Inconsistencies in Overdose Suicide Death Investigation Practice and Potential Remedies Using Technology: A Centers for Disease Control and Prevention Consultation Meeting Summary

Abstract

Introduction:

It is widely accepted that suicides—which account for more than 47 500 deaths per year in the United States—are undercounted by 10% to 30%, partially due to incomplete death scene investigations (DSI) and varying burden-of-proof standards across jurisdictions. This may result in the misclassification of overdose-related suicides as accidents or undetermined intent.

Methods:

Virtual and in-person meetings were held with suicidologists and DSI experts from five states (Spring-Summer 2017) to explore how features of a hypothetical electronic DSI tool may help address these challenges.

Results:

Participants envisioned a mobile DSI application for cell phones, tablets, or laptop computers. Features for systematic information collection, scene description, and guiding key informant interviews were perceived as useful for less-experienced investigators.

Discussion:

Wide adoption may be challenging due to differences in DSI standards, practices, costs, data privacy and security, and system integration needs. However, technological tools that support consistent and complete DSIs could strengthen the information needed to accurately identify overdose suicides.

Introduction

Knowing how and why people die—that is, the cause and manner of their death—is important for public health surveillance and prevention. It can inform decisions related to the allocation of resources for research and prevention of suicide-related morbidity and mortality. Therefore, correctly classifying decedents’ manner of death (MOD; i.e. natural causes, accident [note ¹], homicide, suicide, pending investigation, could not be determined [or undetermined] (1)) is critical.

Suicide—defined as fatal, self-directed violence with the intent to die (2)—is one of the leading contributors to death across all age groups in the United States (U.S.), surpassing motor vehicle deaths since 2009 (3). In 2019, suicide was the second leading contributor of death for 10 to 34 year olds and the fourth leading contributor of death for 35 to 44 year olds (3). From 1999 to 2018, suicide rates increased among all age groups under 75 years, among males and females, and by urbanization levels (4,5). Additionally, age-adjusted suicide rates rose by 35% (10.5-14.2 per 100 000 standard population) during that timeframe (4). Still, multiple studies suggest that suicide rates are widely underestimated—between 10% and 30%— (6 -10) due to MOD misclassification (11, 12).

Misclassification of Suicide Manner of Death

Suicide misclassification may result from stigma, legal, religious and political pressures, limited death scene investigation (DSI) resources, and/or differences across jurisdictions in the laws, training, practice, and burden of proof needed for determining that a person died by suicide (6, 8, 13, 14). Although the National Institute of Justice (NIJ) has guidelines for DSI (15) and the National Association of Medical Examiners has a well-known Guide for Manner of Death Classification (16), uptake of these guidelines could vary given that death investigation is not systematic within or across all jurisdictions (17). Consequently, many suicides are reported as accidental (18 -20) or undetermined MOD (9, 20).

Drug poisonings (overdoses) are one of the causes of death that can be frequently classified as accidental or undetermined rather than suicide, typically because the available information is not sufficient for a suicide MOD determination. This is of critical public health importance given the significant increases in drug overdose deaths in the U.S. over the past two decades. The age-adjusted drug overdose death rate in 2017 (21.7 per 100 000) was 3.6 times the rate in 1999 (6.1 per 100 000), increasing 9.6% from 2016 to 2017 alone. Synthetic opioid drug overdose death rates (excluding methadone) increased 45% between 2016 and 2017 (11). With respect to MOD, between 2008 and 2010, 8% of drug overdose deaths in the U.S. were classified as undetermined MOD, with estimates ranging widely (1%-85%) across states (21). In 2017, 5% of drug overdose deaths overall were classified as undetermined MOD (11). The overlap in risk factors for drug misuse and suicide (e.g., histories of self-harm, substance abuse, mental health conditions) is likely one of the reasons for this potential misclassification (13, 22, 23).

