Medical errors are the third-leading cause of death in the US and they are probably the most common and preventable. MedAware, an Israeli company, is using the power of artificial intelligence to detect potential errors before they happen.
The company was started by Dr. Gidi Stein, MD, PhD, who had spent several years in the tech industry before deciding to use his background in algorithms to make a larger impact. He subsequently obtained his MD from Tel Aviv University Medical School and a PhD in computational biology from Tel Aviv University. He started MedAware in 2012.
MedAware is currently being used at two hospitals in Israel and is in the process of being implemented in four healthcare systems in the US and Israel. In June 2018, the company announced a partnership with Allscripts Healthcare Solutions, a major electronic medical records provider in the US.
To learn more, we asked Dr. Stein a few questions about MedAware.
Cici Zhou, Medgadget: Tell me about the founding of MedAware. What was the inspiration?
Dr. Stein: MedAware was conceptualized following a tragic story: a physician was treating a small boy who had asthma. To treat his asthma, the doctor within his electronic prescribing system meant to click on “Singulair,” a standard asthma medication. Sadly, he misclicked, a mistake that could happen to anyone using a computer, and prescribed the boy “Sintrom,” a blood thinning medication meant to treat blood clots. The physician, the pharmacist, and the parents all missed the error that led to the boy’s death.
I was baffled by this story. How could this happen with all the technology at our disposal today? The mistake that was made was one that any overworked, tired doctor could make. None of us are above it. So, I set out to use my tech and medical backgrounds to make sure to prevent similar mistakes in the future.
Medgadget: How does MedAware use AI to detect potential errors?
Dr. Stein: The AI in MedAware is trained in the same way that it is in other fraud detection industries. The idea is to identify a set pattern that occurs in 99.99% of transactions and then to alert whenever we identify an outlier to that pattern.
MedAware analyzes millions of clinical records, studying what medications were prescribed to which patients, and under which clinical situations. Like in the above example, it identifies “normal” prescription patterns. So, for instance, a small child is likely to receive a prescription to help with an ear infection but will likely not need an anticoagulant. If a doctor accidentally prescribes an outlier medication, MedAware will alert the doctor that their choice may cause harm.
Moreover, when a patient receives his pills from the pharmacy and goes home, he can be exposed to harm and adverse drug events throughout the duration of treatment. MedAware continuously monitors any change in the patient’s clinical records throughout the duration of treatment and assesses whether these changes render one of the active medications suddenly a dangerous outlier to the changed clinical setting.
Medgadget: Say a physician tries to prescribe new medication that will interact with a patient’s current medication. What will happen?
Dr. Stein: MedAware is integrated into current Electronic Health Record (EHR) platforms. When an inappropriate medication is added to the patient’s record, MedAware sends an alert directly to the healthcare provider via the EHR platform. Physicians have learned to trust MedAware warnings, as they are infrequent, very accurate and with high clinical relevance. Moreover, they appear differently than other alerts, and physicians are far more likely to accept them. A recent case study showed that a physician is ten times more likely to accept a notification from MedAware than from current decision support tools.
Medgadget: What are some of the biggest challenges facing the MedAware team?
Dr. Stein: Our primary challenge is getting hospital executives, who are burdened by tight budgets, to focus on patient safety as well. From their perspective, they just invested $500 million in an EHR system and don’t want to add even a small additional component on top of it. However, the added value they receive with MedAware — in both improving patient outcomes and avoiding adverse drug events, their associated costs (in the U.S. alone this cost is over $21 billion every year), and the accompanying lawsuits — pays for the investment. It is our job to convince them of that. Ultimately, although we have a proven ROI of between $5 to $8 saved for every $1 spent, it is our hope that patient safety will always be the priority.
We’re a small Israeli startup trying to enter a massive healthcare market. Sometimes it can be hard to get noticed when you are our size. However, we’ve found that our tech is good enough that the right people are paying attention to us.
Medgadget: Looking to the next five years, what are the biggest goals for MedAware?
Dr. Stein: Firstly, we want to expand from 5 million medical records analyzed to 20 million in the next two years. With every record that our AI analyzes, MedAware improves its already high accuracy, and prevents more types of medication errors.
Secondly, we want to provide medical professionals additional ways to help keep their patients safe. This includes opioid dependency risk assessment, finding gaps in a patient’s care, and projections of a patient’s status based on trends.
Finally, with the growth in consumer-focused mobile health applications, we would like to provide our insights directly to the consumer. This would help circumvent interoperability problems between multiple EHRs in use in different medical institutions.
Link: MedAware website…