Article summary: Firearms bans reduced homicide rates in two Colombian cities
Citation: Villaveces, A, Cummings, P, Espitia, VE, Koepsell, TD, McKnight, B & Kellerman, AL. Effect of a ban on carrying firearms on homicide rates in 2 Colombian cities. JAMA 2000;2808:1205-1209
Between 1983 and 1993, the annual homicide rate in Colombia increased by 366%. Before 1993, Colombians who had permits to own firearms could carry concealed weapons anywhere. In response to escalating firearm violence, Bogotá and Cali city leaders intermittently banned residents from carrying firearms during periods associated with higher homicide rates (weekends after a payday, holidays, and election days) and some other occasions, regardless of whether residents had firearm permits. Fatal injury surveillance systems were established in Bogotá in 1993 and Cali in 1995. A homicide prevention program, DESEPAZ, was launched in Cali in 1993 and operated throughout the study period.
This study used an interrupted time-series design with multiple replications to determine whether the ban on carrying firearms was associated with any changes in homicide incidence in Bogotá and Cali. The data collection occurred in Bogotá from 1993-1994 and in Cali from 1995-1997. Police, coroners, and hospitals reported data on victim demographics, time and place of injury, manner of death (suicide, homicide, or unintentional), and firearm involvement to each city’s fatal injury system. City police enforced the ban by conducting searches at targeted times and locations. The authors used indirect standardization to conduct a stratified analysis. Stratification categories were weekday, payday weekend, other weekend, quarter of year, time of week, and type of homicide (firearm-related or non-firearm-related). The authors controlled for variation by time of day and month, special events, and victim age and sex, as well as the possibility an individual who planned to commit firearm homicide waited to carry a firearm until the ban was no longer in effect. Potential serial correlation among homicides was investigated.
On average, 82.1% of Cali homicides and 90.7% of Bogotá homicides occurred on non-intervention days. After adjustments, the regression model found homicide incidence in both Cali (rate ratio, 0.86, 95% CI, 0.76-0.97) and Bogotá (rate ratio, 0.87, 95% CI, 0.76-0.99) was significantly lower on days when the ban was enforced than during nonintervention periods. There was no evidence age, sex, or serial correlation were significantly associated with homicide incidence during the intervention periods and seven-day periods after each intervention. The effect of the intervention was not significantly different for firearm and non-firearm homicides.
A major strength of this study is its design. The intermittent application of the intervention made it reduced the likelihood additional factors that influence homicide rates could have selectively reduced incidence when police enforced the ban. The mechanism through which the intervention prevented some firearm and non-firearm homicides is unclear. Data on time of assault were missing for 21.6% of Bogotá deaths, so the authors imputed missing values using an approximate Bayesian bootstrap method.
In two Colombian cities, intermittently banning residents from carrying firearms was associated with significantly lower homicide incidence, compared to periods where the ban was not in effect.