Most unintentional injury deaths among young children result from inadequate supervision or failure by caregivers to protect the child from potential hazards. Determining whether inadequate supervision or failure to protect could be classified as child neglect is a component of child death review (CDR) in most states. However, establishing that an unintentional injury death was neglect related can be challenging as differing definitions, lack of standards regarding supervision, and changing norms make consensus difficult. The purpose of this study was to assess CDR team members' categorisation of the extent to which unintentional injury deaths were neglect related. CDR team members were surveyed and asked to classify 20 vignettes—presented in 10 pairs—that described the circumstances of unintentional injury deaths among children. Vignette pairs differed by an attribute that might affect classification, such as poverty or intent. Categories for classifying vignettes were: (1) caregiver not responsible/not neglect related; (2) some caregiver responsibility/somewhat neglect related; (3) caregiver responsible /definitely neglect related. CDR team members from five states (287) completed surveys. Respondents assigned the child's caregiver at least some responsibility for the death in 18 vignettes (90%). A majority of respondents classified the caregiver as definitely responsible for the child's death in eight vignettes (40%). This study documents attributes that influence CDR team members' decisions when assessing caregiver responsibility in unintentional injury deaths, including supervision, intent, failure to use safety devices, and a pattern of previous neglectful behaviour. The findings offer insight for incorporating injury prevention into CDR more effectively.
- Unintentional injury death
- child death review
- child neglect
- public health
Statistics from Altmetric.com
Child death review (CDR) is a process that involves a multidisciplinary team of professionals sharing information to better understand the circumstances surrounding the death of a child. Almost every state in the USA has a CDR program; some states have one state-level team, while other states have multiple CDR teams serving local or regional communities.1 2 CDR team members typically include representatives from law enforcement, medical examiner/coroners' offices, public health, medicine, social services, the courts, and emergency medical technicians or other first responders. At a review, team members share information from their agency about the child, the family, and events leading to the child's death. The goals of the review are to fully understand the circumstances of the child's death, more accurately classify cause and manner of death, and identify risk factors as well as prevention strategies.3
Most unintentional injury deaths among young children result from inadequate supervision or failure to protect the child from potential hazards in the home environment.4 Documenting these and other factors that contribute to a child's fatal unintentional injury, as well as identifying the potential contribution of child neglect to the death, is a component of CDR in most states. This is done to ensure accurate classification of cause and manner of death, as well as to improve agency response to specific deaths, and provide input for prevention. Nevertheless, reaching agreement on the role of caregiver responsibility and child neglect is often a challenge for CDR team members due to the influence of changing social norms and lack of standards of minimally adequate care or appropriate supervision that can objectively be applied to every situation.5
Several additional factors contribute to the challenge CDR team members face when examining the circumstances of a child's death and reaching consensus on adequacy of supervision and whether neglect was involved. Fundamentally, definitions of child neglect differ across states, disciplines, agencies, and purpose (eg, criminal or civil legal proceedings, public health surveillance, or research). There is also disagreement on conceptual aspects of neglect definitions, specifically whether they should be parent focused—parent or caregiver fails to act, resulting in harm to a child—or focused on a child's needs not being met, regardless of parental (in)action or other contributory factors such as cultural beliefs or financial resources.6 Furthermore, a number of attributes are often considered when determining child neglect. These attributes might be found explicitly in neglect definitions or implicitly in legal or agency interpretations, and include poverty, intent, child age, and chronicity (whether similar risk to the child has been documented in the past).
Although one of the strengths of CDR is the different perspectives the participating agency and professional members bring to the review, the different agency definitions of neglect, lack of consistent standards for determining adequate care and supervision, and changing social norms all likely contribute to the challenge of reaching consensus when it comes to agreeing on the extent to which a child's caregiver was responsible for the fatal unintentional injury and documenting whether the death was neglect related or not.7 These same issues have led to the suggestion that child neglect might better be classified along a continuum with options on one end ranging from a momentary lapse in parenting that results in harm to a child, to the other end—a parent who deliberately acts in a manner that puts their child at risk.6 Unfortunately, CDR team members are typically asked to determine whether inadequate supervision or neglect contributed to the death without the opportunity to incorporate uncertainty or contributing factors in their classification.