Inclusion of social and psychological evidence for MOD determinations is helpful, particularly for overdose deaths where intentionality may be less overt than for other self-harm methods (e.g., hanging) (12, 24). Before, when physical autopsies were insufficient to determine MOD, psychological autopsies (i.e., process of gathering in-depth psychological and psychiatric information to retrospectively assess the state of mind and intention of decedents (25)) were sometimes used to help determine MOD. However, these could fail in providing critical information regarding intent among some subgroups. For example, because they may not always include an analysis of social networking sites (SNS; e.g. Facebook), they may fail to identify decedents (e.g., adolescents) who might have used these to express their thoughts prior to suicide (e.g., “e-suicide note”) (26 -28).

As technology advanced, new “social-mobile autopsies” can help expand DSIs by accessing and analyzing messages, photos, videos, and conversations saved to the decedent’s computer, mobile devices, and SNS accounts as part of medicolegal investigations (26). Even when a decedent’s devices are inaccessible (e.g., password-protected), publicly available SNS information has proven helpful in determining MOD in potential suicide investigations (28).

Use of Mobile Technology in DSI

Mobile devices, software, and applications can help investigators navigate the process of collecting and documenting circumstances surrounding causes and manners of death. For example, the Forensic Institute for Research and Education at Middle Tennessee State University developed the Checklist App for Scene Examination (29), a smartphone app that assists first responders in capturing and preserving crime scene evidence (e.g., text, photos, video, audio, global positioning system [GPS], dates, times). Additionally, the U.S. Fish and Wildlife Service Forensic Laboratory developed a tablet app to guide agents through the process of documenting and processing crime scenes, reducing the time traditionally spent by agents per scene from 4 to 5 hours to 45 minutes (30).

Private businesses have also developed technological tools for DSIs. GoCanvas created the Death Scene Checklist Mobile App that allows first responders to reference standard DSI procedures (31), and an electronic Death Investigation Tools and Equipment Checklist to help medical examiners, coroners, accident investigation teams, and law enforcement officers have the proper tools and safety equipment to conduct investigations (32).

Mobile tools can help address some of the challenges resulting in the misclassification of suicide overdose deaths as unintentional or undetermined intent, by providing investigators access to job-aids and training to improve their preparedness, quality, and efficiency in processing DSI information. Furthermore, their wide adoption could reduce the misclassification of suicide by improving the standardization of medicolegal death investigations (e.g., terminology, definitions, procedures), potentially resulting in the implementation of evidence-based overdose-related suicide prevention strategies and communication improvements across disciplines and jurisdictions (33).

Death Scene Investigations Meetings—Overdose Suicide MOD Determinations

In March 2015, the Centers for Disease Control and Prevention (CDC) convened a meeting of DSI professionals and subject matter experts to explore individual, systems-, and place-related barriers to accurate MOD determinations, and to brainstorm potential solutions to address these challenges. At the meeting, participants acknowledged the need to reduce inconsistences in MOD determinations across jurisdictions, and noted the potential value of technology for this purpose (33).

As a follow-up, during Spring-Summer 2017, CDC convened a series of virtual and in-person meetings with professionals representing diverse DSI fields (e.g., law enforcement, coroners, medical examiners, first response, toxicology, research). These aimed to obtain a better understanding of the types of information collected during overdose DSIs and the potential features, challenges, and opportunities for developing and implementing a mobile tool to improve DSI information collection and subsequent suicide MOD determinations.

Methods

This article collectively presents a summary of individual participants’ thoughts, ideas, and perspectives from the Spring-Summer 2017 meetings. Their individual perspectives should not be interpreted as consensus, guidance, or recommendations. This project did not seek to provide recommendations to the field, nor it followed a systematic data collection process.

Virtual Meetings

Five two-hour virtual meetings were held with DSI professionals from Alaska, Florida, Ohio, Maryland, South Carolina, and Utah during March-April 2017. These states were selected because of their participation in CDC’s National Violent Death Reporting System (34) and/or their variation in proportions of undetermined or suicide deaths. National Violent Death Reporting System experts, prior 2015 meeting participants, and other experts nominated DSI investigators as potential meeting participants from each of these states. Additionally, an experienced medical examiner and a member of the planning team were present at all virtual meetings. A maximum of eight participants were invited, allowing increased virtual participants’ interaction. Participants included representatives from medical examiner and coroner offices (n = 14), law enforcement (n = 6), toxicology (n = 4), and suicidology/research (n = 3).

