Research, Practice, and Data Informed Investigations of Child and Youth Suicide: A Science to Service and Service to Science Approach
The National Institute of Mental Health (NIMH) conducted a two-day hybrid workshop to understand the challenges of investigating child and youth deaths. The purpose of the workshop was to discuss past, current, and future research on child and youth suicide, as well as the methods used to investigate and certify their deaths. The workshop looked at data collected by state and local groups such as Child Death Review, National Violent Death Reporting System (NVDRS), and suicide cluster investigation teams. By bringing this information together, workshop attendees hope to create a feedback loop that informs the work of everyone involved in efforts to determine, assess, and ultimately mitigate the rate of suicide death in children and youth.
Research, Practice, and Data Informed Investigations of Child and Youth Suicide – Day One
Research, Practice, and Data Informed Investigations of Child and Youth Suicide – Day Two
National Institute of Mental Health
Additional Event Information
Background and Purpose of the Workshop
Joshua A. Gordon, M.D., Ph.D., Director, NIMH
Shelli Avenevoli, Ph.D., Deputy Director, NIMH
The National Institute of Mental Health (NIMH) convened a two-day meeting to address the escalating crisis of child and youth suicide, particularly in response to a disturbing increase in suicide rates among Black youth and the impact of the COVID-19 pandemic on youth mental health. Past NIMH workshops on child and youth suicide underscored challenges involved in suicide death determinations in children. Therefore, the aim of the workshop was to stimulate a collaborative response between researchers, death investigators, and data collection teams to build a science-to-practice and practice-to-science approach toward understanding and ultimately preventing this tragic yet growing problem.
The workshop was designed by a co-planning team of members from NIMH, the National Center of Health Statistics, the National Center for Injury Prevention and Control, and the Health Resources and Services Administration's Maternal and Child Health Care Bureau, Workshop session summaries are available on this website, along with a list of resources for investigators at all levels within the science to practice approach.
Session 1: Overview of Child and Youth Suicide Statistics
and Challenges with Data Collection and Surveillance Activities
Margaret Warner, Ph.D., Director, Collaborating Office of Medical Examiners and Coroners, Centers for Disease Control and Prevention (CDC)
Dr. Margaret Warner provided an overview of national statistics on suicide deaths among children and youth. She reviewed the process of collecting data from death certificates, including the role of medical certifiers, and how this information feeds into national, state, and local databases such as the National Vital Statistics System (NVSS), the National Violent Death Reporting System (NVDRS), and the Child Death Review (CDR). She emphasized the important role of death investigations by medical examiners and coroners as they are medical certifiers who determine whether the manner of death is suicide.
Death certificates record demographic, geographic, and cause-of-death information. A funeral director typically enters demographic information on the death certificate. The cause of death section is reported by a medical certifier who is a physician in the case of natural deaths or medical examiner or coroner in the case of sudden and unexpected deaths, including suicide. The information is coded according to the International Classification of Disease (ICD-10).
Manner of death is a checkbox on the death certificate which includes accident, homicide, suicide, and undetermined intent, and is critical for the collection of information on suicides. During the death investigation, the medical examiner and coroner may review medical reports, mental health reports, legal history, physical examination and/or autopsy, toxicology reports, and other lab tests to determine cause and manner of death. The death certificates are filled out in a state's Electronic Death Registration System (EDRS), which then feeds into CDC's National Center for Health Statistics (NCHS). States also compile information from death certificates and will also feed that information into state and local public health and public safety data systems, such as CDR and NVDRS. States then report these data to CDC’s NVSS through a cooperative agreement.
Dr. Warner highlighted CDC’s efforts to collect reliable, quality data. The number of reported suicides among children under 10 years of age reported on death certificates is quite low. CDC does collect data on these deaths and they are included in the National Vital Statistics data. However, CDC does not typically report these numbers in annual statistical publications due to statistical reliability issues. Additionally, recent data linkage studies found racial and ethnic misclassifications accounting for 3 percent among Hispanic individuals, 3 percent among Asian/Pacific Islander individuals, and 34 percent among American Indian/Alaska Native individuals. These misclassifications emphasize the need to improve data collection at its entry points. CDC has been working closely with funeral directors to collect information more systematically. Additionally, CDC’s Collaborating Office of Medical Examiners and Coroners (COMEC) is a new program that provides resources to medical examiners and coroners towards improved data collection.
Dr. Warner summarized that funeral directors, medical examiners, and coroners are important for the reliable collection of suicide data, which in turn can be used to inform research and practices that ultimately prevent future suicides.
Session 2: Overview of Current Approaches to Child and
Youth Death Investigation When the Manner is Potentially Suicide
Session Introduction: Marcus Nashelsky, M.D., University of Iowa
Session Discussant: Jan Gorniak, D.O., MHSA Medical Examiner, Clark County Office of the Coroner, NV
Optimal Medicolegal Death Investigation Systems
Joyce deJong, D.O., Professor and Chair, Western Michigan University Homer Stryker M.D. School of Medicine
Dr. Joyce deJong provided an overview of death investigation systems and the characteristics required to ensure consistency, objectivity, and integrity. Medicolegal death investigators are tasked with determining the who, when, why, and how of a death, as well as using this information to complete the death certificate. This information is gathered specifically to classify deaths in order to better focus public health efforts.
The death investigation is a complex process that involves gathering data from multiple sources: the scene of the death, the circumstances of the death, the decedent's medical and mental health history, social and family history, autopsy reports (if performed), toxicology testing (if performed), and additional testing as needed. This information can sometimes be challenging to obtain because of medicolegal protections or because it involves information about a living family member. Despite this complex investigation process, death investigators often experience criticism about or challenges to their findings. Accreditation and certification programs and systematic quality control help ensure the quality of death investigations and improve processes.
Accreditation by the National Association of Medical Examiners (NAME) and the International Association of Coroners and Medical Examiners (IACME) promotes consistent, high-quality death investigations by setting practice and certification standards for medical examiner and coroner systems. Accredited systems must have safe and secure facilities, data security measures, rigorous staff training and certification, and continuous improvement efforts. These processes help maintain objectivity in operations and ensure the system is independent from undue influences from other entities. Just over one half the death investigation systems in the U.S. are accredited.
Beyond accreditation, NAME guides much of forensic pathology medical practice through its Forensic Autopsy Performance Standards. NAME AND IACME accreditation and adherence ot NAME standards meet the minimum standards of practice and do not cover all practices that might be in operation, such as the use of a CT scanner or 3D reconstruction. However, these guidelines comprehensively promote quality assurance in death investigations.
Dr. deJong added that one critical gap in medicolegal death investigation is the presence of mental health processionals in every system to help support families, especially in the context of a sudden or traumatic death such as a suicide. Support services offered by these mental health professionals could reduce the risk of suicide by family members.
Data Collection for Determination of Manner of Death
Lauri McGivern, M.P.H., F-ABMDI, Office of the Chief Medical Examiner, Burlington, VT
Ms. McGivern provided more detail about the work of medicolegal death investigators. Although death investigators work independently, they are part of a larger team of individuals gathering information guided by a common purpose—to help medical examiners and coroners determine cause and manner of death and to complete the death certificate. Their work begins with the collection of basic data at the time of death notification and at the scene of death. Common data points are collected on all cases and include basic demographics. This is followed by focused, detailed information centered on the events leading to death. Death investigators work closely with law enforcement, and interview witnesses, bystanders, and family who may be at the scene. They obtain medical records from primary care and mental health care providers and review recent hospitalizations, past medical history, and medications.
When a suicide is suspected, interviews with families experiencing the sudden loss of a loved one can be very difficult. Collecting the data at the time of the death is important and requires patient, empathy, and compassion. Interview questions may seek information about behavior changes, previous attempts, suicidal ideation, or preparations to end life. Other questions may focus on recent legal, relationship, work, or personal conflicts. A child or adolescent suicide may require additional information gathering on school problems, bullying, or social media conflicts.
Death investigators often confront challenges in the field, such as a lack of information at the scene, grieving families, difficulty obtaining medical and mental health records, and a lack of time due to high caseloads. Death investigators must often prioritize which data to collect, balancing the immediate need for basic data to determine cause of death with more specific data that may inform public health approaches.
Ms. McGivern emphasized the need for social workers trained in trauma support to help grieving families. Social workers can assist families by providing resources for grief counseling, information on how to obtain reports, and what to expect in the days and weeks ahead. They may also assist in gathering additional information or data that may not have been obtained in the initial interview to assist with informing public health. Public health nurses can provide support and education for families through individual and community outreach. Death investigation offices require additional funding opportunities and resources to obtain these services. Currently, there are limited funding opportunities for death investigation offices to obtain these services.
