Dr. Gidi Stein is the CEO & Co-Founder of MedAware, which utilizes AI-driven outlier detection technology to mitigate medication-related risks and evolving adverse drug events.
He is an award-winning physician, researcher, technologist and an expert in medical informatics. Early in his career, he served as CTO and Chief Architect of several algorithm-rich startup companies in Israel.
Summary
In this episode, Dr. Stein starts off by giving us a context of medical errors and patient safety. He then gives us eye opening examples of how unintentional but yet simple medical errors like an incorrectly prescribed drug, an overdose of a drug or a treatment pathway that doesn’t match the profile of the patient being treated may have catastrophic consequences.
The AI based patient safety platform that Medaware has built helps safeguard patients from these possibilities while not overburdening clinicians with irrelevant alerts during the course of care.
Audience Questions
[41:04.69]: How should the decision rationale of AI used in healthcare applications be checked i.e. should it be vetted by human health care professionals?
Michael Scott Long — A PhD scientist, writer, journalist and editor
[44:02.83]: Just like drug-drug interactions could have adverse effects, drug-nutrient interactions could also have adverse effects. Is there a role that AI can play in detecting and alerting drug-nutrient interactions?
Dr T R Gopalan — Former Dean and Prof. of General Surgery
[45:25.10]: Can the use of AI in spotting, incorrect medications or prescriptions lead doctors to rethink, re-examine the care workflow and overall lead to wider adoption of best practices in the care organization?
Umashankar S — Manager, R & D at HPE
Topical Questions
[4:08.85]: Could you give us a sort of an overall context of what is medical errors? Maybe how frequent is it? How big of a problem is it?
[6:12.06]: In your experience what kind of medical errors have you seen? …. What are the most common ones or perhaps the ones that are most damaging in your experience? What have you seen?
[19:42.36]: If you want to track the life of a patient who came for care and if you want to track the interesting events that happened in their life outside of the hospital setting, how do you do that?
[23:32.85]: So, let’s say this patient then walks in for a particular condition and he’s being treated by a physician. So and the physician obviously takes the history of the patient, makes an assessment and let’s say the physician prescribes a particular drug or a drug combination. So, now what happens or what is happening in the background?
[36:12.74]: I can perhaps think that there must have been some pushback from physicians or clinicians or hospitals or clinics because somewhere this is like, you’re pointing out where the errors are, right.
Quotable Quotes
[7:16.66]: In every point in this circular flow, significant errors or medication related risks can happen. We might reach the wrong diagnosis, we might choose the wrong treatment for the patient, we may forget to treat the patient for his diagnosis, we may miss evolving adverse drug events that are emerging from the medications in the combination of that specific patient.
[8:42.29]: Med-Aware as a company came from my experience as a clinician that I encountered with a case of a nine year old boy that died simply because his primary care physician clicked on the wrong entry.
[11:14.30]: In India I get prescriptions from doctors and one of the hardest things to do is read the doctor’s prescription, right? You never know for sure what the doctor has written and what I’ve seen is that only the pharmacy that is attached to that particular doctor’s clinic or hospital — that pharmacist is an expert in reading that particular doctor’s prescription.
[12:05.61]: …The transition from manual prescribing to electronic prescribing was one we all thought that would reduce the number of errors and catastrophes, but it wasn’t exactly so for various reasons.
[21:26.85]: We need to make sure that the alert burden is very low, the clinical relevance is high and the alerts are actionable.
[22:13.36]: We are taking historical data and learning via algorithms, the behavioural patterns of clinicians, when they are treating their patients and looking for outlined situations.
[47:19.74]: We’re spotting and preventing risk at the specific patient level, the impact could be on an institutional level.
[50:28.69]: As a clinician, don’t take for granted that medical errors happen, there’s nothing we can do about it.
Connect
LinkedIn: Gidi Stein
Website: Medaware
The You+AI Vodcast
A companion video segment full of fun and candid moments.
Check it out here!