Using the Adobe Connect 9 web conferencing system (35), the same experienced facilitator led all virtual meetings, which were divided into two parts: 1) Evidence to Determine Poisoning/Overdose MOD, and 2) DSI Technology Tools and Features. The first part allowed participants to comment on different types of evidence collected during poisoning/overdose DSIs, often relied upon to determine MOD. Subsequently, participants discussed DSI tools currently used, and suggested features for a hypothetical novel tool for DSI evidence collection that could improve information collection and MOD determinations.

In-Person Meeting

Held in Atlanta, Georgia, in July 2017, this all-day in-person meeting expanded on the virtual meetings’ findings, and discussed potential benefits and challenges for the widespread adoption of a DSI mobile information collection tool. In-person meeting participants included representatives from the virtual meetings, as well as subject matter experts (N = 29). It consisted of three main topic areas: 1) Overview of prior DSI and suicide MOD work; 2) Presentation, contextualization, and expansion of feedback received during the virtual meetings; and 3) Challenges, opportunities, and implications for adopting a mobile DSI tool. Facilitated discussions included three individual and group activities.

Results

Evidence to Determine Overdose Suicide MOD

Participants discussed the types of evidence they regularly seek and consider most important for determining MOD of drug overdose fatalities. They distinguished between evidence found while evaluating the death scene, body, and decedent’s profile.

At the death scene

The presence of suicide notes—in paper or electronic format (e.g., text messages, social media, email)—were considered one of the most important types of evidence in determining suicide MOD. Additionally, any suicide-related internet searches found on the decedent’s electronic devices could help determine if an overdose was a suicide. However, participants noted that privacy and security settings across devices, websites, and apps impose access challenges for investigators.

It was also deemed important to document the presence of any drugs or evidence of drug use—illicit, prescription, or over-the-counter—and drug paraphernalia found at the scene, as this evidence helps discern the decedent’s recent and long-term drug use behaviors. The importance of matching valid prescriptions against remaining pill counts and drug supplies helps investigators determine if the decedent used medications as directed.

Participants emphasized the importance of interviewing key informants who were at the death scene. Participants also indicated that the way the interview was conducted should ensure the stigma associated with suicide drug overdoses does not hinder the level of disclosure and insight witnesses can provide.

Other types of evidence noted as important to collect during poisoning/overdose DSIs included medical documents, books, and other reading materials with suicide-related themes, frequently referred to as “suicide manuals.” Additionally, evidence of getting one’s affairs in order and inconsistencies in the setting (e.g., hotel room with no suitcase) help provide the burden of proof necessary to determine if the overdose fatality was a suicide.

Body examination

Visible injection-site scars, unusual odors, and the body’s position at the death scene were noted as helpful for determining MOD. Additionally, it was stated that an autopsy could determine whether there was a mass of pills (e.g., “pill cake”) in the decedent’s upper digestive system or other overdose-related pathologies (e.g., cerebral hemorrhage, pulmonary edema) that could help rule out non-suicide causes of death (e.g., heart attack).

Decedent’s profile

Beyond the evidence found during the postmortem examination and at the scene, participants noted the importance of establishing a profile of each decedent. Such a profile describes who the decedent was, his/her family composition, and any recent stressors (e.g., divorce, job loss) that could have been directly or indirectly associated with his/her overdose death. Additionally, participants noted the importance of establishing decedents’ medical history, both in terms of mental health conditions that could have altered his/her state of mind or caused an unintentional overdose (e.g., dementia) and physical health problems (e.g., painful, incurable conditions) that are associated with increased risk of suicide.

Reviewing the decedent’s prescription drug monitoring program reports was noted as important to establish current substance use disorder or drug misuse patterns (e.g., “doctor shopping,” filling prescriptions early). The decedent’s prescription and illicit drug use history can also be assessed through information provided by relatives, friends, and physicians or as indicated by past hospital and/or arrest records. Participants also noted the importance of documenting the reason for any hospital stays, police encounters and/or arrests, and previous substance use disorder treatment and/or suicide attempts.

