|
Abstract
|
Consistent with previous studies, the Article by the GBD 2019 Police Violence US Subnational Collaborators finds that the National Vital Statistics System (NVSS) data under-report approximately half of all violent deaths of civilians that occur following encounters with law enforcement. The decision not to assess the National Violent Death Reporting System (NVDRS) is unfortunate for two reasons. First, the NVDRS is the most promising surveillance database for tracking homicides by police in the USA; second, the NVDRS is excluded on the basis of the misconception that it does not capture homicides by police involving methods other than firearms.
The authors correctly note that the NVDRS reports 10-20% fewer fatalities than do some open-source databases, namely Mapping Police Violence and The Counted. However, they are incorrect in attributing this discrepancy to under-reporting of non-firearm homicides by the NVDRS. Open-source databases report more non-firearm officer-involved deaths than does the NVDRS because, except for firearm deaths, the NVDRS is not designed to capture deaths that were ruled natural or accidental. For example, in a study by our group, 42% of taser-related deaths and 94% of motor vehicle-related deaths reported by The Counted, Fatal Encounters, and Mapping Police Violence were not captured in the NVDRS, probably because they were not ruled homicides after investigation of the cause of death.
Case definitions matter and account for at least some, and perhaps most, of the discrepancy in officer-involved fatalities reported by the NVDRS and open-source data. It is unclear how the authors define their main outcome of interest, deaths due to police violence. In referring to The Counted as the gold standard, the authors seem to favour the case definition of The Counted, which includes motor vehicle-related deaths. The authors are inconsistent because they also state that when the cause of death was listed as vehicle in Fatal Encounters, the death was excluded, explaining that these incidents would be less likely to involve direct violence perpetrated by police.
The decision to focus on the NVSS rather than the NVDRS was a missed opportunity to assess whether the NVDRS is as excellent a source for non-firearm homicides...
Language: en |