entered into the system is stored on secure servers at MPHI.
The system is child based, and can capture identifiable data on the child, but not identifiable for others involved in the death incident. Extensive data elements are included that address risk factors for most major causes of injury death.
Access to the system is allowed upon the signing of a data use agreement between a state and MPHI and confidentiality statements for all registered users in a state. Users log into and have access to the secure system via passcodes, depending upon one of three levels:
Level 1: individual team users can enter, edit, print, and delete cases and download identifiable data only for the cases reviewed by their team
Level 2: state level users can enter, edit, print, and delete cases and download identifiable data for cases reviewed by teams in the stateii.
Level 3: NCCDR staff can print and download de-identified data for all cases in the system by state
There is a paper form available that mimics the web system, but the web system was developed using a complex system of skip patterns to speed the data entry process. A data dictionary is available via paper and is also linked as a help feature to every data element in the web system.
Thirty-two standardised reports are available for downloading and/or printing at the local and state level. These reports are created using real time data. The reports cover all major causes of deaths and serious injuries. Local and state users are able to download local data at any time into their own software for further analysis. A data code book accompanies the system.
States are able to migrate case reports from archived CDR databases into the NCDR-CRS. A number of states have already done this. Some customisation is available at minimal costs for states. For example, users in Georgia are presented with an additional screen to help them track the state agencies involved in the case and recommended systems improvements.
The system is free to all users. The NCCDR staff enrols users and provides training and help desk support. MPHI programmers and IT staff maintain the system's functionality and servers.
Limitations of the data
There are a number of ways in which this system is unlike typical public health surveillance or vital statistics data. Most obvious is that the case reporting system does not usually include all child deaths occurring in specific jurisdiction and thus cannot be compared one to one with vital statistics data; rates cannot be calculated nor can the data be assumed to be a representative sample of all deaths without detailed analysis. Secondly, the data cannot be compared state to state, and sometimes even team to team within a state, because of variation among teams in the types and timing of death reviews. Third, there can be large differences in the quality of data between teams and states, especially for states new to the system. At first many users leave a large proportion of questions unanswered and data fields blank. We have found that this improves with time. CDR teams can use the form as a quality improvement tool. They find that not knowing the answer to an important question such as ‘were there working smoke detectors in a fire death’ has them gathering this information for their next fire death review.
Some teams also do not routinely access the data dictionary to ensure consistent meaning. NCCDR attempts to work with users to encourage compliance with the data dictionary, but is aware that some states have developed their ‘own’ definitions for a term. There are also a number of relatively subjective data elements, such as ‘was this death preventable?’ or ‘did an act of omission contribute to the death?’. These questions were intentionally included in the tool to encourage discussion, but may be problematic for certain types of analysis. Fourth, the original reporting source for specific data elements is not specified—so that it is not known which agencies contributed information, although the types of agencies participating at the review can be entered for each case. As such the system does not have a primacy rule for selecting the best answer to a question and instead relies on the CDR teams to determine primacy when there is dispute among agencies. The system cannot determine if the team or the person entering the data selected an answer.
Strengths of the data
Despite the limitations, the case information provided by local and state CDR teams provides valuable information on the complexities involved in many child deaths, and much of this information is not available from any other single source. For example, data entered on infant sleep related suffocations describe with whom, on what surface, and where the child was sleeping at the time of the death. This can be cross matched with detailed information on the child's supervisor to better understand the circumstances of these deaths. With pool drowning deaths, data record how the child entered the pool area, what barriers they may have breached, and why those barriers were not working. Box 1 describes the type of data that could be entered for a teen motor vehicle crash. For all deaths, comprehensive information on caregivers, supervisors, and perpetrators can help describe specific risks to children and improvements to help persons acquire resources to better protect their children.
Box 1 What the case reporting system can tell us about a teen motor vehicle death
Child's demographic information
Age; sex; education and employment; disabilities, health, substance abuse, mental health, delinquency, and child maltreatment and family violence histories.
Child's primary caregivers (up to two)
Age; sex; income; education and employment; primary language spoken; on active military duty; disabilities, health, substance abuse, mental health, delinquency, and child maltreatment and family violence histories; prior child deaths.
If needed and for person responsible for supervision: age; sex; income; education and employment; primary language spoken; on active duty in military; disabilities, health, substance abuse, mental health, delinquency, and child maltreatment and family violence histories; prior child deaths; specific impairments at time of supervision.
Time, place, emergency response, child's activity at time, number of other deaths.
Types of investigators, persons declaring cause of death, types of forensic tests conducted, reviews of child protective services records.
Manner and primary cause of death
Information on crash circumstances
Number and types of vehicles involved in crash, position of child, collision type, primary causes of crash, driving conditions, location of crash.
Information on drivers, occupants, pedestrians
For child, child's driver and other drivers involved in crash: licence status and violations to graduated licensing regulations; for all vehicles in crash: number of total occupants, teen occupants and teen deaths; protective measures—for example, seat belts needed, present, used, used incorrectly or not used.