The purpose of this study was twofold. First, we wanted to assess CDR team members' categorisation of the extent of caregiver responsibility for a child's fatal unintentional injury based on vignettes that described the circumstances of the injury. We also sought to determine whether adequacy of supervision, social norms, and other attributes commonly considered when classifying child neglect would influence the CDR team members' categorisation of whether the unintentional injury death was neglect related or not.
We surveyed CDR team members to assess their classification of 20 vignettes that described the circumstances of a child's unintentional injury death. Study participants were asked to document the extent to which they believed the child's caregiver was responsible and the death was neglect related. The choices provided were: (1) caregiver not responsible/not neglect related; (2) some caregiver responsibility/somewhat neglect related; (3) caregiver responsible/definitely neglect related. Respondents were given the option to provide additional comments for each vignette.
The 20 vignettes were presented in 10 pairs. The first vignette in each pair (vignette a) included at least one attribute typically considered when categorising a child death as neglect related. These attributes included adequacy of supervision (vignettes 1, 5, 6, 8, 9), social norms regarding the use of safety devices (vignettes 2, 4, 10), and social norms regarding infant sleep environment (vignettes 3, 7). Poverty was included as an additional attribute in vignettes 7, 8, and 9. Changes in the second vignette of the pair (vignette b) were highlighted in bold and represented different or additional attributes that might influence the CDR team member's classification of caregiver responsibility/neglect. These included chronicity (vignettes 1, 7), adequacy of supervision (vignettes 2, 3, 10), poverty (added in vignettes 5 and 6, removed in vignettes 8 and 9), and child age for operating an all terrain vehicle (ATV) (vignette 4). Intent was included as an additional attribute in vignettes 6 and 7. By pairing the vignettes in this way, we sought to determine if the additional attribute or change in detail from the first vignette to the second resulted in a change in the CDR team members' classification of caregiver responsibility and whether the death was not, somewhat, or definitely neglect related. Each vignette is reproduced in full and the attributes of interest included in the first vignette and changed in the second are specified in table 1.
Although the vignettes were hypothetical, they were realistic, represented common causes of unintentional injury death among children, and contained information commonly available and discussed during CDR team meetings.
The data were collected using SurveyMonkey, an internet based survey tool (http://www.surveymonkey.com). In addition to the vignettes, respondents completed demographic information including their age, gender, professional affiliation, and details of their CDR experience.
The survey was available for completion between 1 September and 11 December 2009. CDR coordinators in 11 states were contacted and invited to participate. These 11 states were chosen to represent states with different programme attributes (eg, local CDR teams versus only a state level team). All CDR coordinators agreed to participate and were asked to forward an email to all CDR team members in their state. This email, from the Director of the National Center for Child Death Review, provided an explanation of the study, ensured participant confidentiality, and contained a link to the survey. Up to four reminder emails were sent to state coordinators; once a survey was completed by someone in their state, no additional reminders were sent.
Data were downloaded from SurveyMonkey and analyses were completed using SAS for Windows version 9.1. Fischer's exact test was used to determine if there was a significant difference in classification of neglect across each vignette pair. Respondents who did not respond to any vignette were dropped from the analysis. The study was exempted from review by the University of Missouri Health Sciences Institutional Review Board.
A total of 294 people from five states initiated the survey; seven did not respond to any of the vignettes, leaving 287 respondents who provided data for analysis. Although all 11 state coordinators agreed to participate and multiple reminders were sent, CDR team members from only five states completed the survey. Because the invitation to participate was forwarded from the state coordinator, and the number of people to whom the email was forwarded is not known, we are unable to calculate an accurate response rate. The 287 respondents represent the most common professions serving on CDR teams, and most reported serving on CDR teams in local jurisdictions for 3 years or more (table 2).
The respondents' assessments of caregiver responsibility for each of the paired vignettes are presented in table 3. For 18 (90%) vignettes, a majority of respondents assigned the caregiver at least some responsibility for the child's death. The exceptions to this were vignettes 4a and 10a. Vignette 4a described the death of a 15-year-old in an ATV crash with the goal of assessing the extent to which social norms regarding helmet use in an adolescent of legal age (15) to ride an ATV would influence classification of whether the death was neglect related. Comments provided by respondents indicated a hesitancy to assign responsibility to the parents given the child's developmental age and the sentiment that adolescents do what they want and parents cannot watch them all the time. Vignette 10a assessed social norms regarding the use of a smoke detector by describing the death of two children in a fire caused by faulty wiring in a rental residence without a smoke detector. Respondent comments for vignette 10a indicate that respondents thought the landlord was responsible for the faulty wiring and should have provided smoke detectors.
