Florida death certificate

ABBREVIATION. ME, medical examiner.

Firearm injury is the most lethal of all types of Violence (1) and the second most common cause of death for young people. (2) Prevention of pediatric gun death requires a clear understanding of violent injury--homicide, suicide, and unintended gun death--and its preventable causes and consequences. Epidemiology provides the foundation for that understanding and all subsequent injury prevention planning and policy initiatives. (3,4)

A key epidemiologic source for fatal firearm injuries is medical examiner (ME) reports, completed for each firearm-related death. In Florida, the manner of death reported by the Office of Vital Statistics on death certificates is determined by the ME except in cases in which there is a pending police investigation. MFs determine and code the cause, mechanism, and manner of death. In the case of firearm fatalities, manner of death is typically coded by the ME as homicide, suicide, or accident. (Note: "Accident/accidental" are terms generally used by the ME and in this article refers to such ME manner of death classification.)

Manner of death determination is a subjective conclusion based on facts concerning the circumstances surrounding the death, the autopsy findings, and laboratory tests. (5) Previous studies have documented notable differences of opinion among MEs regarding manner of death classification. (6-9) A 28-year review of pediatric firearm fatalities revealed that individual MEs vary considerably in their classification of accidental manner of death. (6) A review of adolescent self-inflicted deaths suggested racial and age-related bias in manner of death determination. (9) ME and coroner criteria for manner of death also differ by jurisdiction and death investigation practices. (10) Thus, individual subjectivity and regional variation in manner of death classification may affect the reporting of unintentional pediatric firearm fatality incidence.

The definition of manner of death classification accounts for intent, wherein homicide and suicide are deliberate and unintentional death is not, regardless of evidence of negligence. (11) However, in practice, intent is not uniformly considered in manner of death coding of firearm fatalities in Miami-Dade County (Joseph Davis, MD, Miami-Dade County ME, personal communication, November 30, 2000). Barber et al (12) also found differences between ME manner of death coding policy and practice in Massachusetts.

Unintentional firearm death is estimated to be 2.82% of all firearm deaths. (13) This translates to 824 unintended gun deaths annually. This figure may lead clinicians to discount the importance of prevention counseling and minimize recommendations to parents regarding risk of unintentional injury, yet, as with many unintentional injuries, children and adolescents are disproportionately involved: unintentional gun death among 0- to 19-year-olds is 2.5 times higher than for the general population. (13) The aim of this study was to compare the proportion of pediatric unintentional gun death in Miami-Dade County based on manner of death coding by the ME with an intent-based classification approach.

METHODS

A retrospective study reviewed ME and police department records for all pediatric (0-17 years of age, inclusive) firearm fatalities in Miami-Dade County from January 1, 1994, to December 31, 1998. The Miami-Dade County Medical Examiner's Office provided summary data sheets of all pediatric deaths by firearms for the 5-year period specified. Victim identifiers including name, date of birth, date of death, and ME case number were used to locate the corresponding police records. All available material in both ME and police records was reviewed between May and August 1999. ME and police records were reviewed at the ME office and police departments, respectively. In addition, death certificates were requested from the Florida Office of Vital Statistics for all study subjects.

The ME and police departments record information that includes medical history, autopsy reports, police reports, toxicology screens, summaries of detectives' and MEs' communications with family, documentation of previous Department of Children and Families investigations, suicide notes, photographs, and press clippings. The records reviewed included cases investigated by several different MEs and investigators employed by the Miami-Dade County Medical Examiner's Office and detectives from 3 different police departments within Miami-Dade County (Miami-Dade County, City of Miami and Hialeah Police Departments). Extracted variables included demographics of victims and suspects; details of the incident circumstances, including the preceding and subsequent historical accounts; the victim's medical and mental health well-being; and social history information, including drug use or gang involvement by the victim or shooter.

The investigators used an intent-based classification of the deaths according to whether intent to harm or intent to threaten harm was expressed or implied. (14) A threat to harm would include a situation in which a gun was used by a perpetrator to intimidate, such as during an argument, even if the actual firing of the weapon was unintended. Such a death would be considered intentional, in that the weapon was used to threaten harm.

Unintentional deaths were determined to occur when the person who fired the gun did not intend harm or intend to threaten anyone with a firearm, and included deaths that occurred when the shooter believed that the gun was unloaded, was engaged in cleaning the gun, was in play, or otherwise demonstrated curiosity or bravado. Russian roulette deaths were excluded from the analysis, as determination of intent in such fatalities remains controversial.

The first 3 authors conducted the record review and independently classified the deaths as intentional or unintentional. Classification of intent required unanimity among the 3 authors. When there was a discrepancy of opinion between the reviewers, full records were independently reviewed again. When disagreement still remained, the investigating police department would be queried regarding determination of intent, and this was to be accepted as final. Interclassifier reliability was high: the 3 reviewers agreed unanimously on 97.5% of the cases on the first review. Agreement was reached on the remaining 3 cases when the complete records were reexamined.

All deaths coded as unintentional, intentional homicide, and intentional suicide by the investigators were compared with those coded as accidents, homicides, and suicide by the ME. For comparing these proportions, the [chi square] test of hypothesized proportions was used.

RESULTS

During the 5-year period studied, the ME examined 124 firearm deaths of people younger than 18 years. One case was excluded as the deceased died in a foreign country without ME resources and was sent to Miami for an autopsy. The remaining 123 deaths all occurred in Miami-Dade County. Death certificate manner of death coding corresponded in all cases with ME coding for the same.

The mean age of pediatric firearm deaths was 13.9 ([+ or -] 3.9), and the median age was 16. Seventy-one percent of the deaths occurred among children between 14 and 17 years of age, 21% occurred in children between the ages of 6 and 13 years, and 8% occurred in children 5 years of age or younger. Twenty-three percent of the victims were female. More than half (52.8%) of pediatric gun deaths occurred in black children, and 35% of the deaths occurred in children with Spanish surnames (includes both black and white). This is a proxy variable for Hispanic ethnicity, which is not required and only occasionally recorded by the ME.

Among the 123 pediatric gun deaths enrolled, a total of 22 cases previously classified by the ME as homicide (n = 16) or suicide (n = 6) were reclassified as "unintentional." In those cases, there was no evidence of intent to harm or to threaten harm. A significant difference between the ME's coding and our intent-based classification was found for homicide (94 vs 78; P = .037) but not for suicide (25 vs 18; P = .175). A significant difference was found between the ME's category of accident and our reclassified "unintentional" deaths (4 vs 26; P < .0001).

The ME had coded 4 deaths as accidental, and our review concurred that they were unintentional; these cases were not different in circumstance from the reclassified cases (Table 1). In each of the 4 cases, manner of death was coded by a different ME. Two cases (1.6%) of Russian roulette were placed in their own category (Table 2).