Overseas Vietnamese 'artificial intelligence' Tran Dang Minh Tri reveals the future of digital healthcare in Vietnam: Before a person leaves the X-ray room, AI has already 'caught' the disease

For many years, this remote island commune has had only one doctor. People who want to get an X-ray diagnosis have to go to the mainland because there are no medical staff capable of reading films. With the latest X-ray machine combining the newly equipped Annalise CXR Edge technology, doctors can identify abnormalities through images and support diagnosis right on the spot. If necessary, this image will be transmitted to doctors at higher levels and people will quickly receive results without having to travel far. The funding for the machine is supported by the ASIF (Australasia Impact Foundation) - a non-profit organization of which Tran Dang Minh Tri is Vice President. The two "artificial intelligence" brothers Tran Dang Minh Tri and Tran Dang Dinh Ang were once famous in the Vietnamese media for their AI (artificial intelligence) technology applied in the field of artificial insemination, invented by their company Harrison.ai. Annalise.ai, a subsidiary founded by Harrison.ai in collaboration with Australia's largest imaging diagnostics group I-MED, has launched the world's first comprehensive AI-powered chest X-ray diagnostic software with 124 imaging markers. The product is currently being used in more than 400 medical facilities in Australia and the UK, including many leading hospitals, and supports the diagnosis of more than 1 million patients worldwide each year. Tran Dang Minh Tri revealed that he will bring this technology to hospitals across Vietnam through cooperation with Viettel Solutions Corporation.

Having been applied in leading hospitals in the UK and Australia, why did you first apply Annalise in Vietnam not at a large hospital, but at the Thanh An island commune health station? I had the opportunity to work with the Vietnamese healthcare sector before founding Harrison.ai through teaching at Hanoi Medical University, and working with dedicated colleagues at the Vietnam Center for Health Care Improvement Research. From there, I saw the enormous pressure on the Vietnamese healthcare sector. Every day, healthcare workers have to take care of up to more than a hundred patients. Although the product was developed abroad, I am Vietnamese, I always hope that the product can be applied in Vietnam to support Vietnamese healthcare workers. Through cooperation with the Ho Chi Minh City Department of Health, Annalise.ai was first applied in Vietnam in Thanh An island commune because I wanted to prove that this product is not only effective in leading hospitals like abroad but also really useful for places with difficulties and lack of medical services.

Specifically, how does this product help doctors and why can it do so? The Annalise CXR product supports comprehensive diagnosis of X-ray films and can identify 124 signs on the film image. For example, when evaluating an X-ray image, Annalise CXR can identify 6 signs related to pneumonia and COVID, 7 signs of lung cancer, 17 signs related to pneumoconiosis, 22 signs related to tuberculosis. This technology is groundbreaking because current AI products in the world are very narrow and can only diagnose a few signs. Why can it do so? An AI software development requires 4 ABCD factors. A is algorithm, B is big data, C is supercomputer system (Computing) and D is demand. Annalise's advantage is that its team of doctors are also AI engineers who develop in-depth algorithms combined with clinical knowledge. The collaboration with I-MED gives us the opportunity to access their huge dataset that has been built and stored for over 20 years. With the diversity that Australia is a multi-ethnic country, this dataset can be applied to many countries. This data is used to train AI on the most modern supercomputer system in the region in the field of medical AI. And finally, about factor D - demand, when making the product, with a large team of Australian and Vietnamese doctors participating in the project, we understand very well how AI can solve practical problems in the medical industry.

What are the results of the application in practice? After more than a year of implementation in hundreds of hospitals around the world, the effectiveness of Annalise CXR has been recognized by the medical community in many countries. Currently, more than a quarter of Australian radiologists are using this product. At the end of November, Annalise.ai was awarded the Best New Imaging Technology Company Global by AuntMinnie magazine at the Radiology Society of North America RSNA Conference. A few days ago, a hospital in the UK that had just deployed Annalise's AI product shared a story with me. This hospital had been severely lacking in radiologists for many years. Therefore, the list of patients waiting for results was up to hundreds of people. When Annalise.ai's AI application was first opened at the hospital, within seconds, the AI ​​provided results for all patients. Some people have just finished taking a scan, and have not even left the X-ray room, and have already had results confirming lung nodules and are immediately sent for lung cancer screening. Evidence is very important in the medical field. From research published in the world's leading medical journal Lancet Digital Health Annalise.ai helps increase the accuracy of chest X-ray diagnosis by 45% and reduce reading time by 12%. When applied in practice in Australia, we have a research article published in the journal BMJ Open showing that in every 30 chest X-rays, AI will help radiologists find an important sign that was missed. With the application of AI, hospitals will increase the accuracy of diagnosis and enhance their brand position with patients, while taking care of more patients. Hospitals will increase revenue and reduce costs, creating a double factor to increase profits. Only with profits can they continue to reinvest in medical staff and new technologies. Are there any difficulties in the process of training doctors?

When we deployed it in Australia, we found that doctors only needed 1-2 training sessions and 1-2 days of using the software to be proficient. I believe that when deployed in Vietnam, it will be very fast. However, because it is so easy to use, many people just jump right in and use it. This makes it difficult to understand and take full advantage of its features, so we have to organize training through webinars and instructional videos .

After Thanh An, how will Annalise enter the Vietnamese market? AI in the healthcare sector requires a lot of evidence and time to gain acceptance from regulatory agencies and staff at healthcare facilities. We are very ready for that and do not consider it a problem. The implementation of AI requires a fairly high level of healthcare information technology, such as electronic medical records, image storage systems, and strong server platforms. We cannot just bring AI technology but must bring a whole package of solutions. To do that, we have just announced that we will cooperate with Viettel Solutions and the first place to apply this technology on a large scale is Hong Ngoc General Hospital, expected in February 2023.

