Advanced Crash Data Used to Predict Auto Crash Severity

GM, OnStar and Dr. Stewart Wang, director of the University of Michigan’s International Center for Automotive Medicine (ICAM), have completed a research project to help form the next potentially beneficial connection between vehicles with crash sensor technology and the emergency response community. As the first research study to match real-life injury outcomes with crash telemetry data, the project’s preliminary results show the technology could help predict severity of injury.

"This is a groundbreaking research project which may help improve the outcomes for certain crash victims," Dr. Wang says. "Our goal is to make injury severity predictions more accurate. We have been working with GM and OnStar toward an algorithm that could be used more broadly across the industry someday."

When a vehicle crash takes place, several factors influence the possibility of occupants sustaining injuries - as well as the severity of any such injuries. Injury Severity Prediction (ISP) is a feature that may alert OnStar Advisors to the severity of a crash and provide important information that can be relayed to first responders in the vital minutes and seconds that precede their arrival on the scene.

The factors that make up the ISP algorithm to determine the likelihood of a severe injury in a crash include:

  • Delta V (maximum change in velocity)
  • Principal direction of force
  • Seatbelt use
  • Age and gender of vehicle occupants (verbally obtained by advisors)
  • Multiple events (impacts)
  • Vehicle type

These factors are used in conjunction with a computer algorithm to determine the likelihood of a severe injury (the predicted “injury severity”) in the crash, also known as ISP. This instant computer analysis of the crash data determines whether there’s at least a 20 percent chance that a vehicle occupant sustained a severe injury. In validating the algorithm in the recent research project, Dr. Wang and his team have matched each ISP rating with the corresponding police report, medical records, EMS data and CT scan data. All personal information in this study was removed, so injured occupants were kept anonymous.

"If ISP predicts a greater than 20 percent risk of severe injuries, then the PSAPs and First Responders should be informed so that the appropriate transport destination — the type of hospital or trauma center — is chosen," Dr. Wang said.

With this information, OnStar Emergency Advisors can be connected into the vehicle to inquire about the occupants’ condition. If occupants need emergency assistance — or are injured and cannot respond — the OnStar Advisor can immediately contact the appropriate Public Safety Answering Point (PSAP) and provide the vehicle’s exact GPS location along with the important information supplied through Injury Severity Prediction. In the meantime, EMD-certified OnStar Advisors can provide Medical Priority Dispatch System (MPDS) protocol instructions to victims until help arrives.

"Other information that can be very helpful to EMS is if there was more than one vehicle involved or if there were multiple occupants in the vehicle," Dr. Wang said. "Urgent complaints — difficulty breathing, chest pain, witnessed seizures, etc. — that are volunteered are important to relay as well."

OnStar advisors have been verbally providing crash data since 2004. Future plans to adopt the protocol may require:

  • EMS Medical Directors and providers be educated on the usefulness of crash data and
  • Implementation of criteria-based transport policies

Future standards also need to be developed across the telematics industry for the transfer and use of ISP data in the field and among PSAPs using Next Gen 911 technology.

When EMS providers are alerted to the possibility of severe injuries, they can be prepared to manage field triage more effectively and direct the most severely injured parties to the appropriate trauma centers. "For critically injured patients, the mortality rate can be decreased by 25 percent if they are transported to a Level-1 trauma center versus a non-trauma center," Dr. Wang says.

This helps connect the right patients with the right treatment, and also helps first responders apply their resources in the most effective and efficient manner while validating minor crashes that may not require a full EMS response.

"It allows EMS systems to appropriately deploy the proper resources to those who most urgently need them," Dr. Wang says. "It may allow them to respond more safely to lower-risk crashes."

Learn more about Automatic Crash Response by visiting or contact us at

Not all vehicles may transmit all crash data.

OnStar service requires vehicle electrical and OnStar systems (including battery), wireless service and GPS satellite signals to be functional, available and operating for features to function properly. Visit for details and system limitations. OnStar acts as a link to existing emergency service providers. Subscription Service Agreement required.

[ Feedback → ]