How AI is changing the face of trauma care

Sign up for the Unity Health Toronto newsletter, a monthly update on the latest news, stories, patient voices and research emailed directly to subscribers. If you haven’t subscribed yet, you can do that by clicking here.

As a Level 1 Trauma Centre, St. Michael’s Hospital receives patients from all over Ontario who have sustained sudden, severe injuries – such as injures from a car collision or a sporting accident. When a trauma patient arrives to the hospital, the trauma team springs into action to provide life-saving care.

Over the past year, the Trauma, Emergency Department, Neurosurgery and Medical Imaging teams have partnered with Unity Health’s applied AI team to develop and use a number of different AI tools to improve care and outcomes for trauma patients.

Predicting when surgery is needed for a brain injury

Traumatic brain injuries (TBIs) happen when a sudden strike to the head damages the brain. TBIs are the leading cause of death and disability in young people globally and 22 per cent of all injury-related deaths in Canada are the result of a TBI.

The type of treatment a patient will receive for a TBI depends on the severity of their injury – some patients may need short-term observation, while others may require an urgent, life-saving brain surgery. Neurosurgeons provide consultation on TBIs to provide an expert recommendation on treatment. At St. Michael’s Hospital, neurosurgeons provide more consultations than all other specialty services combined.

“Every patient with a TBI is triaged in the same way, even though only 20 per cent of patients will require surgery,” says Dr. Christopher Witiw, a neurosurgeon at St. Michael’s Hospital. “Whether it’s the middle of the night, the middle of the day or in the middle of another surgery, the neurosurgeon on-call needs to review the head CT scans and provide a consultation.”

ASIST-TBI is the name of an AI tool that helps to speed up triage for TBI patients by screening the CT scans for patients who come to the St. Michael’s Emergency Department with a head injury. If the model predicts the need for surgery – which it can do just as accurately as a neurosurgeon – the ED physician will consult a neurosurgeon directly, instead of waiting for the radiologist to review the CT scans first.

ASIST-TBI will also be tested and launched to the St. Joseph’s Emergency Department to help even more patients.

Streamlining care for brain bleeds

An intracranial hemorrhage (ICH), more commonly known as a brain bleed, can happen from an external cause, such as a TBI, or from an internal cause, such as a stroke. Brain bleeds can cause permanent brain damage and even death. Fast diagnosis and treatment is critical to recovery.

The brain bleed detection tool predicts whether a patient has a brain bleed. Neuroradiologists typically need to review 25 to 50 head CT and MRI scans a day, and this tool helps to streamline the process.

“Implementing the brain bleed detection tool has enhanced our ability to quickly and accurately diagnose intracranial hemorrhage,” says Dr. Errol Colak, a radiologist at St. Michael’s Hospital. “This technology empowers our radiologists to focus on the most critical cases first with the aim of improving patient outcomes and saving lives.”

The AI solution is already being used in the St. Michael’s Emergency Department and correctly identifies more than 85 per cent of cases where the patient has a brain bleed.

Personalizing trauma resuscitation in real-time

Trauma resuscitation is the act of reviving someone from a life-threatening injury. It is extremely fast-paced and task-oriented. Trauma teams rarely have time to review bloodwork results or consider how a patient is trending so they can anticipate their future needs. Trauma AI is the name for a group of tools that will be used to support the resuscitation of severely injured patients in the St. Michael’s trauma bay.

Trauma AI will use data from monitors and devices that measure a patient’s physical state, such as their heart rate, blood pressure and oxygen levels. It will also use data from the electronic medical record, such as how the patient was injured, the patient’s demographics and – importantly – bloodwork and lab results. Blending these different types of data, Trauma AI aims to make predictions and recommendations that will help the trauma team to personalize their care approach in real-time.

“Trauma AI will be like having a hematologist in your corner, helping to interpret some of the patient’s blood tests,” says Dr. Brodie Nolan, a trauma team leader and scientist at St. Michael’s who is one of the clinical leads on this AI project. “It will also have a predictive trajectory for patients who are becoming unstable so we can mobilize additional resources in the hospital, like getting the operating room ready.”

Hematologists are doctors who specialize in blood and bleeding. The other clinical lead on the project is Dr. Michelle Sholzberg, a hematologist and scientist at St. Michael’s and a leading expert in her field. Getting bleeding under control is one of the central ways to improve outcomes for trauma patients. Care interventions for bleeding are an area of trauma care where there is a lot of opportunity to improve and personalize care.

“The more we can tease out really individualized treatments for patients is probably the next step for where trauma resuscitation is going,” says Nolan.

Trauma AI recently received a grant from the New Frontiers in Research Fund and planning is underway to make these tools a reality over the next two years.

By: Robyn Cox