Challenges in Manner of Death Determination in Youth
Kathryn Pinneri, M.D., Director, Montgomery County Forensic Services, TX
Dr. Pinneri talked about the role of forensic pathologists in death investigations. Manner of death is based on the circumstances of a death and can be classified as natural, accident, suicide, homicide, or undetermined. A suicide death requires an intentional act, which cannot be determined by autopsy alone. For instance, a contact gunshot wound to the head requires death scene investigation to determine intention and manner of death. A comprehensive investigation to clarify the circumstances of the death is needed, especially in complex cases such as child or youth suicide.
Forensic pathologists must consider factors such as mental health history, medication use, or illicit drug use. Content from electronic devices is particularly useful in child and youth deaths because it may contain social media posts, journal entries, notes, communication with peers, evidence of bullying or blackmail, and passwords for other devices. However, it can be challenging to view content of electronic devices in some jurisdictions, and it is also important to quickly return the devices to the family because they can provide connection and insight about the deceased individual. There can be other signs that help indicate intent such as cut marks or scars, evidence of prior suicide attempts, or sudden changes in behaviors. Evidence of drug use at the time of death or plans for the future are not necessarily factors that preclude intent.
There are special circumstances in data collection. First, race and ethnicity can be incorrect on the death certificate. CDC’s efforts to train funeral directors to collect accurate data will ultimately help public health efforts address potential disparities. Additionally, sexual orientation and gender identity are not captured on the death certificate, and it can be challenging to mine this information from autopsy, death investigation, law enforcement, or case management reports. The information is often not collected routinely or uniformly, and it can be challenging to broach the subject with family members. However, it is critical information, considering data that suggest transgender individuals have a 40 percent lifetime prevalence of suicide attempts, with the highest risk among transgender adolescents and young adults.
Dr. Jan Gorniak invited the panelists to join a question-and-answer session.
Question: Are religion or sect captured in demographic statistics in the US and on the death certificate?
Answer: Dr. Warner said that religion is not captured on the death certificate.
Question: Can you address the age cutoff and the ability for children to understand the consequence of their actions? Why do we not believe a very young child who talks about suicide?
Answers: Dr. Gorniak said that very young children can understand the concept of death —children often use phrases such as my cell phone is going to die or my tablet died. The youngest age that she has concluded was suicide was eight.
Dr. deJong emphasized the importance of considering each death individually. There is often a lack of objectivity for child suicide, with some investigators excluding information about suicide in an effort to not hurt the family further. A related topic is whether the death certificate contains the manner classification. A few states offer the family an administrative version of the death certificate that omits the cause and manner of death; however, the complete certificate contains cause and manner and is used for statistical purposes. The core death certificate should be completed objectively without any effort to omit or mask information.
Dr. Pinneri added that it might be doing a disservice to families to exclude information about suicide. Mental health conditions may be familial; hiding that information might take away an opportunity for prevention and intervention for another family member. Family survivors of suicide are also at higher risk of suicide.
Question: What are the biggest challenges you face in making a suicide determination?
Answers: Dr. Pinneri answered that electronics and social media are the number one source of useful investigative information, but access can be difficult to obtain. Parents often do not know of all the sites their children access.
Dr. Gorniak agreed that electronics and social media are important sources of information and that some families were forthcoming with access and others were not. Sometimes a family can be very engaged and accepting but there is still not enough information to make a determination of suicide. Social media may lead you one way, but other information such as previously concealed photos might change that path of investigation. It is difficult to ask for that information, but it is important.
Question: Why is more attention not paid to family members who may be at increased risk of suicide?
Answers: Dr. deJong answered that there is a lack of resources for professionals who can respond to families.
Dr. Pinneri added that she had experience in an office that did have a social worker and it made a huge difference. In addition, having a chaplain or other spiritual counselor involved in the death notification can also be very helpful for the family.
Question: Why are public health nurses specifically used for infant and child deaths instead of social workers?
Answer: Ms. McGivern answered that they used public health nurses from their maternal and child health group specifically in infant deaths, which are often related to an unsafe sleep position or environment. It is an opportunity to provide education to families and resources such as a crib for the future. However, they use social workers from their law enforcement office for other violent deaths, which had been extremely helpful.
Question: What are the factors that perpetuate the disparities in accurately determining manner of death across certain racial and ethnic groups?
Answer: Dr. Pinneri answered that these data come from the funeral home and conversations with the families. The funeral homes might not be asking those in-depth questions or perhaps the family should be asked the question on a form rather than verbally. There are death investigation offices in jurisdictions with high populations of American Indian/Alaska Native individuals who may be a better resource for this question. Now that they understand how important the data is to collect, they will get better at collecting it. There are more boxes for race and ethnicity on the death certificate, but there are also complexities with self-identification.
Session 3: Overview of Research on Child and
Youth Suicide: Epidemiology and Prevention Considerations
Discussant: Jane Pearson, Ph.D., NIMH
Preschool Predictors of Early Adolescent STB
Joan Luby, M.D., Washington School of Medicine
Dr. Joan Luby reviewed her research on clinical depression and suicidality in preschool-aged children. Although eight years of age has been used as a cutoff for a suicide determination, there is evidence that suicidality can occur in children between the ages of four and seven. Signs of clinical depression can arise as early as age three. Dr. Luby and her research team followed 300 preschool aged children for nearly 20 years and found that young children have recurrent rates of depression, alterations in brain development, and biological markers that are similar to the characteristics of clinical depression in adults. Although this initial cohort showed very low rates of suicidality, a more recent study of preschool aged children found increasing rates of suicidal expression.
As a result, her research team aimed to identify what depression looks like in preschool aged children, why these signs are often missed in clinical settings, and how their suicidal thoughts were expressed. They found that the children experienced internalized feelings of anhedonia, guilt, and other negative thoughts. In a task in which children were to complete a story stem involving conflict, the children who expressed suicidality were more likely to resolve the conflict with violent themes or people dying or killing themselves. They also found that children who expressed suicidality showed a more mature understanding of death than their healthy peers or children with depression who did not express suicidality. Further, preschool aged children who expressed suicidality were more likely to later experience suicidality, suicidal thoughts or behaviors, or suicide attempt as an adolescent. The research team found similarly high odds among preschool aged children who were depressed. The odds remained higher even after the research team controlled for externalizing comorbidities and family history of suicidality. Importantly, similar to findings in mood disorder research, parent and child reports of suicidality in early childhood do not match very well; it is important to assess symptoms directly from the child.
Dr. Luby conceptualized suicidality as a feeling of intense, intolerable distress that leads people to use rash behaviors to solve the problem. Her team developed and conducted a clinical trial of a preventive intervention called the Parent-Child Interaction Therapy Emotion Development (PCIT-ED), which used the parent as their child’s external emotion coach and regulator. The treatment did not directly target suicidality, but rather improved the children’s ability to accurately label their own emotions, identify emotions in others, and regulate their intense emotions. Importantly, the parent also learned to tolerate their child’s expression of intense emotion. The treatment was effective in treating depression and also reduced the likelihood of suicidality in preadolescence.
Dr. Luby suggested that the increasing rates of suicidality in very young children may be influenced by increasing rates of intolerance, bullying, violence, and exposure to lethal means. She stressed that clinicians should be more vigilant about early onset mood and depressive symptoms as a marker for later suicidality. Further, their study showed that the emotion development intervention was a promising treatment for young children.
Youth Suicidal Behavior: Trends, Facts, and Figures
Arielle Sheftall, Ph.D., University of Rochester Medical Center
Dr. Arielle Sheftall reviewed trends in child and youth suicide and non-suicidal self-harm behaviors over time, shifts in lethal mean use, and risk factors. In 2021, suicide was the 3rd leading cause of death among children and youth aged 10 to 19, accounting for more than 2,900 deaths. Although suicide among both male and female children has increased, the increase has been sharply higher among female children—rising by 129 percent between 2000 and 2021. The prevalence of self-harm has also increased among children aged 10 to 19, with emergency department visits for self-harm surging by 181 percent since 2001. Additionally, suicide has emerged as a prominent cause of death among young children aged 5 to 12, rising 204 percent between 1990 and 2021. Emergency department visits for self-harm among this age group have increased by an alarming 411 percent over the past 20 years.
Mechanisms of suicide have also evolved, with differences across gender. Among female children aged 5 to 19, the leading cause of suicide continues to be hanging suffocation, however suicides by firearm among female children have increased by 80 percent. Among male children in the same age range, the leading cause of suicide was firearms, which has increased by 59 percent, followed by hanging suffocation.
Risk factors for suicide and self-harm include factors that can be changed, such as emotion regulation and environment, and factors that cannot be changed, such as family history, genetics, age, sex, gender identity, race, and ethnicity. Known risk factors for suicide and self-harm include being male, LGBTQ+ sexual orientation and/or gender identity, access to lethal means, and a history of sexual or physical abuse. Social media has been associated with increased suicidal behaviors but could also be used as a platform for positive interventions. In younger children, risk factors for suicide and self-harm include impulsivity, limited cognitive skills, and difficulties in emotional regulation, as cited in the literature. These factors can make it challenging for children to develop coping mechanisms that may affect their immediate distress, envision a future beyond their current circumstances, and effectively regulate their emotions.