A majority of respondents classified the caregiver as definitely responsible for the child's death in eight (40%) of the vignettes—1a, 1b, 2b, 5a, 5b, 6a, 6b, and 7b. Vignette 1a assessed adequacy of supervision by describing an infant left without adult supervision in a bathtub and its pair, 1b, assessed chronicity by adding a prior Child Protective Services (CPS) substantiation. Vignette 2b assessed social norms regarding incorrect use of a car seat (from vignette 2a) with the addition of an impaired caregiver. Vignettes 5a and 5b assessed the adequacy of supervision in children who were left alone at night with the addition of poverty in 5b. Vignettes 6a and 6b assessed adequacy of supervision by describing an infant being left alone in a hot car, with the addition of poverty and intent in 6b. Vignette 7b assessed social norms regarding infant sleep environment and poverty with the additional attributes of chronicity and intent.
The change in distribution of responses across each vignette pair was statistically significant for eight of the 10 pairs, indicating that the change in attributes across the pair influenced the respondent's classification of the extent to which the caregiver was responsible and the death was neglect related. The two pairs without a statistically significant change in the distribution of responses were vignettes 5 and 9 (table 3).
Results from Fischer's exact test provide information on whether the distribution of responses for each pair differed significantly between vignette a and b. It does not, however, provide information on whether or how individual respondents changed responses across vignettes. Eight of the 10 vignette pairs (vignettes 1–7 and 10) were ordered so that one or two additional attributes were included in the second vignette (b). It was anticipated that if the attribute(s) added to the b vignette in these eight pairs influenced the respondent's categorisation of whether the death was neglect related, they would assign more caregiver responsibility in the b vignette. In vignettes 8 and 9, an attribute (poverty) included in vignette a was removed in b. The expectation here was that the assignment in caregiver responsibility would decrease in the b vignette if the removal of poverty influenced the respondent's categorisation. To assess the influence of the addition (or removal) of these attributes commonly considered when classifying child neglect, the proportion of respondents who changed their classification of caregiver responsibility across vignette pairs was evaluated (table 4). The direction of the change in assigned caregiver responsibility was as anticipated across all vignette pairs, except for vignette 9.
The three vignette pairs with the greatest proportion of respondents who changed classification across the pair were vignettes 2, 4, and 10, with 54.7%, 62.7%, and 54.8% changing classification across the vignette pair, respectively. Although 33% of respondents indicated the caregiver was definitely responsible for the child's death in vignette 2a, the addition of the caregiver's elevated blood alcohol concentration resulted in 152 more respondents classifying the death as definitely neglect related, a 160% increase. The difference in child age across response pairs in vignette 4 was more influential in assigning caregiver responsibility for this ATV related death than the fact that the 15-year-old was not wearing a helmet in 4a and the 12-year-old was wearing a helmet in 4b. In vignette 10, 46.7% of respondents assigned more caregiver responsibility in vignette 10b then 10a, indicating the addition of detail that the children's mother knowingly left her children with an inappropriate (disabled) caregiver was influential in the primary shift in categorisation of this death from not neglect related to somewhat neglect related.
Vignettes 1, 5, and 9 had the lowest proportion of respondents who changed their classification across pairs: 12.5%, 15.0%, and 11.2%, respectively. For vignettes 1 and 5, this was largely because most respondents (>60%) classified both deaths in each pair as definitely neglect related, although in vignette 1b, 29 (10%) additional respondents classifying this death as definitely neglect related were likely influenced by the mother's prior CPS substantiation. In vignette 5, respondent comments indicated they felt strongly that children should not be left unsupervised during the night, regardless of the reason. The difference in socioeconomic status described in vignette 9 (lake/pond drowning) did not influence many respondents to change their classification across the pair. The higher proportion of respondents that classified 9b as definitely neglect related was unanticipated. Respondent comments explained this finding by noting that the pond in the trailer park might have been out of sight, whereas parents would definitely have known about the lake at the resort (table 4).