Importance of key informant interviews

Participants frequently noted the importance of having guidance for investigators on how to conduct family/friend/witness interviews and to collect scene descriptive information. The use of systematic modules for information collection (e.g., creating a “data collection wizard”) was considered potentially beneficial to experienced and new investigators. Key informant interviews should be conducted soon after the death, as perceptions may change over time. Such interviews may provide important information otherwise unavailable to investigators.

Challenges in Determining Suicide MOD Among Drug Overdoses

Several participants emphasized that suicide MOD determinations require affirmative criteria—information to support the claim. Otherwise, these cases are likely to be categorized as unintentional or undetermined—views consistent with reports in previous literature (9, 18 -20).

Aside from suicide notes, most participants said documented mental health issues, previous suicide attempts, and recent life stressors were the types of information that could best help identify a suicide-related overdose. However, investigators face diverse challenges in obtaining these and other types of information needed to correctly categorize suicide-related MOD in the context of a drug overdose, such as:

Reclusive Decedent and/or Lack of Cooperation by Key Informants:

At times, key informants are reluctant to disclose information that could provide sufficient information for a death to be ruled a suicide (e.g., mental health history not documented elsewhere).

Delayed Discovery of the Body:

As time passes, the decedent's body decomposes. Therefore, corporal evidence that can help determine the MOD could be lost or prove inconclusive.

Investigator Differences:

Variability in the investigators' experience, training, resources and workload could affect the death scene information collection process.

Restricted Access to Drug/Medical and Computer/Phone Records:

States have different laws, some of which restrict investigators' access to drug/medical and computer/phone records. Nevertheless, this information could help establish the decedent's profile.

On-site Technology Restrictions:

Connectivity and information technology (IT) policy restrictions can impede investigators from accessing resources that can help them best document and understand the evidence found at the scene.

Lack of Uniformity in Burden of Proof:

Some jurisdictions require more information than others to classify a decedent's MOD as a suicide. For example, one jurisdiction may find that having a history of drug-related suicide attempts even though toxicology results were inconclusive is sufficient to determine suicide as the MOD. However, the burden of proof in another jurisdiction may render that evidence insufficient, and classify it as undetermined instead of suicide. In spite of these burden of proof differences, participants noted similarities in the types of information they would associate with a suicide MOD (e.g., suicide note).

Inability to Document what is Not Found on the Scene:

At times, an investigator's greatest finding can be what they did not find (e.g., no suitcase in the hotel room where the decedent was found). However, participants noted that it is not always standard practice to document what is not found (i.e., “non-findings”). While documenting what is not found at the scene could be considered circumstantial and allow for speculation, some “not founds” could alert the investigators to the potential MOD (e.g., suicide, homicide).

Current DSI Evidence Collection Tools

Participants noted digital cameras as investigators’ primary tool to take pictures of the body, location, and death scene evidence. These are preferred over cell phone cameras, due to the higher quality images produced, and because taking pictures with a cell phone can affect their interactions with family, friends, and witnesses at the scene (i.e., looks “unprofessional”). Notwithstanding, participants also acknowledge the value of using cell phones to access online resources at the scene.

Investigators are frequently take notes and fill out multiple forms during DSIs. Specifically for overdoses, they usually need to complete inventory lists of all drugs found at the scene—specifying where they were found in relation to the body, the type of drug, amount, and prescribing physician, if available. Paper and pencil/pen are the tools commonly used to collect this data, as well as for note taking. Other specialized equipment for mapping and substance collection/testing are also used.

Participants noted that personal protective equipment (e.g., gloves, protective clothing) is often essential at overdose DSIs. However, its use may make it difficult to use technological devices for information collection (e.g., gloves impede use of touch-screen devices). Efforts to collect all available information during a DSI should not compromise investigators’ safety.

Brainstorming: A Mobile DSI App for Information Collection

In light of the information needed to determine suicide MOD of fatal drug overdoses and the challenges of collecting such information with currently available tools, participants envisioned a new technological tool for DSI information collection. They conceptualized a hypothetical DSI tool as a cell phone, tablet or computer app, rather than a stand-alone device. Potential helpful features were identified in six different categories: hardware, software, data security, data import/export, information access, and support tools (Figure 1).

Figure 1:

Summary of potential helpful features for a death scene investigations (DSI) Mobile Tool. GPS indicates global positioning system; ID, identification; IT, information technology; Mac, refers to Macintosh or Apple operating systems; MP, mega pixels; Win, Windows operating system.