Challenges in Ascertaining Suicide Risk in Child and Youth Suicide
Rhonda Boyd, Ph.D., Children’s Hospital of Philadelphia
Dr. Rhonda Boyd talked about challenges involved in ascertaining suicide risk among children and youth, including discrepancies in reporting, stigma, limited information about young children, and racial misclassification. How a question of suicide risk is asked, when it is asked, and by whom is it asked can significantly influence a child’s answers. Consequently, there can be large discrepancies between the child, their parents, and their teachers.
The discrepancy in reporting suicidal thoughts has been well-documented in the literature. For example, Dr. Boyd and her research team used data from a large database of youth aged 11 to 17 to evaluate the extent of this discrepancy. They found that 50 percent of parents were unaware that their children had suicidal thoughts in their lifetime and 75 percent were unaware that their child had thoughts about death or dying. These discrepancies were more prevalent among racially minoritized populations. Dr. Boyd added that a different study found that racially minoritized youth were more likely to omit answers to questions about suicide attempts than white youth.
Stigma is one factor involved in failing to disclose suicidal thoughts. A survey study found that perceived stigma from family members was the most common type of stigma related to disclosure. Fear about disclosing related factors such as relationships, sexuality, gender identity, trauma, and bullying could also cause a child to withhold information from their families. Some social beliefs and media depictions further stigmatize suicidal thoughts and behaviors. For example, some may believe that asking a young child about suicide will distress the youth or put the idea in their head and increase risk, despite the evidence that this is inaccurate. For example, in one study, the majority of all youth aged 10 to 21 surveyed in emergency departments about thoughts regarding universal screening in EDs were in support of it. Further, a meta-analysis on the impact of suicide-related content found that asking about it has no negative effect on level of distress or functioning.
Dr. Boyd added that there was a lack of suicide data on children under the age of 10—states often do not collect these data and it can be difficult to determine intent among young children, especially when the presence of a suicide note is required. Compared to teens (of whom 30 percent leave a note), children are less likely to do so. Additionally, the growing rate of multi-racial and multi-ethnic children may increase racial misclassification in suicide mortality data.
Intervention Targets for Bereaved Families, Schools, and Communities
Stephen O’Connor, Ph.D., Chief, Suicide Prevention Research Program, NIMH
Dr. Stephen O’Connor addressed what is known about suicide bereavement and its outcomes and potentially helpful interventions, noting that these topics need more research. He also discussed that what is known is not broadly incorporated in practical prevention strategies. In terms of scope of the problem, he cited the work of Dr. Julie Cerel, who estimated that for every person who dies by suicide, 135 people are exposed to the loss. Exposure to suicide increases the risk for suicide. Studies have found that parental loss of a child to suicide is a risk factor for suicide death, with stronger effects in deaths that occur in early childhood. Although this risk is high for both parents in the first month after a child’s death, it remains markedly high for women across the first year. Additionally, 1 in 5 youth are exposed to a peer death by suicide, which has been shown to increase the risk of suicide ideation and attempt.
There is a continuum of effects from suicide exposure, ranging from larger, more loosely connected groups of individuals to smaller, more intimately linked groups of individuals who suffer prolonged bereavement. However, because the risk of future attempt in youth is not attenuated by the closeness of someone to the person who died, this suggest that universal prevention strategies in schools and communities would be beneficial. At the same time, those experiencing prolonged bereavement may require a tiered approach to address distress separate from suicide risk.
Suicide “postvention” strategies could therefore be organized across a socio-ecological model. At the individual level of the socio-ecological model, prevention strategies could target meaning-making, grief, depression, guilt, shame, and perceived avoidance. At the school level, strategies could target factors such as social connectedness, social affiliation, belongingness, and a supportive learning environment. At the community level, strategies could address economic supports, protective environments, coping and problem-solving skills, and safe reporting about suicide.
Prevention interventions could also be organized across the universal-selective-indicated model. Universal strategies affect populations and may focus on healthy and empowered individuals through organizational and systems-based approaches. Selective strategies target at-risk individuals and may focus on community and caregiving systems for effectively responding to a scene of a suicide. Indicated strategies target individuals who are currently experiencing negative outcomes and may focus on accessible, effective, and coordinated treatment and support services.
Dr. O’Connor highlighted NIMH funding opportunities to develop novel therapeutics or to adapt existing therapeutics for loss survivors.
Dr. Jane Pearson invited the panelists to join a question-and-answer session.
Question: What would be helpful for capturing multiple races in suicide data?
Answers: Dr. Boyd answered that the current NIH categories are not sufficient for capturing multiple races and there are ongoing discussions to correct this. It is an important topic because people should not be forced into boxes they do not want to be in.
Dr. Sheftall added that one of the requests they have had is to look at multiracial youth, who are one of the largest growing populations in the U.S. Data on multiple race only goes back to 2018 and there is not enough data to understand the trends in suicidal behaviors and risk factors, which might look different than other populations.
Question: There have been multiple age groups presented for children and youth. What is the distinction between child and youth? Are there standard age groups or categories?
Answer: Dr. Pearson answered that defining age groups is a common challenge that NIMH faces when writing funding announcements. There might be an announcement for preteen children, and the age range needs to be defined for the paricular study purposes. The categories might be defined by school grade levels for one project, but a project on systems of care may want to expand the definition to young adults to capture pediatric care. It is a challenge and the only solution currently is to ask people to define their age groups. There may be implications for guidelines, such as cognitive status in very young children. This variability also emphasizes why an age cutoff is not a good policy.
Question: Are there certain screening tools that seem to be more accurate for children? How might this vary across settings, such as schools versus clinical practice?
Answer: Dr. Boyd answered that her team at Children’s Hospital of Philadelphia uses the PHQ-9. The American Academy of Pediatrics recommends depression screening for children aged 12 and older. Screening is more challenging for young children. The PHQ-9 is generally acceptable across a lot of different environments and it has a suicide item, but there is some controversy in how well it identifies risk, particularly outside the context of depression. Her team also uses the Columbia Suicide Severity Rating Scale in their outpatient clinic. Their pediatric emergency department uses the ASQ.
Dr. Luby added that CDI is a screener for depression that is appropriate for children as young as age 7. Her team also developed the Preschool Feelings Checklist, which is a brief, validated tool that includes a suicidality item and can be used for children as young as age 3.
Question: In addition to story stems and puppets, could books or videos also be helpful in a therapeutic setting to understand a child’s thinking?
Answer: Dr. Luby answered that books and videos may be good intervention tools, but not necessarily pragmatic screening tools. The screening is better done with a caregiver.
Question: Is there any research or data on the role of siblings in terms of services or treatment?
Answers: Dr. Luby answered that there was not a lot of research on siblings. Some of the parent-child interventions probably radiate to the siblings because the parenting style changes.
Dr. Sheftall said that her team looked at the narratives of youth who had died by suicide and they noticed that there had often been an argument with a sibling prior to their death. It would be important to consider how to incorporate siblings into interventions and prevention programs so that they were equipped alongside parents.
Dr. Gorniak talked about having to call children’s services for a suicide of a young person, which was a good resource for understanding a sibling’s perspective because they help the surviving children. People from children’s services were also helpful to have on child death reviews to provide that broader perspective.
Dr. Wilcox commented on her experience in conducting psychological autopsy interviews with parents and siblings of youth who died by suicide. Although they are there to gather information about short-term risk factors, they find that the families derive great benefit just from being able to talk through their experiences. She had received feedback from one interview that it had helped the child’s sibling in the grieving process.
Question: Are questions about adverse childhood experiences (ACEs) a way to ask about suicidality or should the questions be more targeted?
Answers: Dr. Luby said that ACEs are a robust risk factor for various negative outcomes, but probably are not useful as a specific risk factor for suicide in itself.
Dr. Boyd said that ACEs include experiences that are common in ethnically and racially diverse youth, such as community violence, racism, discrimination, or incarceration of a parent. Bullying does not necessarily capture those experiences. It would be important to understand these experiences as context and they might not come out in interviews with families.
Question: Can some of these protocols or screening tools be adapted to a postmortem setting? How concise can those questions become while still getting accurate information?
Answer: Dr. Luby said that this was a great idea that should be explored. There might be a certain point in the postmortem setting or the grieving process in which a caregiver would be able to give an effective answer. It could be important information to understand.
Question: With the need for more resources to fund services such as social workers working in death investigation offices, could services such as telehealth be leveraged to connect to families?