Each vignette received multiple comments from respondents, ranging from 22 comments on vignette 10a to 68 on vignette 5a. These comments provide important insights into the respondent's decision processes. For example, chronicity (eg, evidence of ‘a pattern’ of neglect) and intent were frequently mentioned in the comments, and respondents identified them as highly influential attributes when considering whether a death was neglect related. In addition, the comments document wide variation in judgement and opinion among CDR team members and frequently note that the circumstances described in the vignettes were ‘tragic’, ‘unfortunate’ or ‘freak’ accidents. A sample of the comments is included in table 5.
In this exploratory, descriptive study, current members of CDR teams classified the extent to which a child's unintentional injury death was neglect related based on information provided in 20 vignettes describing common circumstances of unintentional injury deaths of children. A majority of CDR team members assigned at least some caregiver responsibility for 18 of the vignettes. In eight of the vignettes, over 60% of respondents classified the caregiver as responsible and the death definitely neglect related. Three of these vignettes were the first vignette in the pair (the ‘a’ vignette) indicating that the attribute assessed in these three initial vignettes, inadequacy of supervision, was sufficient to classify the death as neglect related. The remaining five vignettes where the majority of respondents classified the death as definitely neglect related assessed the addition of: (1) chronicity to adequacy of supervision (1b) and social norms regarding infant sleep environment (7b); (2) adequacy of supervision to social norms regarding use of safety devices (2b); (3) poverty to adequacy of supervision (5b, 6b); and (4) intent to adequacy of supervision (6b) and social norms regarding infant sleep environment (7b).
The 20 vignettes represented 10 vignette pairs that described similar circumstances, where the second (b) vignette included a change in attributes that might influence classification of neglect. A change in assessment of caregiver responsibility across vignettes reflects the potential for different attributes to influence the respondents' judgements. The change in assessment of caregiver responsibility was statistically significant for eight of the 10 vignette pairs. The two vignette pairs without a significant change in the assessment of caregiver responsibility both assessed the attributes of adequacy of supervision and poverty. Vignette 5b added poverty while vignette 9b removed the poverty attribute. Other vignette pairs where the poverty attribute was included or changed did have a significant shift in distribution across pairs (vignettes 6, 7, 8). It is not clear from these mixed results whether or how poverty might influence CDR team members when assessing caregiver responsibility and the role of neglect in a child's unintentional injury death.
The three vignettes with the highest proportion of respondents who changed their classification of caregiver responsibility from vignette a to vignette b assessed the addition of disabled or impaired caregiver (vignettes 2 and 10) or young age and legal mandate to social norms regarding the use of safety devices (vignette 4). The three vignettes with the lowest proportion of respondents who changed their classification (vignettes 1, 5, and 9) all initially assessed adequacy of supervision. These findings suggest that the presence of additional attributes may be an important influence when the CDR team is attempting to determine caregiver responsibility and the role of neglect in unintentional injury deaths that involve social norms related to use of safety devices. In contrast, when there is consensus that the death was neglect related in light of inadequate caregiver supervision, the presence of additional attributes may have little influence. Importantly, the addition of chronicity or intent always resulted in a statistically significant change in distribution of caregiver responsibility across vignette pairs and an increase in the proportion of respondents who classified the caregiver as responsible and the child's death as definitely neglect related. This was also supported by the respondent comments.
Survey respondents were given the option of including comments on each of the 20 vignettes. Because comments were not required of or provided by all respondents, we did not conduct a formal qualitative analysis. However, the numerous comments included for each vignette provide important insight into respondents' thoughts and decision processes, and demonstrate the wide range of opinions held by CDR team members. Moreover, the comments highlight the importance of chronicity and intent in assessing the role of neglect in unintentional injury deaths, clearly documenting that some team members are reticent to call a death neglect related without evidence of chronicity (eg, ‘a pattern of neglect’) or caregiver intent.
The distribution of the survey results, with the majority of responses indicating some but not full caregiver responsibility, and the wide disparity in opinion provided in the respondents' comments, document the challenge for CDR teams in reaching consensus when attempting to determine if a child's unintentional injury death was neglect related. This provides support for using a continuum to characterise better the role of child neglect in these deaths.
