WISH Documentation: Hospitalizations Resulting from Nonfatal Motor Vehicle Crashes
This query module provides data for hospitalization stays with motor vehicle (MV) injuries and some of the circumstances of the crash associated with these injuries. This provides an opportunity to look at how these injuries are associated with risk and protective behaviors related to motor vehicle traffic safety. Hospitalization data are linked with crash report data.
Datasets and Data Linkage
The injury dataset for this match is based on the Injury Hospitalization subset used for all other injury queries (see injury technical notes). In short, this subset is limited to stays with a principal diagnosis of injury (any injury, initial encounters only). Furthermore, for this query, the subset of data is restricted to stays that note an external cause of morbidity related to a motor vehicle traffic injury.
The MV injury subset is then linked with crash report data. The Wisconsin Department of Transportation (DOT) has developed a crash report data system with reports entered by law enforcement officers. These reports contain information on the crash, vehicle, all persons affected/involved, and other circumstances surrounding the crash.
The link of hospitalization and crash data was performed using deterministic methodology. Currently, records were considered an exact match if the first and last name initials, date of birth (DOB), sex and zip code of residency matched. Other combinations were used to match any remaining records by the following criteria:
- First and last name initials, DOB, sex and the first three numbers of residency zip code;
- First and last name initials, DOB, and sex; or
- DOB, sex and first three numbers of residency zip code.
Additionally, matches were limited to hospitalizations within six days after the crash date (only matches within the same calendar year included; for instance, date of crash, admission date and discharge date all within the same calendar year).
Limitations of Linkage
If a hospitalization with an MV injury code does not match with a crash report (or visa versa), these stays will not be included in the query. There are several reasons why not all hospitalizations with an MV injury will match with a crash record. The DOT database captures only crashes occurring in Wisconsin. This means that if a Wisconsin resident was involved in a crash in another state, that crash record will not be part of the WI DOT database. Additionally, motor vehicle traffic crash reports are limited to events causing injury or significant property damage that involve motorized vehicles on public roadways. Such events include crashes of one or more vehicles and also instances of vehicles impacting pedestrians, pedal cyclists or stationary objects. The DOT reports do not include incidents occurring on private roads, off-road trails, sidewalks, paths, driveways, parking lots, and similar non-public roadways.
Finally, the linkage is performed on health visits with matching admission and discharge years that link with a crash occurring no more than six days prior in that same calendar year. For hospitalizations, we estimate fewer than 2% of nonfatal stays with an MV injury were excluded because year of crash and year of discharge from hospital were not the same.
Due to the reasons just listed, the match is not a perfect one. Additionally, there is no way to be certain that the hospitalization noting an MV injury is directly linked to the specific crash report with which it is matched. Because the timing of the hospitalization and crash are close and the hospitalization includes a code for MV traffic as the cause, these records are associated but it is possible that some of these stays are not the result of the crash.
Based on the details provided above (non-fatal stays with crash and hospitalization within the same year), for the period 2017-2019, there were 8,200 nonfatal hospitalizations with an MV injury noted in the record; 6,360 (approximately 78%) matched with a DOT crash record. Approximately 1% (83) of stays with an MV injury noted in the record were inadvertently excluded from the match. Of these 83, it is likely that not all of these would have matched with a crash record.
In Step 3 of the query, the data user can select one of three major injury types: Traumatic Brain Injury (TBI) of the internal organ (i.e., brain), spine or back injury excluding contusion or superficial wound, and abdomen injury to an internal organ. The specific codes for body region and nature of injury can be found at the Center for Disease Control and Prevention's National Center for Health Statistics Tools and Frameworks. Scroll to Injury Diagnosis Matrices. These injury types were selected as they are often more likely to result in longer recovery time, greater cost, and more significant impact on quality of life. Furthermore, they are noted in more than 50% of hospitalizations with an MV injury. This Step does not include all injury types.
Injury Severity Coding
Severity is based on an Abbreviated Injury Scale (AIS) score which is an anatomically based injury severity scoring system that classifies each injury by body region on a six-point scale. The scoring system was developed by the Association for Advancement of the Automotive Medicine (AAAM). The software TraumaBase, V9 by EOS was used to attribute an AIS code to each ICD-10-CM injury code (though not all codes have a corresponding AIS score; for instance, under-dosing and adverse or toxic effects of drugs are not considered trauma codes due to the type of medical services they receive–i.e., internal medicine/critical care services instead of surgery services). If there were no diagnosis codes with an AIS score for any single visit, the visit was categorized as mild to moderate severity (less than half a percent—0.5%–of all visits from 2017-2019 had no AIS score). The highest AIS score is selected to represent the overall severity. We have grouped the six AIS categories into the following two groups:
- Minor to Moderate (AIS score of 1, 2 or missing)
- Serious to Maximal (AIS score of 3 through 6 representing serious, severe, critical and maximal)
Personal Safety Equipment
Type of safety equipment is related to the type of MV injury (for instance, motorcycle helmet use only applies to Step 1 selection of Motorcycle). Use of safety equipment is recorded in the DOT crash record. The values represent personal use of the relevant safety equipment or absence of use (this could be “no” or it could be missing/unrecorded information at time of crash). Approximately 1% was unknown for MV occupant use of seatbelt or child-seat; 1% unknown for pedal cyclist helmet use; and 4% unknown for motorcyclist helmet use.
These behaviors are recorded in the DOT crash record. Alcohol and drug use means that law enforcement suspected at least one driver or non-motorist involved in the crash had used alcohol (which includes under the legal limit and at or over the legal limit) or drugs. Use of alcohol or drugs does not necessarily apply to the individual who is receiving health care (i.e., hospitalized) but instead, the behavior was present among someone involved in the crash at the time of the crash. It does not imply that drugs or alcohol were directly responsible for the crash. Absence of alcohol or drugs indicates no involvement or missing/unknown information (not recorded for the crash). Approximately 5–6% of hospitalizations linked with a crash record had missing information on drug and/or alcohol use.
For speed, this could be excessive or inappropriate speed for the situation and indicates that speed was a factor in the crash. This is a crash variable and does not imply that the person hospitalized was responsible for inappropriate speed. Absence of this risky behavior indicates that speed was not a factor or that this information was not recorded for the crash. Approximately 5% of hospitalizations linked with a crash record had missing information on speed.