Answer: Dr. O’Connor answered that telehealth could be very beneficial. It is known that there is often a window of openness when interacting with people who have just survived a serious suicide attempt, which tends to close over time. Similarly, there could be a window of openness in someone who just experienced a loss and an opportunity to provide near-term support. However, it would be important to understand the guardrails needed for the services to be effective. It would not be good to open people up but not deliver services. When 988 was expanded, it was very important to have a sufficient workforce that could take on the increased utilization. Adequate resourcing is an important issue because it is not helpful to have recommendations if the resources were not available to make things happen.
Session 4: Overview of Coordinated (Post Suicide Determination) Data Collection Approaches
Discussant: Karin Mack, Ph.D., CDC
Child Death Review: A Comprehensive View of Youth Suicide
Abby Collier, M.S., Program Manager, National Center on Fatality Review and Prevention
Ms. Abby Collier reviewed the functions of the National Center on Fatality Review and Prevention, which provides programmatic and data technical assistance for Child Death Reviews across the U.S., supporting more than 1,500 local teams. Child Death Reviews are conducted both locally and statewide, and the structure, funding, resources, and legislation supporting these reviews vary widely between programs. The Center’s National Fatality Review Case Reporting System is a free web-based tool that collects a wide range of data from multiple sources and is used in 47 states. The case reporting system currently contains more than 2,600 variables and states can use the tool to access, interpret, and visualize their data. One of their newer sections is based on life stressors, which provides valuable context about variables such as relationships and home environments.
Collecting this comprehensive data when manner of death is suicide often takes more time. In the spirit of informing child death prevention programs, a feature was added to allow the CDR data collection team to indicate a suspicion that the child intended to harm themself, even if the death was not ruled a suicide on the death certificate. This feature allows them to collect additional, suicide specific risk factors that otherwise would not have been captured.
During the COVID-19 pandemic, the Center quickly adapted its program to incorporate pandemic-specific questions to identify direct and indirect relationships with COVID-19. Between January 2020 and March 2023, 29,711 deaths were entered into the case reporting system. Of the 42 percent of deaths that were indirectly related to COVID-19, some included cases in which children were not able to obtain timely mental health or substance use treatment. The Center also observed that a subset of children whose suicide death was considered indirectly impacted by COVID-19 had experienced significantly higher numbers of life stressors. Across all child suicide deaths during this time period, children had experienced an average of 3.2 life stressors. However, children who died by suicides deemed indirectly impacted by COVID-19 had an average of 4.5 life stressors. These stressors were commonly related to transitions, criminal charges, mental health needs, and virtual schooling.
More than a year ago, the Center identified five recommendations that could improve the review of child suicide deaths. Ms. Collier highlighted two: the need for consistent access to records and standards for death scene investigation related to suicide. She underscored the complex nature of the work, emphasizing that the infrequency of child suicide death could present further challenges without consistent access to data and standardized processes.
National Violent Death Reporting System Overview
Janet Blair, Ph.D., M.P.H., Team Lead, Mortality Surveillance Branch, CDC
Dr. Janet Blair provided an overview of the National Violent Death Reporting System (NVDRS), which was established in response to a 1999 Institute of Medicine Report. The system began collecting data in 2003 and gradually expanded to all 50 states; Washington, DC; and Puerto Rico by 2018. NVDRS is an online surveillance system that monitors violent deaths—linking data from death certificates, coroner or medical examiner reports, and law enforcement reports. Violent deaths include suicide, homicide, deaths of undetermined intent, legal intervention deaths (excluding execution), and unintentional firearm deaths. No personally identifying information is collected and trained abstractors enter the data according to CDC guidance.
NVDRS data provides comprehensive detail about the incident (e.g., injury characteristics), decedent information (e.g., demographics), toxicology, and the circumstances leading up to the death (e.g., history of suicide attempt or thoughts, school issues, suicide of a peer, or exposure to disaster or traumatic event). More than 600 data elements are collected in the system and two narratives are written by the abstractor—one based on the law enforcement report and the other based on the coroner/medical examiner report. Dr. Blair reviewed some of the characteristics of child suicide from 2020 data, which showed that family problems, school problems, arguments and conflicts, a history of suicide attempts, and intimate partner problems were common among children aged 10 to 17 who died by suicide.
NVDRS data can help identify emerging issues (e.g., increasing use of sodium nitrate), reveal important variations and patterns, examine risk factors, and inform and evaluate prevention efforts. States such as California, Nevada, and New Jersey are using NVDRS data to support suicide prevention initiatives. NVDRS also collaborates with various partners, such as the International Association of Chiefs of Police, the National Sheriff’s Association, and the American Public Health Association, to enhance the quality and accessibility of data. NVDRS provides restricted access to data for researchers and supports the Web-based Injury Statistics Query and Reporting System (WISQARS). CDC also publishes an annual surveillance summary on NVDRS data.
Suicide Cluster Investigations: The Role of Mortality and Morbidity Data
Steven Sumner, M.D., M.Sc., Senior Advisor for Data Science & Innovation, CDC
Dr. Steven Sumner presented an overview of CDC’s process for conducting youth suicide cluster investigations using a case study that occurred in Stark County, Ohio in 2017. CDC does not have access to real-time death data and must rely on local communities to raise concerns about unusual patterns in their area. Stark County had experienced 12 suicides by middle and high school students over approximately eight months, which resulted in intense concern among the local community, school officials, and media. The Ohio Department of Health formally requested CDC support for an investigation and recommendations.
CDC first confirmed that there was a suicide cluster. They used the Anderson-Darling test on data provided by Stark County to confirm that the number of suicides were occurring more frequently than would be expected by chance. CDC also looked at emergency department suicide-related visit data to determine whether a suicide cluster could be detected in absence of real-time suicide death data. Although Stark County experienced an increase in youth emergency department visits over this time period, it was not the county in Ohio that experienced the highest increase—emphasizing the complexity and unpredictability of suicide clusters.
The next phase of CDC’s investigation considered the causes and risk factors associated with the suicide cluster. Students in Stark County were highly connected through friendship networks. Therefore, CDC deployed a secure survey to more than 15,000 students in more than 30 schools to assess the spread of suicide and suicide cluster-related risk factors and to provide linkage to care. Notably, they found that both viewing and posting about social media material related to the suicide cluster was associated with increased risk. Having a close friend die by suicide was another significant risk factor. A large proportion of students also expressed suicide-related thoughts, demonstrating the need for broad prevention strategies.
The last phase of CDC’s investigation was to monitor for resolution. Because the local community often learns about a suicide death before public health officials do, proxy data such as real-time emergency department data may help public health agencies monitor and proactively respond. Dr. Sumner reiterated the importance of supplemental data to identify opportunities for prevention or improved use of suicide death data.
Dr. Karin Mack invited the panelists to join a question-and-answer session.
Question: What are the challenges and needs for conducting a suicide death investigation for children and youth who live in rural areas?
Answers: Ms. Collier said that the National Center on Fatality Review and Prevention can analyze data by location type. They are in the process of releasing a brief on health equity considerations, for which geography is a significant component.
Dr. Sumner added that it can be particularly difficult to determine whether a suicide cluster is occurring in a rural area and CDC might need to triangulate information from a variety of data sources, such as emergency department data, to make a determination.
Dr. Blair said that youth suicide in rural areas can be evaluated using NVDRS data. For example, a study was conducted using NVDRS data that found that suicide deaths in rural areas were associated with less mental health care utilization and greater use of firearms as means.
Question: How do you categorize data that is not necessarily yes/no, such as when a risk factor is considered undetermined and it is unknown why it is considered undetermined?
Answers: Ms. Collier answered that it was a common concern and that they try to differentiate “unknown” as an answer from a blank answer. It relies on the user to read the data dictionary, which might not be happening. They have been analyzing the barriers to data collection and what might be needed to obtain more concrete answers.
Dr. Blair said that it was a struggle with NVDRS data as well. For instance, mental health treatment is likely undercounted because of the difficulties associated with obtaining medical records. Gender identity and sexual orientation data are often missing. NVDRS data could be supplemented, but that could require additional time and memorandums of understanding. Although there is a wait for NVDRS data, it is a rich source of information and they continually add fields when they see trends, such as suicide deaths in cemeteries. Dr. Blair added that CDC offers self-care to their abstractors because of the disturbing nature of the information they are often exposed to.
Question: Are you able to capture information about medical conditions related to sexual health, such as sexually transmitted infection?
Answers: Ms. Collier said that the questions for that information are in the case reporting system and so it could be added if the information was available.
Dr. Blair said that NVDRS does have the ability to capture mental health diagnoses and there is standardized guidance about how to enter it. For instance, if an antidepressant is identified in a toxicology report, it does not mean that depression should be checked off as a condition. Other conditions, such as HIV infection or chronic pain, may be mentioned in the narrative if the conditions were considered a circumstance. If a key word shows up often, they may decide to add it as a new variable.