There are several potential limitations to this study. The survey respondents are a convenience sample of CDR team members, and the response rate is not known. However, it is important to note that the respondents represent the most common professions serving on CDR teams, a range of years of experience, rural and urban areas, and local and state teams. Furthermore, based on two of the authors' (PGS, TMC) over 20 years of combined experience serving on local and state CDR teams and providing technical assistance and consultation in CDR, the comments made by respondents represent the full range of opinions typically expressed during review and classification of caregiver responsibility and the role of neglect in unintentional child injury deaths, lending some degree of confidence in the results. The focus of the study on the role of neglect in unintentional injury deaths is another potential limitation, as the results do not provide information on classifying deaths related to other types of neglect (eg, medical, physical). Although examining other types of child neglect mortality is important, inadequate supervision is the most common type of neglect reported to CPS agencies8 and a contributing factor to most injury deaths among young children.4 9 To keep the survey completion time reasonable, only unintentional injury deaths were included.
In spite of these potential limitations, the results of this study document attributes that influence CDR team members and contribute to the challenge of reaching consensus when determining whether unintentional injury deaths are neglect related. Efforts to document child neglect might be viewed by some as piling on blame and punishment to a family who is already suffering with the loss of their child, and by others as an effort in futility. However, many risk factors for child neglect also increase child injury risk,10 and understanding the circumstances and risk factors for child injuries are important antecedents to developing and implementing effective prevention strategies.11 Therefore, the CDR process of discussing the circumstances of child injury deaths and identifying contributing factors has significant potential to facilitate development of prevention strategies that will effectively reduce all child injury deaths, regardless of whether neglect is ultimately identified as a contributing factor in the death.
Notably, most CDR programmes cite prevention of child deaths as a key programme function.1–3 The findings of this study offer valuable insights into the challenge of incorporating injury prevention into CDR more effectively. Specifically, the persistent references to ‘accident’, ‘tragic accident’, and ‘freak accident’ in respondents' comments belies decades of effort on the part of public health professionals to facilitate injury prevention with the message that ‘injuries are not accidents’. Clearly, there is still important work to do at a fundamental level in framing unintentional injuries as preventable. The findings also highlight the need for training CDR team members to enhance their ability to recommend and facilitate prevention efforts aimed at improving caregiver skills in protecting children from injury. Training should especially be focused in the areas of injury prevention related to improved supervision of young children, reducing substance use while caring for children, and the consistent and correct use of legally mandated safety devices. Moreover, guidance to teams on strategies for developing or adopting a continuum on which to classify whether a child's death was neglect related might reduce the challenge in reaching consensus. Use of such a continuum might also assist teams in determining appropriate agency responses, such as CPS and law enforcement. Given the depth of information on circumstances of child injury deaths, their systematic review of these deaths, their multidisciplinary nature, and their identification of prevention as a key programme function, CDR programmes are logical injury prevention partners. Working with CDR team members to ensure understanding of the tenets of injury prevention has the potential to further advance existing injury prevention efforts at the local, state and national levels and should be a public health priority.
What is already known on this subject
Identifying the potential contribution of abuse or neglect to a child's death is a component of child death review (CDR) in most US states. Determining if a death is neglect related is particularly challenging.
A number of attributes are considered when determining child neglect, including poverty, intent, child age, and chronicity—whether similar risk to the child has been documented in the past.
Different definitions of neglect, lack of standards for minimally adequate care and appropriate supervision, and changing social norms can lead to a lack of consensus among CDR team members.
What this study adds
CDR team members assigned some but not full caregiver responsibility for half of the unintentional child injury death scenarios reviewed, indicating use of a continuum or scale might help characterise the role of neglect in child injury deaths.
Team members often hesitate to call a death definitely neglect related without evidence of intent or ‘a pattern of neglect’.
Persistent references to ‘tragic’, ‘freak’, and ‘horrible’ accidents made by study respondents indicate there is still important work needed at a fundamental level to frame unintentional injuries as preventable.
The authors thank all the CDR team members who took the time to complete this survey, and the CDR coordinators who forwarded the study invitation to team members in their state: Kimberly Day (ME), Heidi Hilliard (MI), Maurine Hill (MO), Susan Rodriquez (TX) and Diane Pilkey (WA). Thanks also to Dr Vincent Palusci, Karen Amaranth, Heather Dykstra and Esther Shaw for helpful reviews of the survey instrument, and Ms Katharine Ball for SurveyMonkey programming.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
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.