Why did you choose to cooperate with Viettel Solutions ? In Vietnam, it is difficult to find a partner that is both a telecommunications service provider with strong bandwidth, covering the whole country, as well as a partner with advantages in information technology, and a digital solution provider that has connected with many medical facilities like Viettel Solutions. Annalise is an international company, when cooperating, we will learn very carefully about the brand value, the trust of people in the brand. And with any criteria, cooperating with Viettel Solutions is the number 1 choice. Firstly, although Annalise's products have been applied in the international market, when entering Vietnam, they still face their own challenges. That is, the population is very large, the hospitals are very overloaded. When approaching new technology, we have to find a way to adapt to the working process of doctors. Therefore, we want to cooperate with partners who have worked deeply with the medical industry. Second, in terms of technology, we also have to think about how hospitals do not have to invest heavily in IT infrastructure but can still use AI and return AI results quickly to patients. To do that, we need strong telecommunications bandwidth. When deploying AI in Australia, the UK, and soon the US, we cooperate with the largest cloud computing companies such as Amazon or Microsoft... In Vietnam, Viettel is the strongest enterprise in cloud computing infrastructure. Viettel has also invested in hospital management software and software for image diagnosis management and Annalise can easily integrate AI to deploy in the medical field. How did this cooperation story begin? 5 years ago, when I started participating in the field of medical innovation in Vietnam, I learned about Viettel Solutions. Through many seminars, I realized that when the story of digital transformation and digital healthcare was just emerging in Vietnam, Viettel had a different vision and was ahead of the market. Recently, when I wanted to bring Annalise to Vietnam, I worked with the Australian embassy to connect with domestic businesses. They introduced Annalise to Viettel because of the advantages mentioned above. This was truly a very interesting opportunity.

In your opinion, what are the biggest problems facing digital healthcare in Vietnam? How much of those problems can collaborations like Viettel Solutions and Annalise solve? Vietnam's digital healthcare is developing rapidly due to government policies as well as huge practical needs. Technology companies, startups and healthcare businesses are very creative, bringing many new solutions. It is a very potential market. But the biggest challenge is that development is not synchronized. Each hospital and clinic pursues its own solutions, so it is difficult to have a common voice. Following the world trend, after this "hundred flowers bloom" period, Vietnam's digital healthcare will shift from "creative" to "integrated". It is no longer each place making its own products, but medical staff and healthcare leaders will demand solutions that follow international data standards. People going to the hospital will require “data connectivity” between facilities to avoid waste in testing and diagnosis. Integration means combining the best solutions in the world with the best solutions in Vietnam and providing the best service to patients.

According to the experience of hospitals around the world, how long does it take to see the effects such as patients realizing the benefits of better diagnosis and they come to that hospital more often? Chest X-ray is a popular imaging technique. Most people who have annual health check-ups have chest X-rays. Although it is popular and has low cost, if read correctly, chest X-rays are very valuable because they help diagnose important diseases early, such as lung cancer or cardiovascular disease. In Australia, we have just announced the widespread deployment of the second AI product, which is AI in brain CT scan with 130 signs, helping to diagnose early stroke, cerebral hemorrhage... We will soon bring this solution to Vietnam. Chest X-ray and brain CT are the two imaging techniques with the largest number in the medical industry. Each hospital has hundreds of cases every day. Increasing the accuracy of these two important techniques will increase the outcomes for patients and the brand value of hospitals – whether public or private. In our experience, on the first day of implementation, doctors will see cases with very subtle lesions that AI helps identify early. Does a hospital have to pay a large initial cost when investing in the Annalise solution? We want our product to have the lowest possible cost, so that many patients can access it in Vietnam. Abroad, we often deploy according to the software rental model, instead of requiring hospitals to make an initial investment. When the hospital has patients and revenue, they will pay for AI. The cost of AI is usually about 10 - 20% of an X-ray or CT scan. I think that just adding 10 - 20% of the cost to increase accuracy by 45% and reduce reading time by 12% will have very clear economic and clinical effects.

Back to Thanh An island commune - where you said you wanted to prove that AI can benefit healthcare in remote areas. This is quite similar to what your partner Viettel often talks about. Is that a resonance in corporate social responsibility ? I heard at international medical conferences that we are living in a post-war period, that is, we have just overcome an extremely difficult Covid war and many people are traumatized. Doctors and medical staff are too exhausted after the war and now when returning to a new normal life, they still have to take care of hundreds of patients every day. The fact that many people leave the medical industry is a global challenge. We cannot blame them. A female radiologist in Australia shared with me that the waiting list for results in her department is always nearly 1,000 cases. Even if she works hard, the next day patients still come, and the number of patients waiting is still the same. Looking at the long list, the medical staff no longer have the spirit to deal with it, very bored. Annalise's AI can read all those cases in a few seconds, and filter to put the most urgent cases at the top, like we filter information (sort) in Excel. This way, the medical staff will prioritize handling the most important cases first, and the normal cases will be handled later. Focusing the limited resources of the medical staff on the most important cases instead of spreading them out will greatly reduce the pressure on the staff. I always keep in mind that in business, if you want to go far together, you need to share the same values. Viettel Solutions and Annalise both share the same value of putting serving society first. We want to work together to reduce pressure on the medical industry, to attract and retain people to the medical industry. Thank you for sharing this!

According to Market Pulse