Hardware

It is important for such a new tool to be available in a hand-held, lightweight, and durable mobile device. It should have a long battery life, high screen resolution, and be waterproof, given the variable environmental circumstances faced at different DSIs. While participants previously noted preferring stand-alone digital cameras, in considering a new tool, a ≥12 mega-pixels built-in camera would be desirable. Additionally, having laser distance measurement capabilities could facilitate frequent on-the-scene tasks. A fingerprint scanner could be helpful in identifying decedents at the scene, as well as a fingerprint or other type of device unlocking mechanism.

Software

Given participants’ preferences for a mobile device, having basic smart phone features is important (e.g., address book, note taking, voice recording). Specifically, a DSI mobile app could assign unique ID numbers, GPS coordinates, and time and weather stamps to each investigation. Drop-down menus with pre-loaded options can eliminate handwritten interpretation challenges and can expedite data analysis. Modules that allow for the step-by-step collection of information (e.g., checklist, skip-pattern wizard) could be useful for less-experienced investigators, especially for describing death scenes and interviewing key informants. Language translation features could assist in translating written materials found on the scene, as well as in communicating with some witnesses. Voice recognition software could assist in notetaking and accessing app and/or device tools whenever the investigator is unable to use his/her hands to do so.

Data security

Participants noted the importance of ensuring all information collected at the scene is secure. At the minimum, the device and/or app could have a lock/passcode mechanism; fingerprint access was suggested. Data encryption is essential and the device/app must be able to conform to local agencies’ IT safety policies and requirements.

Data import/export

Any information collected using such a mobile device or app will need to be uploaded onto the investigators’ agency servers for analysis, evaluation, and record keeping. To this point, participants stressed the value of cellular or satellite wireless connectivity for real-time data exports. The ability for pushed export to a server was also noted. Shared connectivity or the ability for multiple investigators to be working on the same case via an app across different devices could be valuable. The participants preferred exports to be in the most commonly used formats for pictures, video, data, and text (i.e., JPEG, MPEG, XLS/XLSX, PDF, DOC/DOCX).

Information access

Beyond real-time shared connectivity, participants noted the potential utility of multiple users being able to access the information collected—even across agencies. The ability to sort databases and query searches was suggested. In addition, given the envisioned mobile nature of this tool, participants mentioned it is important for the app to allow access to procedural guides and forms, as well as contact lists. Given the variety of mobile devices available across agencies and jurisdictions, it is important for the app to be compatible with Windows, web-based/internet browsers, and Mac operating systems—in that order of preference.

Support tools

Considering the diversity in background and professional training across DSI teams and jurisdictions, a DSI mobile app provides an opportunity for diverse users to access similar training features, databases, and resources to support their work. A DSI process checklist tool, consistent with NIJ and National Association of Medical Examiners' guidelines for DSI and MOD classification (15,16), could be an asset.

Challenges for Adoption of a DSI Mobile App

Participants emphasized the diversity in DSI standards and practices across jurisdictions, cost, and integration with current systems as challenges for the wide adoption of a DSI mobile tool or app. Given there is no overarching regulatory DSI structure, it will be challenging to design a tool that complies with all jurisdictions’ DSI standards.

The development, sustainability, and update costs associated with such a mobile app—including the costs for acquiring the app and mobile devices to use it, maintenance of those devices, IT requirements for system integration, and wireless data transmission charges—were concerning. While participants perceived value in investing in such a DSI tool, funding availability and uncertainties across jurisdictions and agencies were perceived to be barriers to adopting a DSI mobile tool. Allowing investigators to download the app to their own personal devices also presents data storage and security concerns, and would place their personal devices at risk of subpoena.

While participants generally supported having a mobile device or app for DSI information collection, its adoption could have negative effects. It could limit non-tech savvy investigators in their work, complicate processes for experienced investigators, and thwart investigator intuition. Among some communities, using a mobile device while conducting a DSI could interfere with building rapport and trust with witnesses.