Dr. Pinneri emphasized the importance of asking for specific data. Death investigators receive packets to fill out before participating in a Child Death Review to help guide the discussion. It can be time consuming to complete the packet, but it helps move the process forward. If there was a question that needed to be answered, it could be embedded in the packet and eventually they will get answers. It is easier to obtain this information in real-time than by going through the narratives a year later.
Dr. deJong echoed the importance of embedding questions or processes for gathering needed data. There is a shortage of forensic pathologists and they are not always able to search very deeply in the records. Including questions in packets or having abstractors in the larger death investigation offices would be helpful mechanisms for obtaining those data.
Ms. Collier said that it is challenging to meet the competing needs of researchers and funders with what is available to collect. Their office has an extensive wish list of detailed variables they would like to collect but fatality review teams often do not have access to that level of detail.
Dr. Pinneri added that it is important to manage expectations—death investigative teams are more about the how than the why.
Question: Is there any developmental variation in suicide clusters? Adolescents are often part of very tight social groups, but has there been any study on younger or older ages?
Answers: Dr. Sumner answered that it is a very important topic that requires more research, particularly in the U.S., where it is unclear how many suicide clusters occur. It would be important in terms of identifying what contributes to the rising rates of suicide among younger children, for example.
Dr. Mack added that they often receive requests for technical assistance for suicide clusters across all ages.
Question: What is your most important partnership? What were the important lessons learned from the pandemic about data capture and sharing?
Answers: Dr. deJong answered that virtual meetings increased the ability for medical examiners to attend multiple death reviews, which was a significant improvement.
Ms. Angie Hayes agreed that virtual meetings increased the number of meetings one was able to attend, but it did not necessarily translate to improvement. It can be harder to manage a meeting virtually than in person. Additionally, she manages the death review team in Harris County, which is very large and quite behind in cases. It might be more productive to have virtual meetings in smaller counties.
Dr. Warner said that CDC’s WONDER system is now collecting a lot more provisional data as a result of the pandemic. There is now only a six-month delay in suicide data, which is much improved from five years prior. The resources made available to address the drug overdose crisis also impacted the ability to obtain more timely data, particularly to the Offices of the Medical Examiners and Coroners. As a result, the reporting of suicide deaths has become more timely.
Question: What are the differences or similarities between US efforts in suicide data surveillance and any other international systems? Are there lessons that can be learned?
Answers: Ms. Collier answered that her office participates in an international group of child death review programs and the most robust programs are in the UK and throughout Europe. The primary difference is data collection. Not only do the UK and Europe have broader access to records, but they also engage the family in the review process. This can provide a perspective that might not otherwise be captured.
Dr. Blair said that her office had enquiries from Canada and Australia about the implementation of NVDRS, and they provided technical assistance to those public health officials.
Question: In terms of suicide clusters, has there been any effort to monitor what is happening in student Chromebooks?
Answer: Dr. Sumner answered that no particular product is completely universal, but that they do use new products and systems in schools. For example, one school had a safety monitoring software program and CDC worked with it to understand what was happening in students’ lives as measured through the objective information the software collected. The software collected information about what the students are being exposed to, such as bullying, suicide-related content, or violent content. It provided a very useful way to understand risk factors in a more granular time frame.
Question: Once the data is collected in the data systems, what kind of interoperability is there with other systems? Can a child be cross-referenced across multiple systems? Or could there be a geocoding function to better understand violence in the community or the impact of public policies?
Answers: Ms. Collier said that it was not possible in the National Center on Fatality Review and Prevention because their data is at a national level and does not have access to identifiers. They are available at the local and state levels, but there is often a lack of resources to conduct that kind of research. She invited researchers to reach out of there was a state they wanted to connect to.
Dr. Blair said that there are no linkages in NVDRS data either because they do not collect personally identifiable information. However, NVDRS data can be linked with other interesting data sources such as juvenile justice.
Dr. Deb Stone added that CDC was collaborating with the Department of Defense (DoD) to link NCDRS data with their Suicide Event Report (DoDSER). They are also conducting another project in Washington State to link mortality data to morbidity data from social services, justice departments, and emergency departments. CDC also uses syndromic surveillance data, which is not a data linkage effort per se, but provides useful healthcare information about suicide morbidity, attempts, and ideation across a range of populations.
Question: Besides additional staff, what other resources are needed to help this work?
Answers: Dr. Pinneri said that the delay in data gathering is significant. A six-month lag would be better than an end-of-the year reporting so that requests for data can happen throughout the year. Six months would help ensure that data collected on cases are completed. Currently, they must wait until the end of the year to start data abstracting.
Dr. Blair talked about their data modernization efforts including a rapid reporting system that allows abstractors to access certain characteristics before the year is closed out. The caveat is that access to certain reports or complete reports is limited, but it does make it easier for states to report earlier. Additionally, NCDRS has an interoperability feature in which an abstractor can import information from the death certificate. It would be helpful to have standardized systems with harmonized data elements that could also be uploaded into NCDRS.
Ms. Collier said that their greatest need is data modernization across Child Death Reviews. The number of staff needed to modernize is a barrier—there are currently around 2,300 users and there are differences in how people answer questions. If there was one reliable source, it would reduce the noise in their data. Additionally, data linkages would expose their teams to less intensive work for data gathering because they would not have to read lengthy reports. The impact of data modernization on the workforce would be significant.
Session 5: Special Populations
Discussant: Stacia Friedman-Hill, NIMH
LGBTQ+ Youth Suicide: Addressing the Data Chasm
John Blosnich, Ph.D., M.P.H., University of Southern California
Dr. John Blosnich discussed the significant gaps in understanding suicide among LGBTQ+ individuals. Although data are available on suicide ideation and attempt, there is a lack of data on suicide mortality—limiting the ability to impact policy, funding, and awareness. The primary obstacle is that the U.S. does not systematically collect data on sexual orientation and gender identity at the time of death. There is not a section to check on the death certificate and it is not part of the standard investigation process.
Despite the evidence gaps, there are some studies that signal a greater risk of suicide death among the LGBTQ+ population. These studies used four types of data: survey data paired with mortality data, national registry data, medical record data, and NVDRS. Survey data paired with mortality data showed inconsistent findings, with one survey identifying greater risk among women who have sex with women and another study not identifying any difference. National registry data studies have been conducted in Denmark and Sweden and have also identified a risk signal, but but differ in the magnitude of difference between LGBTQ+ and non-LGBTQ+ individuals. Medical record data from VHA is limited in its classification of transgender individuals. However, these data did show that suicide was the 4th leading cause of death among transgender VHA veterans as compared to the 10th leading cause of death among cisgender VHA veterans.
Dr. Blosnich and his research team used natural language processing on clinical progress notes to identify LGBTQ+ veterans. They found that suicide was the fifth leading cause of death for this population. NVDRS introduced two fields for sexual orientation and gender identity in 2013, although these data are severely hampered by missingness. One common limitation across all of these data sources is often a significant delay between collecting data and analyzing or publishing findings that may inform prevention efforts—it can take as long as decades to get the information. Changes in identity over time and missing data are also ongoing challenges.
Dr. Blosnich talked about one solution that was underway. They found that most death investigators did not ask about sexual orientation or gender identity, nor had they received training on how to collect those specific data. Therefore, Dr. Blosnich and his team collaborated with advocates and other stakeholders to develop a training protocol for death investigators. Part of the training included role play and case reviews because the team recognized that death investigators conduct very sensitive conversations with bereaved individuals and that obtaining information on sexual orientation and gender identity was not a standardized process. The training program was currently being piloted in Los Angeles County, California, and Utah. Dr. Blosnich emphasized the importance of collecting sexual orientation and gender identity for all individuals and the need for leadership buy-in to help scale this data collection.
Tales from the Field: Collecting Sexual Orientation, Gender Identity, and Other Information
Michael Staley, Ph.D., Utah Office of the Medical Examiner
Dr. Michael Staley described Utah’s process for collecting sexual orientation and gender identity in their death investigations. Utah has a centralized medical examiner program, but its workforce ranges from full-time pathologists and death investigators to a wide network of part-time and on-call death investigators. Training people on a standard process for collecting sensitive data presents a challenge due to the broad and diverse nature of their workforce. He emphasized that being a minority of any kind does not cause suicidality, but rather exposure to external forces such as harassment, discrimination, and other associated stressors are likely what is linked to increased suicidality.
Sexual orientation and gender identity can be captured through three methods: from family or friends at the scene of death, through funeral directors, or through next-of-kin interviews 4 to 9 months after the event. There are some pros and cons to each method. Although obtaining information at the scene of the death is the most timely method, family and friends might not be on scene or may be too distraught to answer questions. Funeral directors can provide standardized information on the death certificate, but the information comes from a single informant and may not be accurate. Next-of-kin interviews provide an opportunity to obtain rich data and connect the family to appropriate resources, but families might not respond to calls or might not provide accurate information.
