AI Innovations Amid COVID-19

Drug Topics®: Thank you so much for joining me today.

Mok: It’s my pleasure.

Drug Topics®: Your research has focused on antimicrobial stewardship and infectious disease, and also on artificial intelligence. Can you provide a little bit of your professional background in antimicrobial stewardship and your interest in artificial intelligence (AI) as well?

Mok: My background is pharmacy, and I had spent 16 years in the world of infectious disease, establishing an antimicrobial stewardship program in a variety of settings and health systems prior to joining in Wolters Kluwer 2017.

Since I joined Wolters Kluwer, I've been engaging in quite a few projects on collaborating with our data science team on how we can leverage technology to improve the delivery of pharmacist-based patient care interventions to improve and optimize patient outcomes.

Even though my background is on infectious disease and antimicrobial stewardship, a lot of the hands-on experience in the world of AI have been my collaboration projects with a lot of the data scientists to gain experience in the past few years and seeing some of the exciting developments in terms of what we can do when you put clinicians and technologists together that can derive the biggest benefit for patients.

Drug Topics®: How do we currently see AI being used in healthcare?

Mok: That is a loaded question, and I will try to address that briefly. In my humble opinion, if you could look at the human body and pathophysiology, the 3 major areas where we see the biggest promise in terms of AI would be in the prevention, diagnosis, and treatment of diseases. If you allow me briefly, I will touch on a few of these.

For example, prevention is getting better and better in understanding what some risk factors are that lead to the development of certain disease states, and this is near and dear to my heart. For a lot of infections, we have a good understanding of what can lead to an increased risk for developing some of these bacterial viral infections.

If we can leverage AI to identify these patients earlier, we may have an opportunity to remove some of those risk factors or change people's behaviors that may result in a low rate of infection, if you will.

In terms of diagnosis, I'm sure some of you have heard, even in the in the general news, a lot of the AI algorithm has now been deployed to read some of the radiology records, look at CT scans, stuff like that, to help physicians with clinical decision making using an AI to see if there's a disease process that they may have missed, or to confirm some of the diagnosis that they have already made.

Similar applications of AI can be used in the world of laboratory testing. There are other records that are in the patient's charts, where the technology can help identify patients with diseases a little bit earlier. That’s where we see AI being applied for the diagnosis part of human diseases being the most promising.

Finally, the last part I mentioned, treatment is another area. If we can prevent and then diagnose a patient with disease, we’re interested in what treatment can help patients we cover or manage with these disease states. In the world of treatment, we're seeing a lot of the researchers in the academic or in the commercial setting, leveraging AI to help them with drug discovery, looking at new molecular entities and identifying targets or agents that may be of interest and may have promise in terms of accelerating the path into becoming a usable human drug.

Closer to the pharmacists heart: some of these treatments also can have side effects. We're talking a lot about precision medicine these days. How can we leverage technology to understand a patient's specific characteristics, such as their organ function, their genomics, and other characteristics to help identify some of those patients who may be at risk for having adverse outcomes or side effects in some of these treatments, and that may change some of the prescribing decisions. A first line drug that may be good for you may not be so good for me because of the differences that are inherent between us as 2 different human beings.

Using technology has a lot of promise, instead of having - I hesitate to use the term cookbook medicine to describe what is the first-line treatment - the first-line treatment may look different based on the patient’s individual characteristics.

Drug Topics®: How can AI be used to identify individuals at risk for contracting COVID-19?

Mok: That's a great question. In terms of what we're seeing with COVID-19, I think some of the things that we're learning a lot about the disease since it was first identified at the end of 2019: we're learning how it is impacting the human body, and then we're also learning a lot about what treatment works and what treatment doesn't work. I think the greatest area of interest right now is identifying patients who may have been exposed or who may have been at high risk for contracting the virus. We recognize that a significant portion of patients may be asymptomatic, and they don't even know they are carrying the virus, and yet at the same time, they can be transmitting the virus or other people who may be more vulnerable or at a higher risk or having complications.

Some of the technology that we've seen so far: using your phone to identify whether you've been exposed to somebody who might have been identified as a carrier. That can facilitate contact tracing, which is such a buzzword now. It warms my heart, as an ID specialist, this is a term that I know, but now has become a common lingo in the in the lay community.

How can we use technology to do this very labor-intensive contact tracing? If we could have the ability to actually put in a magical number that said, “Oh, this person is diagnosed with COVID-19 and then using the technology to roll back the 10 days, 7 days before, where the individual had interacted with other people and then notifying them that they may have been exposed or that they can take the proper precautions as far as getting tested.” That is very, very exciting and promising.

Certainly, we're not just talking about proof of concept. This type of technology has already been used in the United States, and we see that recently Europe has reopened, and they have certainly enabled some of this technology in European countries, where people can opt in to be notified as well as opt in to be made aware of the chain of folks that may have been exposed to the infection.

Drug Topics®: You mentioned Bluetooth can be used. Could AI take the form of applications on people's phones for COVID-19?

Mok: Absolutely. Some of the things that we've seen, AI can actually be used without diagnosis. There's certainly a possibility as people do searches, right? If you have a headache or you lost a sensation, we see that searching symptoms is often the first thing people do. Thinking about these symptoms, you have to go through an internet search engine and type in those keywords to look for those things. Mobile device search history can help at the macro level for local governments for better understanding how many people or how many citizens in our locality are searching for these terms.

Even though we may not be able to contact trace 100% of our citizens for this disease, we can leverage some of those searches to help local government agencies understand how often people are searching for these types of terms to have a better understanding of the impact of the disease in your local community.

Drug Topics®: Steve, thanks so much for joining me today and your continued work as well.

Mok: Thanks very much.