Q&A: In-vehicle ECG monitoring could prevent traffic accidents caused by CV events
Click Here to Manage Email Alerts
In-vehicle cardiac monitoring for older drivers may reduce their risk for heart-related traffic accidents, according to a researcher.
In collaboration with the University of Michigan hospital system, researchers at the Toyota automotive company are assessing the feasibility of in-vehicle ECG monitoring to predict adverse CV events with the goal of improving vehicle safety and reducing the risk for traffic accidents.
Healio spoke with Pujitha Gunaratne, PhD, principal scientist at the Toyota Collaborative Safety Research Center in Ann Arbor, Michigan, about the development of this technology, how it works and when it may be made available to the public.
Question: How likely is a cardiac anomaly while driving?
Gunaratne: We looked into several crash analysis databases, most of which are police reported so they have some indication about the driver condition. They are not actual clinical records, but they give some indication based on what patients were reported and any medical treatments they received in relation to a medical condition.
These datasets that we investigated, that were actually reported about 8 years ago, showed that among patients over 65 years of age, 4% reported causes of traffic accident as medical conditions and a large portion of those from heart-related conditions.
We also looked at the population as a whole. Based on a report from the U.S. Census Bureau, published in 2018, by the year 2050 a quarter of the U.S. population will be over 65 years old.
Heart diseases are very prevalent with that segment of the population. Of the individuals aged 60 to 79 years, 70% to 71% are reported to have CVD. Among those 80 years and older, that percentage increases to 83% to 87%.
What we have right now are mainly police-reported crashes, but there are a lot of other events like near misses and unreported crashes and we see that CVD is very prevalent in that population. That’s where our motivation came from to conduct this research.
Q: How was this in-vehicle cardiac monitoring model developed and how does it work?
Gunaratne: It in not yet developed to an applicable, implementable system right now, but we can see a hope of a light at this point.
In general, ECG monitoring outside of a clinic is a huge problem. Ambulatory monitoring inside the clinic is the ideal condition because you have less interference from other devices, electromagnetic disturbances or motion artifacts.
We wanted to begin by investigating the feasibility of cardiac monitoring inside a vehicle because there are many unknown parameters. We started developing models using the clinical databases from the University of Michigan hospital system in addition to publicly available medical data.
To monitor the ECG inside of vehicle, we selected a particular segment of that patient population who were discharged from the university hospital and provided with an FDA-approved wearable ECG patch. The patch had a GPS and markers so that we can trace where and when they were driving. We were convinced that we would at least detect a decent signal, but it was not necessarily as good as a clinical signal. However, it poses us the challenge of developing more sophisticated algorithms that are more robust to noise.
Q: In its current state, the monitor is a wearable patch, but are there plans to integrate this technology with the interior of the vehicle?
Gunaratne: At this point, we have not decided on one particular modality. We are investigating different ways of monitoring the ECG inside a vehicle such as the steering wheel or any other places a driver may touch. It may be entirely non-touch based. These are very futuristic ideas, and we are actively exploring all avenues.
Q: When the monitor detects an anomaly, will it interfere in the action of driving?
Gunaratne: The first stage is to inform the driver and make them aware an anomaly is happening before an accident occurs. The intention is to predict the events a few minutes before they happen. We would like to allow enough time to inform the driver and let them take action.
We are considered informing first responders at the same time. In fact, one of the platforms we launched in Japan in 2018 has this type of call center contact mechanism for when an emergency situation arises inside a vehicle. Those modalities are being considered.
Q: What will prevent the device from sending false alarms in situations such as sudden stops or road rage, when a driver’s heart rate might suddenly and unexpectedly change?
Gunaratne: We have already seen those types of interferences, so to speak. There are effects on the ECG when you have an emotional episode or when you are drinking coffee, and these things can interfere with your heart rate. But what we have seen from critical heart conditions, especially when looking at the electrical impulses of the heart, conditions such as arrhythmias show a very distinct type of signal. When you have these conditions, the signals tend to have a different pattern. In atrial fibrillation, you will have a repeat heartbeat and it is quite different from what you would see in an episode of anger, road rage or coffee drinking.
Using these distinct features, we can train the algorithm to eliminate these types of noises, differentiating them from an actual heart condition. This involves a huge machine learning component.
Q: Will this technology be made available to all models?
Gunaratne: I cannot speak to how it will actually roll into the fleet, but I can say that any reliable safety components will crawl into the global Toyota safety platform. For example, for the Toyota safety sense system we have already integrated advanced driver assistance components. This type of research, detecting heart conditions inside the vehicle, is very early stage. But, when it is developed to a more robust, implementable level, we will use similar models of implementation.
Q: How soon can something like this be rolled out?
Gunaratne: It's hard to predict such a large-scale deployment, but based on what I see from the current state of the evolving consumer electronic market, we will likely see something within the next 5 years.
Q: Will in-vehicle cardiac monitoring technology require FDA approval?
Gunaratne: I assume there will be some regulations that we come across. It all depends on what the final product is, how it will be presented and how the user will use it, whether it is for monitoring or as some diagnostic guide.
We are actively engaged with organizations, even during our research, to publish our work, get expert opinions and help navigate our research to meet regulatory and other requirements that we might need in the future.
Q: Is there anything else you would like to mention?
Gunaratne: Physiological monitoring inside a vehicle will become more relevant and important in the future because of the growing older population. We are seeing a trend that, after retirement, people want to be mobilized. We want to provide the mobility and we cannot exclude people with some conditions or for having health issues. This type of physiological sensing may become a requirement in a number of products in the future.