Dr. Staley noted that they had collected sexual orientation and gender identity information for nearly all of the 1,400 next-of-kind they contacted in 2021. Only 11 percent refused to provide information. Since the program began in October 2018, there had been few concerns from next-of-kin about the additional questions and those concerns had been easily addressed.
Suicide in Minority Youth
Arielle Sheftall, Ph.D., University of Rochester Medical Center
Dr. Sheftall shared data on the increasing rates of suicide among 5- to 19-year-old children and youth across race and ethnicity. Between 2020 and 2021, American Indian/Alaska Native youth had the highest rate of suicide. However, the largest increases were seen among Black youth (100 percent increase), followed by American Indian/Alaska Native youth (62 percent), Asian/Pacific Islander youth (52 percent), and White youth (44 percent). Although data on multiracial youth was largely missing, the available data indicated a 7 percent increase from 2018 to 2021. There were also differences across race and ethnicity by gender. The increase in the suicide rate among American Indian/Alaska Native male youth (19 percent) was significantly higher than in female youth (7 percent). The increase in suicide rate among Hispanic male youth (71 percent) was significantly lower than in female youth (118 percent).
In 2021, suicide was the 3rd leading cause of death for Black youth. The increase in the suicide rate among Black male youth (57 percent) was significantly lower than in Black female youth (402 percent). Suicide was the number one leading cause of death among Black female girls aged 12 to 14. There is also an age disparity among Black youth. Black youth aged 5 to 12 were approximately twice as likely to die by suicide as compared to their White counterparts. This trend reversed at age 13. The circumstances leading to a suicide death also varied among Black youth. Black youth aged 5 to 11 and 12 to 14 were more likely to die at home and experienced a family or school problem or argument just prior to death. Black youth females aged 15 to 17 were more likely to have experienced a relationship problem; more likely to die within 24 hours of an argument; more likely to have a history of suicidal behavior or mental health treatment; and more likely to be diagnosed with depression dysthymia, or anxiety. Black youth males aged 15 to 17 were more likely to have experienced a criminal or legal problem and be diagnosed with ADHD.
Dr. Sheftall emphasized the need for understanding culturally-relevant risk factors for Black youth. There may be unique theories of suicide, pathways from ideation to behavior, and intervention needs for this population. Mixed method research approaches and engagement with community resources such as faith-based organizations and coaches are promising ways forward.
Autism Spectrum Disorder, Intellectual Disability, and Suicide: Conceptualizing Symptoms in Context
Joette James, Ph.D., ABPP, District of Columbia
Dr. Joette James talked about the complexities of autism spectrum disorder (ASD) and the risk of suicide and self-harm. ASD is generally associated with challenges in social-emotional reciprocity, non-verbal communication, and forming and maintaining relationships. Common ASD-related behaviors include repetitive movements, rigid thinking, special interest fixations, and either hypo- or hypersensitivity. ASD is commonly comorbid with intellectual disabilities, ADHD, depression, and anxiety—further complicating the landscape of risk.
Some of the risk factors for self-injury among children with ASD include a change in routine that precipitates behavioral or emotional outbursts, impulsivity, or darting away from an adverse sensory experience. Children with ASD often use self-injurious behaviors such as head banging, hair pulling, or hand biting as a distraction to their distress. There is also a high prevalence of bullying among children with ASD. Children with ASD have difficulties sharing information about their inner thoughts and feelings, creating a challenge for assessing suicide risk.
Diagnosis is an important barrier in addressing suicide and self-harm among individuals with ASD. Male children are more likely to be diagnosed than female children. White children are more likely to be diagnosed and diagnosed earlier than their Black or Hispanic counterparts. There is a wide range of symptom presentation and varying abilities to mask or learn acceptable behaviors. Dr. James emphasized the challenge of diagnostic overshadowing, in which ASD is focused upon over the presence of other conditions that might help us understand better the risk of suicide. There is a need for nuanced assessment tools, and research to better understand the relationship between ASD, comorbid conditions, and suicide.
Suicide Among Youth in the Child Welfare System
Cynthia Fontanella, Ph.D., Ohio State University
Dr. Cynthia Fontanella addressed the importance of focusing suicide research on youth in the child welfare system. These youth are at higher risk for mental health disorders, substance use disorders, poverty, child maltreatment, parental psychopathology, and exposure to community violence than the general population. Although there is a research gap in this at-risk population, Scandinavian data show that youth in child welfare or the juvenile justice systems are between 3 and 5 times more likely to die by suicide, 4 times more likely to have attempted suicide, and 2.5 times more likely to seriously consider suicide than youth in the general population.
Youth in the child welfare system are exposed to certain risk and protective factors across the socio-ecological model. At the individual level, they are more likely to have mental illness, substance abuse, and impulsivity. At the relationship level, they are more likely to experience conflict, violent relationships, or a family history of suicide. At the community and society levels, they are more likely to have poor access to health care, higher availability to lethal means, and greater exposure to unsafe media portrayals of suicide. These risk factors could be offset by protective factors such as psychological or emotional wellbeing, family and school connectedness, exposure to a caring adult, and coping skills.
Dr. Fontanella and her research team conducted a study in Ohio, integrating 2010-2020 data from multiple sources for more than 2.2 million youth aged 5 to 25. Suicide death among youth in the child welfare systems was 2.5 times higher than the general population and nearly 3 times higher than the Medicaid population. Suicide death was 3.4 times higher among female youth aged 5 to 14 in the child welfare system than the general population. Suicide deaths among youth removed from their home due to parental issues were 3.7 times higher than the general population. Type of living environment was a risk factor for suicide, at 5.2 times higher among children living in detention centers and 3.4 times higher among children living in congregate care. In their study, non-Hispanic Black youth were 72 percent less likely to die by suicide than their White counterparts.
Dr. Fontanella highlighted their finding that 90 percent of youth in the child welfare system who died by suicide had a physical or mental health visit within six months of death—providing a window of opportunity for intervention. Out-of-home placement, mental health diagnoses, substance use diagnoses, and seizure disorder also increased the odds of suicide death. Children in the child welfare system who had a history of self-harm were 16 times more likely to die by suicide than their counterparts. Dr. Fontanella said that there was a lack of surveillance data on suicide among youth in the child welfare system because the data tended to be siloed across disparate systems. She suggested that suicide in this population is likely underreported and advocated for the use of universal screening in the child welfare system.
Suicide in Juvenile Justice Systems Involved Youth
Gail Wasserman, Ph.D., Columbia University
Dr. Gail Wasserman talked about the complexities of the juvenile justice system and its implications for mental health services and suicide prevention. There are significant challenges in working with youth who are in the justice system, such as fragmented and siloed systems, difficulties with follow-up, and confidentiality restrictions. The juvenile justice system has a complex infrastructure consisting of courts, probation departments, and detention centers—each with its own system of records, making it difficult to track a child as they move through the system. Further, each state and county has its own system with varying processes.
Nevertheless, it is important to track youth in the juvenile justice system because they face higher risk of mental health issues and suicide. For instance, this population is more likely to experience challenges in their family or neighborhood, issues within their peer networks, and school problems. Contact with the justice system is itself a risk factor for suicide. Starting in the 1990s, the justice system began consistent universal screening initiatives, screening approximately 72 percent of youth for mental health. However, only 22 percent of those youth received a referral to mental health services. Among those 22 percent, only 15 percent initiated treatment.
To improve the linkage between screening and referral, Dr. Wasserman and her research team developed an initiative called Project Connect that slowly began to improve referral rates. Eventually, they developed a web-based clinical decision support system that classifies clinical need and provides real-time county-level referrals for behavioral health services. Since its initiation, there has been an 11-fold increase in the number of youth with behavioral health needs who receive a referral and a nearly 17 percent increase in treatment initiation.
Dr. Stacia Friedman-Hill invited panelists to join a question-and-answer session.
Question: What is your approach to using epidemiological data to incorporate multi-level, multi-factor perspectives (i.e., contributing factors) and what information or tools are needed to make these analyses easier?
Answers: Dr. Blosnich talked about hidden factors such as intergenerational poverty, systemic racism, historical racism, or hetereosexism—likening the study of these factors to astronomers studying planets they cannot see but can measure using proxies. It is important to discuss these kind of findings in a way that is interpretable and incorporates context. In a field that cannot even collect LGBTQ+ data, there are certainly no valid instruments for measuring these large-scale, largely hidden factors that are founded within discriminatory structures aimed at vulnerable and marginalized populations. He added that sexual orientation and gender identity is not in itself a risk factor.
Dr. Luby noted that the commonality among high-risk populations is that they are marginalized and invalidated and there has been an increase in organized efforts to further marginalize these individuals. It is a very important factor that needs to be measured.
