A pioneer of AI in medicine and translational biomedical informatics, Professor Jack Li reflects on his stellar career, the impact of AI in patient safety and prevention, and the value of international cooperation for biomedical informatics development across the globe.

What initially sparked your interest in medicine?

I was pretty intrigued by human biology and anatomy when I was in high school. It sparked my interest in attending medical school in Taipei.

When did you become involved in medical AI?

When I started to pursue my PhD in 1991, I decided to study AI in medicine because, like many people, I was also inspired by Edward Shortliffe’s work on MYCIN when he was still at Stanford. Back in 1991, AI wasn’t really what it is today, but I think there are many potential ways to use AI to solve many medical problems.

When I was a medical intern in the hospital, I realized that we needed to spend a lot of time memorizing many things. However, we can’t always remember everything – what we need to consider and be careful about – when treating a patient because of the complexity of the procedures and steps. There are just too many things that we need to remember to make it right. So, I think there are so many possibilities for intelligent computer programs like AI to be a big help in terms of the complexity of the healthcare process.

You are a keen advocate for AI in patient safety and prevention. Where, in particular, are you seeing artificial intelligence making an impact in this space?

The processes in healthcare are now getting more complicated. No one can always remember every step without any assistance, and perform perfectly in all the procedures and steps whenever they face a patient. That’s why I feel like having an intelligent computer program, using AI in terms of helping the healthcare provider at the right time, is a key to actually ensuring patient safety.

On the other hand, healthcare providers are left alone in this complicated healthcare environment. As humans, providers are bound to make mistakes. And these mistakes could end up in death or health hazard for patients, which is the last thing we want. So I believe that, if we can use AI to provide timely assistance to medical professionals in this environment, it will make a significant difference.

Also, on the prevention side, it will rely more on detections. So if we can predict what happens in patients with specific risks, we can prevent it from happening. Nowadays, many AIs have been applied to predict specific events, like falling prediction. If you predict a patient will have a high risk of falling after a certain surgery, you could set up an environment to prevent falling earlier. We could use AI for all the different types of risk according to each patient’s risks to reduce unnecessary and preventable injuries.

In what areas do you see the next big advances?

One of the significant advances we have harvested and enjoyed is the result of machine learning for medical images, like X-ray, CT, MRI, sonogram, and pathology images.

I think the next significant advance will be the medical record itself, such as free texts, numbers, coded data, and other non-imaging data. If we can totally take advantage of these medical records, we will achieve much more than only using imaging AI. For example, the predictive AI that I mentioned earlier could be realized by extracting all the phenotypes from the medical record. Also, we’ll be more accurate in detecting and preventing if we can get more personal data from patients’ IoT devices, like a smartwatch or other wearable devices. Then, we don’t actually have to treat the disease – we could stop the diseases before they threaten the patient’s health. We could also alleviate or prolong the healthy period before a patient develops an actual disease.

If we could take advantage of the non-imaging medical data, along with the literature and other free text, human knowledge will have tremendous advances in medicine AI.

You have led international cooperation for biomedical informatics development across the globe. What most excites you about this collaborative effort?

I have been conducting medical informatics projects in Africa, Asia, and the US. What amazes me most is not about how complex the technology is. It’s about using the right technology at the right place and time. By doing so, you will provide the most beneficial help.

One example is a project called Lab Push in Eswatini, Africa. They had the largest proportion of AIDS patients in the world. Their national lab took a long time to deliver the T cell checkup reports back to the clinic where the patient has been serviced. Because they didn’t have a sound road system, and the reports were all paper-based, sometimes it took several months to get the reports. Some AIDS patients can’t even survive to see the reports.

However, in Eswatini, everybody has a cell phone, and they can get the text message (short message service; SMS) at a very affordable price. Since lab results were text-based, we built up a system to read the reports once they came out and disassembled them into the text message.

So we effectively text each patient’s results back to their physicians in the local clinic almost in real-time, shortening the report from weeks or months to minutes or seconds. It really helped these patients because they could not get the right treatments without the lab reports, and they still suffered a lot from their disease and could die at any time while waiting for the report.

So it’s not about using the newest technology. It’s about identifying the real needs of the local society and healthcare system and delivering the right technology. We also published a paper about this project to share the experience.

What advice would you give someone starting their career in biomedical informatics?

I will welcome them to a field with a rewarding future, and I would also like to tell them that they need to get into the details of the daily work of medical professionals. They need to know how physicians, nurses, pharmacists, and other medical staff in a healthcare organization work – they need to be in their shoes. They don’t have to do it for a long time, but they need to do it sometime. They need to go into the ICU, the operation room, to observe what they’re doing and understand why they’re doing it.

After you have a big picture and the details, you will know every step of the way, such as how they interact with the patient, what patients would think when you hand in information to them, and how the information has been exchanged among healthcare professionals. Once you get a deeper understanding of these processes and the interactions, you will have more potential to become a good healthcare professional because you now know the right place and time to use a specific type of technology that would help.

The worst approach is having a hammer in your hand and trying to find nails. That’s kind of like putting the cart before the horse, which would never work. So I would advise you to understand the details of healthcare processes. You will get inspiration of what IT can help and what AI can help from them.

Who’s been the biggest influence on your career?

Dr Homer R Warner and Dr Peter Haug from the University of Utah totally changed my perspective when doing a PhD at Utah. They made me realize how technology and biomedical informatics can make a difference in healthcare and physicians’ daily work. I learned a lot from them, and I really appreciate their help which changed my whole career.

What’s the best piece of advice you’ve ever received?

I think that would be the motto from my mentor, Dr Homer R Warner: work hard, play fair and have fun.

It would be best if you always worked hard. And you all need to work with people to achieve big goals. So you need to play fair and share the credits.

The most important thing is to have fun with what you’re doing. If you always have fun with what you’re doing, you’ll never get tired. If you never get tired, you can consistently achieve incredible results, enjoy the result and enjoy your life.

If you could return to the past, what would you change or do differently?

If there’s anything I could change, I would expose myself to computers even earlier. My first exposure to computers was Apple II when I was in senior high school. If I could do that earlier, I might have some experience with the IBM mainframe, the IBM Punch cards, and turning the switches. But still, I do appreciate my family offering me the chance to expose myself to personal computers in my early life.

Also, I do appreciate that I went to medical school first before studying medical informatics. I went to medical school with some computer skills and insights. I wrote some software programs, tried to make differential diagnoses, and tried to do nutritional planning. We also published some software.

That is a very precious experience that I have had. I was not only surprised by the capacity of the computer, I also had the chance to explore the human body and see patients in the hospital who showed me their illness and suffering.

After medical school, I went to the PhD program at the University of Utah directly. That was the happiest three and a half years of my whole life.

As one of the top 2% scientists in the world, Professor Li is a pioneer of artificial intelligence in medicine and translational biomedical informatics. He has devoted himself to evolving the next generation of Al in patient safety and prevention (“Earlier Medicine”). He has also been deeply involved in international cooperation for biomedical informatics development in Asia, America, Europe, and Africa. 

He currently serves as President of the International Medical Informatics Association (IMIA) and previously served as Vice President of IMIA and President of the Asia-Pacific Association for Medical Informatics. In addition, he has been elected as a fellow of the Australia College of Health Informatics in 2009, the American College of Medical Informatics in 2010, and the International Academy of Health Science Informatics in 2017.