Information on acts of omission or commission
Types of acts contributing to the death and information on the perpetrators of these acts (same as for supervisor).
Services used, needed, referred or recommended as a result of the death
Recommendations on actions to prevent other deaths
Includes a wide range of options—including education, environmental modifications, legislation, product safety; status of implementation of recommendations.
Information on the case review
Attendees, issues preventing a comprehensive review, summary of outcomes.
Dissemination of the data
Ideally, any review findings should be easily disseminated for use by government, organisations, and the public to keep children alive. However, the NCDR-CRS is first and foremost a system for use by local and state CDR teams and programmes. This is in keeping with the fact that CDR is best as a local process—people closest to the death event coming together to share the story of the death in order to keep other children safe from harm. In fact, according to the terms of the data use agreements with participating states, the data entered into the system is the property of these states. NCCDR only serves as the custodian of the data.
Most local teams are not accessing the data download feature, relying instead on the standardised reports. They are able to generate up to 32 of these, incorporate them into an annual report template, and thereby produce a report on their CDR findings and process to share with their community.
Most states participating in the system are downloading their data on an annual basis and generating extensive annual reports on all deaths reviewed or specialised reports on specific types of deaths such as suicides or drowningsiii. Most states have legislation requiring that reports on state CDR be presented annually to state agencies, legislators and/or governors. Some states are now linking their CDR data to their birth, death, and other records for more enhanced analysis.
Aggregated multi-state, de-identified data analysis generated by NCCDR staff is available to federal agencies and other researchers in accordance with the NCCDR data dissemination policy. Recently a number of agencies in the US government have shown interest in accessing the data to inform national policy. For example, a request has been made to generate data on the circumstances in child passenger deaths which may explain why caregivers fail to use child passenger seats. One federal agency is interested in comparing the number of child maltreatment deaths identified through this reporting system to the number generated in the federal child abuse reporting system. Mental health agencies are interested in access and compliance issues for prior and current mental health services in suicide deaths. A federal childcare licensing agency is interested in analysing unintentional deaths occurring in licensed day care centres. Federal child welfare has requested data on the quality of supervision in all injury deaths to understand the role of supervision and caregiver neglect in these deaths.
The US Centers for Disease Control and Prevention (CDC) are funding two projects to utilise the case reporting system as a means to better understand sudden unexplained infant deaths (SUID) and violent deaths. In the former, an expanded version of the case report tool that includes additional questions on SUID deaths is being piloted in seven states with support to ensure the review of 100% of all SUID deaths. Their data are being shared with the CDC as the pilot for a national SUID Case Registry. Data on violent deaths is being matched with data from states participating in the CDC's National Violent Death Reporting System.4 This probabilistic match will inform both the National Violent Death Reporting System and CDR as to the completeness of their violent death data and enrich understanding of these deaths. The US Maternal and Child Health Bureau is funding a secondary data analysis of infant sleep related deaths, using NCDR-CRS data from over 3000 SUID deaths in nine states, to understand the risk factors in these deaths.
A number of non-federal researchers have also made enquiries as to the availability of the data for research purposes. A formal application must be submitted and approved by the NCCDR Data Dissemination Committee for access to the de-identified database. Part of the application is agreement on the limitations of the data for surveillance purposes. The committee includes representatives from participating states and members of the NCCDR National Steering Committee. Data are not available from NCCDR that counts specific data elements by an individual state—for example, ‘100 of the 1000 deaths are from New York’. Requests for state identified data are rarely approved and if so must be approved by the participating states through a separate process.
Efforts will continue to enrol the remaining 16 states into the NCDR-CRS and to improve data quality. Especially important are: increasing the completeness of information, reducing inconsistencies in interpreting definitions, providing training and technical assistance for all users, and enhancements to the software to allow for customisation and automatic pre-population of data from agency case records. Most importantly, efforts will continue to assist child death review teams to interpret and use their data to prevent child deaths and to keep all children safe and healthy.
The author would like to thank and acknowledge the senior data analysts at the NCCDR, Heather Dykstra, MPA and Esther Shaw, MSIS, for their contributions to NCDR-CRS and review of this document; and to the 30 state CDR leaders who worked tirelessly to design the three versions of the case reporting tool.
Funding Maternal and Child Health Bureau (Title V, Social Security Act), Health Resources and Services Administration, Department of Health and Human Services (grant 5 U49MC00225-09-00) and the US Centers for Disease Control and Prevention.
Provenance and peer review Not commissioned; externally peer reviewed.
↵i The Center, including the development and management of the NCDR-CRS, is funded in large part by the Maternal and Child Health Bureau of the Health Resources and Services Administration of the US Department of Health and Human Services.
↵ii A few states have elected not to have access to case identifiers from local reviews.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.