Dr. Fontanella said that it was important to consider both individual and community level factors. At the community level, some of the tools that are helpful are the area resource file, the census data, the area deprivation index, and geographical location data.
Dr. Boyd explained that all sociodemographic factors, whether race, ethnicity, or gender identity, are not in themselves a risk factor. They are characteristics. In addition, there is a socioeconomic poverty impact on these populations based on systematic structure that needs to be addressed.
Dr. Friedman-Hill agreed and added that there had been some criticism in the field about too much focus on individual factors and not enough acknowledgement of systemic factors such as racism. Instead of a focus on race, there should be a focus on the systemic racism, perceived discrimination, and time trends following events such as racial violence or policies that stigmatize LGBTQ+ youth. It is an area of research that has only begun to advance.
Dr. Warner agreed about the need for research on structural racism and added that it was still important to accurately collect information at the individual level. From an epidemiological perspective, there is a need to accurately identify an individual and avoid misclassification bias. If there is an identified risk, it is important to understand the magnitude of that risk.
Dr. Blosnich clarified that the idea was not to ask death investigators to collect a large amount of contextual data but that it was incumbent on the research community to identify these factors and link them to mortality data to understand the associations.
Dr. Wasserman raised an idea to overcome the discomfort of asking sensitive information about suicide. People tend to be much more comfortable talking to an anonymous computer, even if they know that a human will review the information later. This is particularly true when the person who needs to ask the questions is a law enforcement or probation officer. A pre-recorded set of questions might take away some of the discomfort and ultimately result in more standard, reliable data.
Ms. McGivern added that death investigators ask these questions in a systematic way and the issue might not be about discomfort asking the questions, but rather the timing of asking those questions within an emotionally charged environment. It can be an unsafe environment as well, with a lot of deferred anger from families who might push back on questions related to suicide or gender identity. Based on Dr. Staley’s presentation, it seems that there is more success asking these questions in a follow-up period.
Dr. Staley said that they would continue the follow-up program even if they obtained all the information from families at the scene. Sometimes repeating questions is helpful because families might learn something that they were not previously aware of. It is also an important postvention practice that could help prevent future suicides.
Question: Where is the “so what” in the field? How should the field move from data to action?
Answers: Dr. Sheftall answered that it is important to focus interventions on the risk factors that are uncovered. This depends on the quality of the narrative, which is variable from state to state. A comprehensive narrative helps researchers understand what is happening in the child’s life and where there might be intervention points.
Dr. Blosnich said that is important to get policymakers to the tables to demonstrate the importance of medical examiners and coroners in identifying emerging risks and the need for funding, personnel, and training to leverage this important resource.
Dr. Pinneri noted that most death investigation offices are county-funded, most federal funds given to states do no trickle down to counties, and staff are often tasked with writing grant proposals for the resources they need. However, the staff are already over-tasked and are not grant writers. The offices also tend to lose sight of their public health role because they are under-resourced and often tasked with additional duties such as testifying in court. Despite the critical importance of the data generated from death investigators, the offices need support and resources to do their work.
Dr. Warner talked about CDC efforts to help strengthen their ties with the medicolegal death investigations, as well as their broad data modernization efforts.
Dr. Pinneri noted that the fact that understanding the people who die can help people who live tends to get lost. There may be those who hesitate to fund efforts collect data on those who have died and it is important to convey how important this information is.
Question: Is there follow-up with children who have mental health diagnoses to ensure that they are receiving treatment and are compliant with medications and appointments? Are they lost to follow-up?
Answer: Dr. Sheftall said that NVDRS data shows whether a child is in current or past treatment, but not whether there is compliance. It is an important topic that could help improve data collection.
Session 6: State-Based Approaches to Child and Youth
Suicide Investigations and Suicide Prevention Programming
Session Introduction: Diane Pilkey, RN. MPH, National Center for Fatality Review and Prevention, MCHB
Discussant: John Ackerman, Nationwide Children’s Hospital
The Utah Youth Suicide Research Project
Michael Staley, Ph.D., Utah Office of the Medical Examiner
Dr. Michael Staley talked about his role as a psychological autopsy examiner for the Office of the Medical Examiner in Utah. His position was created to investigate the causes of suicide and his work eventually focused on youth suicide. His work had both community and legislative champions, who sought to improve Ohio’s prevention and intervention strategies. He also collaborated with the University of Utah School of Medicine, which provided a pool of medical students and residents to help conduct interviews with families.
Dr. Staley conducted a two-phased study of 201 children aged 10 to 17 who died by suicide between 2016 and 2019. The first phase aimed to de-silo and aggregate data from multiple sources. The second phase was to interview at least one parent or guardian and up to three friends for every youth who died by suicide. By asking the same questions to each person, they were able to both validate the story and also see different perspectives. Of the 201 children who died by suicide, the team conducted at least three interviews for 34 cases and one or two interviews for 78 cases. There were 30 next-of-kin who refused to participate and 59 who did not respond to the invitation to participate. Dr. Staley noted that some of the families had moved or changed their phone number rather than actively rejecting participation.
The results of the study found that 1 in 4 children were in a romantic relationship that had just ended or was on the verge of ending. Nearly half felt that they did not belong in their community. Half had been bullied and more than half had witnessed a friend being bullied. According to parents, 1 in 4 were bullied online, but Dr. Staley emphasized that this number will likely be higher after they conclude their interviews with peers. He added that the most difficult conversation to have with parents was not about sexual orientation but rather physical, emotional, and sexual abuse. Parents either did not know or felt powerless to stop it. Just under half the children were reported to be impulsive decision makers.
Notably, there was evidence of suicide contagion and diffusion, with 1 in 5 who had experienced the loss of a friend, 1 in 10 who lost a best friend, 1 in 3 who lost a family member to suicide, and 2 in 5 who knew someone who died by suicide. Approximately two-thirds of the suicide deaths involved a firearm, two-thirds of which involved a firearm that was stored unlocked. Notably, over half of these youth had at least one school change- e.g. moving from public school to private charter school and in some cases back to public school, and nearly a third of youth dying by suicide had a record of school truancy or extensive absences, with a third having a history of school disciplinary action that was proximal to the time of death.
Going forward, Dr. Staley and his team will complete this study and hope to be funded for another generation of data collection. They are also sharing their data with the Utah Suicide Child Fatality Review Committee, NVDRS, and State Unintentional Drug Overdose Reporting System and plan to share their study model with other medical schools.
Maryland’s Approach to Youth Suicide Investigations and Suicide Prevention Programming
Holly Wilcox, Ph.D., Johns Hopkins University, MD
Dr. Holly Wilcox reviewed Maryland’s youth suicide prevention initiatives. Although Maryland has one of the lowest overall suicide rates in the U.S., a high number of deaths in Maryland are classified as undetermined, highlighting the need for improved data collection and adjudication strategies. Their centralized Office of the Chief Medical Examiner provides access to real-time data, which show that the percentage of suicides in Maryland that occur in youth has risen from 3 percent to just over 6 percent coinciding with the years of the COVID-19 pandemic. Notably, there is an alarming upward trend of suicide ideation among middle and high school students.
This growing crisis prompted Maryland to embark on several suicide prevention initiatives. For instance, the Office of the Chief Medical Examiner and the Suicide Prevention Office developed a Suicide Data Warehouse, initiated psychological autopsies for some cases, and recently helped establish legislation to create a suicide fatality review. They have also worked with their Health Information Exchange to gather data across various sources and integrate them with data from the Office of the Chief Medical Examiner. These data can be used for predictive analytics and to inform the suicide fatality review process. They also plan to examine suicide by race and ethnicity, as well as manner of death.
The first 11 cases that were analyzed by psychological autopsy showed some common themes related to social interaction, interpersonal conflict, and social isolation. The data also identified alarming video game behavior in which the individual was intentionally letting themselves be killed within the video game. Another major trend was access to firearms. Half the suicide deaths in Maryland are associated with firearms and the psychological autopsy team found that families generally did not consider a firearm in the home to be a risk. They also do not have access to information about firearm safety but might have taken action had they been informed.
Dr. Wilcox reviewed various legislative efforts that could positively impact the state’s suicide rate, such as a bill to end child poverty, improvements to the implementation of extreme risk protection orders, and the Blueprint for Maryland’s future that allocates funding for schools. Additionally, Maryland closed its long gun loophole in its firearm legislation and passed Jaelynn’s Law that extends penalties for parents who do not safely store a firearm away from their children aged 17 and younger. Maryland’s Center for Guns Policy Solutions developed a safe storage map that identifies locations for people to safely store their firearms when an at-risk person is visiting their home. Finally, Maryland recently created a gun violence navigator position that will interface with people in the healthcare system who have firearms.
