Firearm-related morbidity and mortality by injury intent: Analysis of medical, criminal, and vital records in Seattle, WA

Brianna Mills | 2017

Advisor: Ali Rowhani-Rahbar



Due to the reliance on administrative records in firearm injury research, individual studies do not present an integrated picture of firearm injuries. Non-fatal and fatal injuries are frequently studied separately, as are injuries of different intents. Certain risk factors for assault-related firearm injuries, like characteristics of the neighborhoods where they occur, the injured person’s involvement with criminal activity, or their mental health may also indicate an increased risk of self-inflicted, unintentional, or legal intervention firearm injury. However, identifying differences or similarities between intent-specific firearm injury groups requires the inclusion of all intent-specific firearm injuries, fatal and non-fatal, in a single study, using all available data, to present a more integrated picture of firearm injury. Clarifying the specific associations between firearm injury risk factors and assault-related, self-inflicted, unintentional, and legal intervention firearm injuries can inform targeted interventions.


This study clarifies risk markers for intent-specific firearm injury both at the neighborhood level and the individual level. To examine relations between neighborhood-level characteristics and rate/density of intent-specific firearm injury, we conducted an ecological study of firearm injuries in Seattle from 2005-2014, geocoded by residence (for injury rates) and occurrence locations (for injury densities). Using the Seattle Neighborhoods and Crime Survey, counts of intent-specific firearm injuries per census tract were modeled as a function of “code of the streets,” fear of crime, handgun ownership, and negative opinions of police, adjusted for neighborhood disadvantage. To examine relations between arrest history and risk of intent-specific firearm injury, we used a case-control study to compare 10-year arrest histories of individuals with intent-specific firearm injuries to a control group of motor vehicle passengers with unintentional injuries in Seattle. All initial injuries occurring 2005-2014 among adults (21+ years) were included. To examine relations between arrest, substance use disorder, and mental disorder histories and risk of intent-specific firearm injury we used a case-control study to compare two-year diagnosis and arrest histories of intent-specific firearm injury patients (cases) to unintentionally injured motor vehicle collision (MVC) passengers (controls). We included all initial fatal and non-fatal firearm and MVC injuries occurring 2010-2014 in Seattle or to Seattle residents, limited to individuals ages 13+. For both case-control studies hospital and death records were probabilistically linked to statewide arrest and (for the second case-control study) hospitalization records. Multinomial logistic regression was used to compare odds of each arrest or diagnosis-related exposure among intent-specific firearm injury cases relative to controls, adjusting for age, race, and gender.


Based on 1368 firearm injuries occurring in Seattle 2005-2014, we found specific neighborhood characteristics were associated with the rate/density of different intent-specific firearm injury categories, including associations with neighborhood scores on “code of the streets,” fear of crime, handgun ownership, and negative opinions of the police. In examining the 10-year arrest history of 1211 adult firearm injury cases and 522 controls, we found individuals with assault-related or legal intervention firearm injuries were more likely than controls to have been arrested for a variety of crimes, including drug-related felonies and violent crimes. Individuals with self-inflicted firearm injury were less likely than controls to have a history of felony arrests but were more likely to have an arrest history involving firearms. We found no significant differences in 10-year arrest history between individuals with unintentional firearm injury and controls. In examining the 2-year arrest, substance use disorder, and mental disorder diagnoses of 763 firearm injury cases and 335 controls ages 13+, we found no significant differences between unintentional and self-inflicted firearm injury cases and controls. Assault-related firearm injury cases were more likely than controls to have a prior felony record, regardless of prior substance use or mental disorder diagnoses. Legal intervention firearm injury cases were more likely than controls to have had a prior felony arrest as well as a diagnosis involving alcohol use, cannabis use, depression, psychosis, or conduct disorder.


Intent-specific firearm injuries display distinct patterns of association with neighborhood characteristics as well as individual-level arrest, substance use disorder, and mental disorder histories. Associations with neighborhood characteristics vary when examining place of occurrence versus place of residence, suggesting place-based interventions should consider place of residence and place of occurrence separately. Subjects with different intent-specific firearm injuries have distinct patterns of prior substance use, mental disorder, and arrest. Many of these injuries occurred after a series of prior encounters with law enforcement and medical professionals, suggesting these prior encounters are underutilized settings for firearm injury prevention programs.