How AI played a role in saving a patient's life
A 50-year-old resident of Nahariya arrived at the "Galilee" Medical Center for a CT scan with complaints of headache – the patient underwent the examination and was sent home to rest.
Results usually take a week or two, that's the standard for such an examination. But if something really serious happened, the results could be fatal.
But this time, software integrated with artificial intelligence (Viz.ai – in addition to CT, can also analyze echocardiograms and perform other types of scans) analyzed the data and alerted doctors to a possible likelihood that the patient was suffering from intracranial bleeding and needed immediate help.
The doctors immediately sent a vehicle to the patient and operated on him. During the operation, the doctors confirmed that the AI had correctly identified how serious the situation was. The patient's life was saved thanks to rapid diagnosis that shortened the time between the scan and treatment by two weeks. Dr. Don Pass from "Galilee Medical Center" said: "Without this, the patient would probably have come to us too late".
This is not an isolated case, and the next example is "Dr ChatGPT". A mother spent three years trying to find a solution to her son's constant pain symptoms. After unsuccessful consultations with 17 doctors during this period, she finally turned to ChatGPT for help. The frustrated mother created an account and shared with the AI everything she knew about her son's symptoms and all the information she managed to obtain from his MRI.
"We visited many, many doctors. At some point we even reached the emergency room. I continued to search for information persistently", she says. "I really spent a night (at the computer)... I went through all this information".
Therefore, when ChatGPT suggested a diagnosis of tethered cord syndrome (a neurological condition in which the spinal cord is abnormally attached to the lower part of the spinal canal, which limits its movement), "it made a lot of sense", she recalls.
- No matter how much medical technologies have improved, timely diagnosis and treatment are still challenging. Without AI, it's difficult to significantly accelerate diagnosis and treatment
- AI is gradually being integrated into healthcare. Yes, it's still an evolving technology and there are skeptics, but delaying implementation may prevent patients from receiving faster and more effective help.
Karim Karti, CEO of RapidAI which uses AI for rapid diagnosis of stroke cases, aneurysms, pulmonary embolism and other medical problems, described the role of AI in saving lives in medical emergencies in an interview with Healthcare IT News.
Interview highlights:
- Dr. Greg Albers, founder of RapidAI, and his colleague Dr. Roland Bammer developed fully automated AI-based image processing software for CT and MRI. Such software can quickly and accurately create brain and heart images, which can contribute to quick decisions and appropriate treatment.
- AI can help doctors quickly triage patients who need emergency treatment through rapid and accurate assessments, so that the right team can be assembled and the right treatment determined. For example, when a patient has stroke symptoms, AI can leverage the process of creating quality scan images and deliver images to doctors' or paramedics' mobile devices, send alerts to relevant medical teams, and pull patient data and medical history. AI can perform the entire workflow in just a few seconds.
- An important use of AI is saving time in stroke cases. When a person's brain begins to lose brain tissue because it's not receiving oxygen, delay can lead to severe outcomes like death or paralysis. Artificial intelligence can save precious time, allowing the medical team to provide necessary treatment in the shortest time.
- Doctors can miss problems in brain aneurysm scans on CT due to the irregular shape of aneurysms. AI can provide more accurate and detailed images, enabling more efficient and faster decisions.
Our company actively works with the medical field and specializes in developing and implementing medical information systems. We're ready to help you create medical information systems, set up integrations with insurance companies, create databases and other tasks. See one of our case studies on outsourcing experts for a medical portal
What prevents broader AI implementation?
Despite the proven benefits of AI, its implementation in healthcare is lagging due to certain factors.
According to Sachin Patel, CEO of Apixio which provides AI-based health analytics data, "the most prominent obstacle that hinders AI implementation in highly regulated industries like healthcare is the highest level of accuracy required to achieve implementation".
As an example, in the case of a streaming service – "if their algorithms aren't accurate, in the worst case the customer won't like the recommended movie or show. Although this can be inconvenient or even annoying, it doesn't endanger lives".
"But if AI in healthcare isn't accurate enough, it could lead to an ineffective treatment plan, which endangers the patient's health".
There are also trust issues: the idea of software providing information and images of the patient's body, especially in emergency cases, is still difficult to accept. Patient lives are at stake, and AI still doesn't have enough trust, at least not everywhere.
According to David Frieda (VP of Strategic Partnerships at DrOwl) "it can also be difficult to trust a machine to make predictions – this is a common problem in the healthcare industry. Doctors are used to relying on human relationships to determine the best course of action, which includes assessing the patient's mental state, expectations, medical history and more".
Artificial intelligence is still a relatively new phenomenon, although it's developing rapidly, especially in the healthcare industry, because it enables saving lives. AI has already changed the situation in the field, but it will take time to prove its authority everywhere in the medical community.
But ultimately, the need for better preparation and providing emergency care will lead to AI. The changes won't be smooth and consistent. And if it tips the scales toward the Hippocratic oath "do no harm" then probably over time it will become part of the standard medical equipment.