Highlights from CFRT in Harris County, Texas
Angie Hayes, Ph.D., Harris County Child Fatality Review Team, TX
Dr. Angie Hayes talked about the challenges in and efforts to address youth suicide in Harris County, Texas. Harris County is home to a large number of children, accounting for one-sixth of Texas’s entire child population. Accordingly, there is an average of 250 non-natural child fatalities each year, of which approximately 1 in 11 is a suicide death. Over the last six years, there has been an upward trend in child and youth suicide. Although White children represent 22 percent of the population, they account for 42 percent of suicide deaths.
Dr. Hayes reviewed some of the challenges in addressing suicide in Texas. There is a lack of uniformity in death determinations and standard data collection across Texas. Of the 254 counties in Texas, 16 hvae medical examiner/forensic capacity of some sort. The remaining 238 counties rely on the over 800 Justices of the Peace (JP) within those counties to rule on deaths. These elected JPs are not required to have inquest training and they sometimes need to call on a county judge to help out when they aren't able to attend a death investigation.
There are also serious barriers to access to mental health care, primarily as a result of long distances and a lack of public transit. Harris County houses 25 public school districts and 338 private schools, each requiring long administrative processes prior to implementing suicide prevention programs. Once in the schools, programs are challenged by the number of different languages spoken by the children and the large proportion who are economically disadvantaged. Without communication and collaboration, there are likely several areas of redundant efforts and many missed opportunities for synergy. There is also insufficient funding to expand their trauma-informed forensic interviews with bereaved families, which misses a window of opportunity to link families to services and resources. Dr. Hayes illustrated the risk of suicide related to natural disasters, such as hurricanes, through a case study of a child who died by suicide because he had been displaced from his home and had to live with an abusive parent after Hurricane Harvey.
Despite these challenges, there are efforts in Texas that have improved suicide prevention. For instance, their Houston Area Suicide Prevention Coalition was recently revitalized and brings together different organizations for collaboration. New training has been implemented in Texas Children’s Hospital and the number of beds available for individuals with mental health issues has expanded. Harris County has also obtained funding for their Child Fatality Review Team, which is housed within a research medical hospital—providing an opportunity to blend research, implementation, and review. Additionally, they received a grant to develop a natural language processing algorithm to analyze youth suicide data. So far, these data have identified high-risk situations such as parental argument resulting in loss of cell phone access and recent CPS involvement. Although they recognize that not all data on LGBTQ+ youth is captured, the data still show a high prevalence of suicide among this population. Finally, Texas legislation now permits child fatality teams to share data without requiring an MOU or DUA.
Child & Youth Suicide Death Review and Prevention in Colorado
Kelly Dougherty, M.P.H., Colorado Department of Public Health and Environment
Ms. Kelly Dougherty reviewed how child death reviews are conducted in Colorado, which became a more comprehensive process in 2013 with the passage of the Colorado Child Fatality Prevention Act. There currently are 43 local review teams across 64 counties, which aim to identify the risk and protective factors that could inform prevention efforts across the state. The teams review death of children and youth under age 18 who die by injury, violence, or undetermined cause. Between 2017 and 2021, suicide was the leading cause of death in children and youth in Colorado, with a rate of 6.4 per 100,000—twice as high as the national rate. There are disparities across race, ethnicity, place of residence, sexual orientation, and gender identity. Additionally, males were more than twice as likely to use a firearm than females, providing a possible target for prevention strategies.
Data on circumstances associated with suicide deaths collected by the child death review teams show that more than 25 percent of children and youth had communicated suicidal thoughts or intentions. More than 25 percent had experienced a loss of another person—82 percent of which were suicide deaths. Intimate partner problems were a contributing factor for nearly 20 percent of youth aged 15 to 17—twice as much as their younger counterparts. Interpersonal crises, such as bullying, were more likely among youth aged 10 to 14 and school stressors were more likely among their older counterparts. Divorce and family discord were comparable across all age groups, but a recent argument with a parent or caregiver were more frequent among youth aged 10 to 14. Notably, 34 percent of children and youth had experienced child maltreatment (neglect, physical abuse, emotional abuse, or sexual abuse) as a victim before their death.
Ms. Dougherty reviewed efforts to improve their suicide prevention efforts. In partnership with the Office of Suicide Prevention and the Colorado NVDRS, her team developed a suicide death investigation form that captures data on sensitive topics such as sexual orientation, gender identity, and mental health history. Although there have been some challenges with the deployment of this form, such as inadequate staffing and difficulties asking grieving families to address sensitive subjects, there have also been successes. An evaluation of 10 coroners who used the form found that the form helped them conduct better investigations and better understand the factors involved in suicide. The coroners and medical investigators have also partnered with local schools, hospitals, public health agencies, and local health coalitions to increase their involvement in prevention and postvention efforts. Ms. Dougherty and her team are now focused on incorporating questions about life stressors, holding informal discussions, and presenting data to the communities to identify additional approaches for improvement.
Dr. John Ackerman invited panelists to join a question-and-answer session.
Question: What was the most important step in de-siloing data or processes in your state?
Answers: Dr. Staley said that tenacity was key. It can be a challenge to go up against a structural barrier—after three years of legislation, they are still not all the way there. There are still policies needed to allow for the level of data sharing they need. The legislators do not put up a lot of resistance when the goal is to prevent youth from dying by suicide. It helps to give policymakers reasons to engage.
Dr. Wilcox emphasized the importance of establishing partnerships and trust. Patience is also important because the administrative hurdles, such as MOAs or IRBs, take much longer than expected. It is also helpful to leverage senators and other policymakers who have lived experience with suicide. For instance, one senator was instrumental in creating the suicide fatality review board. Advocacy groups and family members with lived experience are another resource to tap because they often want to help but do not always know how.
Dr. Hayes echoed the importance of tenacity and persistence. Progress is often little by little.
Ms. Dougherty said that the 2013 legislation was critical to their efforts because it provided dedicated funding each year to support reviews at the local level. Prior to this legislation passing, the reviews were much more limited and they were all conducted at the state level. Working at a community level has really improved the quality of their review and the recommendations they make can be more relevant to the local people. Data sharing was another important part of the statute that provided access to data and provides a means to working with other agencies. One thing that is missing is more of the family voice and experience. Some of the coroner’s offices have been hiring family advocates, which would be a great step towards supporting bereaved families.
Question: Recent relationship breakup was a common risk factor across several presentations. Considering that there have always been breakups, it is worth considering what had changed to make them such a high-risk factor now. Are there any other “why now” factors that should be addressed?
Answers: Dr. Wilcox suggested that the high volume of firearms in homes could be a “why now” answer. A youth in crisis may know that there is a gun in the house and how to get to it. There is a continued need to enhance coping skills in schools so that children have tools for managing highly stressful situations and de-escalating themselves.
Dr. Ackerman added that there is reduced connection and support and increased access to a range of lethal means.
Question: What are your thoughts on adapting scales, protocols, and screening tools that are used in a research setting for use in a post-mortem setting?
Answer: Dr. Staley said that his team found that their instrument was no longer valid for children under the age of 14. There are different experiences and there needs to be better tools specific to those populations.
Dr. Gorniak added that there were vast generational differences, but one common phenomenon was the lost of patience. The world continues to evolve and there is a need to understand why some people might respond to an event differently than another.
Summary and Future Steps: Federal Suicide Prevention Partners
Lisa Colpe, Ph.D., M.P.H., NIMH
Dr. Lisa Colpe invited Dr. Warner to share her thoughts on the work going forward. Dr. Warner highlighted several helpful resources to attendees, such as the new Collaborating Office for Medical Examiner and Coroners at the CDC, their Interagency Workgroup on Medicolegal Death Investigation , and CDC’s Data Modernization Initiative with medical examiners and coroners that will establish common data elements and data standards. Dr. Warner shared her key takeaways, including understanding postvention, current ideas on suicide ideation among children, tools and screens for post-mortem settings, and increased awareness to child suicide across a wider audience.
Dr. Pearson reviewed some of the translational research topics in youth suicide research for which NIMH has issued funding announcements. Dr. Abigail Soyombo talked about the NIH NeuroBioBank , a network of six post-mortem brain banks that facilitates research to better understand human brain function and disorders.
Dr. Mack provided an overview of CDC’s efforts in suicide prevention, such as their Comprehensive Suicide Prevention Program across 17 states, their Suicide Prevention Resource for Action report of strategies and resources, and technical assistance to help states with suicide clusters.
Ms. Pilkey mentioned the National Center for Fatality Review and Prevention data availability and train-the-trainer event in 2024 that will focus on pediatric death.
Dr. Ackerman announced an open access book called Youth Suicide Prevention and Intervention: Best Practices and Policy Implications .
Dr. Pearson concluded the meeting by summarizing her key takeaways, including the limitations of missing data, proximal and distal risk factors for suicide in children and youth, opportunities for preventive interventions, measures to improve death investigations, and collaboration between data collection systems and agencies,