Chest pain is a common problem with dozens of causes that range from harmless bruised muscles from coughing to potentially fatal pulmonary embolisms. But for the person involved, the experience can be a frightening one – even when the cause turns out to be relatively minor.
Now an Israeli startup has developed an AI-based platform to help doctors diagnose patients with chest issues, accurately and in real time.
Quai.MD seamlessly connects to a hospital’s Electronic Health Records (EHR) – the digital version of a patient’s medical history – and uses this data, along with triage assessments and expert opinions drawn from medical research, to propose the most likely diagnosis and best courses of action.
For although there are established medical guidelines to help determine and treat multiple causes of chest pain of varying degrees, ER physicians, who are often understaffed and under tight time constraints, can make mistakes in diagnoses.
A December 2022 study by the US Health Department found that more than five percent of patients experience misdiagnosis in American emergency departments.
But Shlomi Uziel, co-founder and CEO of Quai.MD, tells NoCamels that the platform gives doctors a clear set of steps to follow during the examination period.
“The application sits inside the EHR, and helps the physician determine what the next step is for each of those potential diagnoses until they reach the decision to ultimately admit or discharge the patient,” he says.
“What we’re trying to do is help [physicians] align more with the best practices and protocols.”
Uziel explains that every hospital has a list of the best practices that derive from general medical knowledge and research. And the failure of an ER physician to follow protocol, he says, can lead to one of two outcomes:
In the first instance, the physician suspects that the patient has a serious illness and admits them to the hospital, where they can potentially spend several days undergoing various tests only to discover that there is nothing wrong.
This phenomenon costs the US economy an annual $750 billion, according to private healthcare firm PinnacleCare.
The other situation, says Uziel, is that the patient is sent home from the emergency room without realizing that they have a potentially fatal ailment, which could have disastrous results.
Deep Learning Diagnosis
Quai.MD accesses and analyzes the patient’s initial ER assessment and their medical history, even as they are being seen by a physician.
The AI platform then generates several diagnoses, ranging from most likely to least likely, as well as the steps the doctor should take in order to eliminate each possibility.
Uziel gives the example of acute coronary syndrome – a range of conditions related to sudden, reduced blood flow to the heart, including a heart attack. In order to rule this out, he says, the doctor must order a blood test to check for troponin, a protein that is released into the bloodstream during a heart attack.
The Quai.MD platform, he says, can be used to order specific blood tests with just one click.
Sign up for our free weekly newsletterSubscribe
Should the doctor decide to admit a patient, Quai.MD would then generate a report of all the care the complete process of patient care – saving the doctor time and making it easier for the medical billing team to determine the patient’s insurance coverage for the services they received.
Quai.MD is currently focused on diagnosing the various causes of chest pain, which Uziel says has around 60 potential diagnoses – including a handful that are among the riskiest of health conditions.
Beyond possibly saving lives and preventing unnecessary healthcare costs, Quai.MD’s CPO Marcelle Kaspi explains that the technology may also eliminate the biases that some healthcare providers have towards certain groups among their patients.
“We know that there’s a lot of sex-based, race-based, and socioeconomic-based biases in the healthcare system in general,” she says.
Research has shown, for example, that white medical students and residents were more likely to believe that black patients feel less pain and do not need the same levels of pain medication as white patients – even as recently as 2016.
Because Quai.MD’s artificial intelligence learns solely from medical journals and studies, as well as the patient’s medical history and treatments, it doesn’t hold the same implicit biases as doctors. This, Kapsi believes, could potentially solve this issue.
The Ramat Gan-based company expects the platform to become operational in the emergency room at the Medical University of South Carolina at the start of 2024. The startup is currently collaborating with doctors from this hospital and from the Mayo Clinic in Minnesota in order to finish developing the app. It is also due to receive medical records from 10,000 patients to further train the AI.
The Doctor Will See You Now
Quai.MD is not the only Israeli startup trying to improve doctors’ performances. Kahun allows patients to discuss their symptoms with generative AI, and provides their doctor with a summary of their condition and possible diagnoses before their real life consultation. And Navina uses AI to produce summaries of a patient’s medical history for doctors via a smartphone app.
But Uziel says that what sets Quai.MD apart is that it is the only one that automates the entire clinical process, from initial diagnosis through to suggesting care and treatment options.
Quai.MD, which was founded in 2020, has raised around $2 million thus far from venture capital firm Random Forest and seed-stage fund Labs/02. It is now announcing a new $2.5 million financing round led by Good Company, with the participation of two healthcare systems from the US as well as new and existing investors.
Most recently, the startup was one of five finalists in the Asper Prize competition, which recognizes startups using innovative technology to create a global positive impact.
Quai.MD was founded by Uziel, a former VP at multinational computational software company Cadence Design Systems; Prof. Chen Shapira, the former CEO of Carmel Hospital in northern Israel; and Dr. Golan Yona, a machine learning expert and former Cornell Professor at the Department of Computer Science.
“We just wanted to do something that makes this world slightly better,” says Uziel.