Belt use: comparison of NASS-CDS and police crash reports

This study compares belt use in police reports with NASS-CDS investigator-determined belt use. The NASS-CDS cases were analyzed by severity of occupant injury and the crash type.

1993-2007 NASS-CDS was analyzed for occupant injury severity and crash type. Light vehicles were included with model year 1994+. Injury severity was subdivided by MAIS 0, 1-2, 3, 4+F, and fatal. Crash types were frontal, side, rear, and rollover. The NASS-CDS determination of belt use was assumed the gold standard used to determine miscoding by the police.

The fraction of unbelted occupants increased with the severity of injury from 3.8 percent with no injury to 53.9 percent with fatalities in the police reports. NASS-CDS reported no belt use of 7.9 percent (2.18 times greater than the police) with no injury to 58.2 percent (1.08 times) with fatalities. In side impacts, severely injured occupants were unbelted in 46.9 percent of NASS-CDS cases. This was 1.35 times greater than the 34.7 percent unbelted reported by the police. For severely injured occupants (MAIS 4+F), 18.4 percent of the police reported cases with belt use were actually unbelted occupants by NASS-CDS. The reporting error decreased to 5.0 percent for uninjured occupants (MAIS 0). For uninjured occupants, 35.7 percent of the police-reported cases of no belt use were coded as belted by NASS after inspecting the belt system. This difference decreased to 2.6 percent for fatally injured occupants.

For occupants with serious-to-fatal injury, the error in police reporting belt use was 13-18 percent. The police often rely on self-reported belt use, which overestimates actual belt wearing and they do not always conduct in-depth investigation of vehicle, seat belt, and occupant injury to reach a conclusion of belt use. The police generally overreport belt use in motor vehicle crashes.