However, participants acknowledged integrating a DSI mobile tool into existing record management and national electronic data reporting systems could facilitate the uptake of such new technology. Additionally, developers must ensure the tool is efficient, compatible with, and interoperable with numerous IT-systems. Not limiting the technology to one type of cause of death or MOD is essential to facilitate widespread adoption. Finally, participants expressed the need for widespread, low cost/free training on the use of any newly developed DSI tool. The tool must be flexible to accommodate the differences in forensic and law enforcement DSIs, allowing for broader jurisdiction-specific adaptations.

Discussion

Consistent with the findings from a prior CDC expert meeting (33), DSI meeting participants generally acknowledged that many of the challenges faced in overdose-related, suicide MOD determinations are related to the information available to meet the high burden of proof for affirmative criteria requirements, which varies across jurisdictions. Because quantity and quality of information available can vary from case to case, it is important to develop mechanisms and tools to assist investigators in systematically collecting as much available information as possible to best assist in determining each decedent’s MOD.

Technological tools that support more standard and routine evidence collection during DSIs could help gather the strong body of convergent evidence needed to distinguish suicides from other MODs. Such tools—for example, a DSI mobile device or app—could benefit experienced and newer investigators with different training backgrounds across jurisdictions. However, investigators will still face limitations in collecting information, in spite of their best efforts and the availability of DSI resources and tools.

Broader conversations to address existing differences in medicolegal death investigation systems across jurisdictions (e.g., centralized state medical examiner, county-based coroner or medical examiner, hybrid), burden of proof for suicide MOD determinations, and policies to allow investigators access to crucial information sources are needed. The information that could be held by these sources (e.g., digital media accounts) could enhance or serve as proxies for psychological and social media autopsies (25,26), and help determine decedents’ state of mind and intention leading up to the time of death.

Challenges facing the widespread adoption of such DSI information collection tools remain, such as funding availability across jurisdictions, the cost of acquiring the tools and training staff to use them, IT system requirements and interoperability, and adaptability to differences in jurisdictional requirements for overdose and suicide MOD determinations. Until these issues are satisfactorily addressed, the widespread adoption of any DSI tool—and its public health prevention impact on overdose and other suicide MOD determinations—will be limited.

Conclusion

Having these series of meetings with diverse DSI investigators and subject matter experts with different levels of experience and from diverse jurisdictions with different burdens of undetermined MOD allowed for rich discussions on what a hypothetical mobile DSI information collection tool could look like. While their perspectives may not be representative of all DSI investigators in their jurisdiction nor the U.S., their insight can serve as a starting point for additional studies, tool development, and the testing, evaluation, and widespread implementation of tools that could improve standardized data collection across jurisdictions. Such mobile DSI information collection tools could potentially help produce more accurate estimates of suicide across jurisdictions and the U.S. overall. In turn, this improved understanding could inform public health prevention and intervention efforts.

AUTHORS

Melissa C. Mercado, PhD, National Center for Injury Prevention and Control, Centers for Disease Control and Prevention, Division of Violence Prevention

Roles: A, B, C, D, E, 1, 4

Deborah M. Stone, ScD, National Center for Injury Prevention and Control, Centers for Disease Control and Prevention, Division of Violence Prevention

Roles: A, B, C, D, E, 2, 3

Caroline W. Kokubun, MPH, National Center for Injury Prevention and Control, Centers for Disease Control and Prevention, Division of Violence Prevention, Association of Schools and Programs of Public Health/CDC Fellow

Roles: C, D, E, 5, 6

Aimée-Rika T. Trudeau, MPH, National Center for Injury Prevention and Control, Centers for Disease Control and Prevention, Division of Violence Prevention

Roles: A, C, D, E, 5, 6

Elizabeth Gaylor, MPH, National Center for Injury Prevention and Control, Centers for Disease Control and Prevention, Division of Violence Prevention

Roles: A, C, D, E, 6

Kristin M. Holland, PhD, National Center for Injury Prevention and Control, Centers for Disease Control and Prevention, Division of Violence Prevention

Roles: A, B, C, D, E, 6

Brad N. Bartholow, PhD, National Center for Injury Prevention and Control, Centers for Disease Control and Prevention, Division of Violence Prevention

Roles: A, B, C, D, E, 3, 4, 6