Pusan National University Hospital (PNUH) is the medical institution that represents Busan and South Gyeongsang Province. Armed with about 1,400 hospital beds, it boasts a lot of patients and serves more than 800,000 outpatients a year, indicating the strong foundation it has established in the nation’s southeastern region. The hospital has recently embarked on a new challenge -- introducing artificial intelligence into the medical field. At stake was how to keep local patients from traveling to Seoul and its vicinity on high-speed trains to receive medical treatments.
|Professor Kang Tae-woo (right) treats a breast cancer patient using Watson for Oncology.|
Ceaseless parade to Seoul by provincial patients
Last September, PNUH opened a precision medical center. It was to provide differentiated health care services in keeping with new paradigms, especially precision medical services based on regional hubs.
In January, the hospital introduced IBM’s “Watson for Oncology” and “Watson for Genomics.” Gacheon University’s Gil Hospital launched the Watson for Oncology last year, but PNUH has become the first Korean hospital to also introduce Watson for Genomics, which analyzes genetic data.
PNUH is using Watson to treat breast, stomach, colon and lung cancer, among others. Though its multidisciplinary committee has mainly used information provided by Watson, individual medical staffs can also use it anytime they want.
The hospital spent more than six months educating internal staff and supporting staff on how to use Watson to help the medical staff other than members of the multidisciplinary panel make use of Watson. It also conducted a simulated test using the existing medical data.
The biggest reason PNUH introduced Watson was the “medical expedition” (to Seoul) of local patients. As a matter of fact, it was nothing new that patients not just with severe diseases like cancer but also with chronic ailments head toward large hospitals in the capital area, nor is the problem limited to PNUH.
|Chung, Joo-seop, chief of Precision Medical Center in PNUH, decided to use Watson to attract patients who leave for Seoul.|
In 2015, Koreans spent 64.8 trillion won ($57.8 billion) on medical treatment, and spent 33.4 trillion won, or 51.5 percent of the total, in regions other than their cities or provinces. According to the regional medical spending statistics at National Health Insurance System in 2015, the largest amount of medical expenses flew into Seoul’s Jongno District, which has Seoul National University Hospital and other large medical institutions.
That prompted PNUH to decide to contend with high-quality medical service, which in turn forced it to introduce Watson.
Chung Joo-seop, head of the Precision Medicine Center at PUNH, lamented that too many local patients are traveling to Seoul riding KTX trains.
“We chose to introduce Watson out of a sense of mission to help patients receive the best treatment at the best timing based on trust here,” Chung said. “We have long been making efforts to upgrade service quality and thought the fourth industrial revolution could be an alternative. That’s why we set foot in AI, which the U.S. had already introduced, ahead of other Korean hospitals.
People have paid keen attention to the news that PNUH has introduced Watson, as shown by the stream of inquiries from patients, let alone medical professionals, throughout the nation.
|Professor Park Do-youn of PNUH demonstrates the use of Watson for Genomics in front of other medical staff.|
Would Genomics solve problems unsolved by Oncology?
PNUH plans to use Watson not only for treatment but research as well.
Medical staffs can first diagnose complex cancer using Watson for Oncology and can reflect the latest information based on clinical guidelines on treating them later. In case the crews cannot receive treatment options from Oncology for relapsed or metastatic cancer patients, they can get customized treatment options by analyzing DNA sequencing and medical information using Genomics. In short, Watson increases the accuracy and efficacy of treatments for cancer patients, and, at the same time, serves as a medical assistant for physicians by providing a vast amount of research results and data in a short period.
Watson, as a combination of oncology and genomics techniques, is especially useful as a means of providing clinically useful data. “As we introduced Watson for the sake of research, the hospital poured its research budget into it,” Chung said. “Watson prevents diagnostic errors by medical staffs and helps to write theses by inviting their participation in research tasks related to genomic information and AI.”
30 minutes per patient but Watson turns it into fun
Watson has already brought about changes in the hospital. There remained some negative perceptions about AI last year. But medical staffs are enhancing collaborative treatments and showing heightened interests in related research this year, according to hospital officials
|PNUH Professor Kang Tae-woo says Watson’s introduction has significantly changed the atmosphere of his hospital.|
Professor Kang Tae-woo of breast surgery department, one of the doctors who use Watson most frequently, said it usually takes 20-30 minutes to diagnose a patient using Watson. “It’s difficult because it takes a lot of time and energy, but change is necessary,” Kang said.
Currently, the hospital is using Watson in limited ways, such as when there is disagreement among medical staff or the situation seems suitable for the Watson platform. In case the crew chooses to use Watson, they have to ask Oncology for proper treatment after inputting the patients’ underlying and disease data. Based on much available data, Watson then provides information on treatment, classifying them into “recommended treatment,” “treatment worthy of an attempt,” and “non-attemptable treatment.” The doctor will then make a final decision taking into account factors such as whether the treatments are permissible in Korea, whether the drugs are usable here and whether they are covered by insurance.
The hospital also minimized the time for doctors and nurses to input patient data using Watson’s natural language recognition function. The difficulty remains, as Watson provides information in English and fails to take into account Korean reality, including its health insurance system.
“To treat patients using Watson, we first open a multi-disciplinary clinic after diagnosis and surgery to explain the results of the biopsy and the patient's condition, as well as the treatment directions the physician has in mind,” Professor Kang said. “When I look at the information in Watson later, I check whether there is any difference between it and other treatment in a broad framework.” Watson suggests anti-cancer drugs widely used today, but there are times when we need to consider cultural or racial differences in our country."
For instance, Watson may give a treatment as its first suggestion, but it would be difficult to go ahead with it if the drug price reaches 10 million won or so, Kang said. He emphasized that doctors should judge the reason and background in adopting Watson’s recommendation, particularly when domestic hospitals have long ceased to use the drugs or used them decades of years ago.
Nonetheless, he said Watson gives doctors a more critical eye in making diagnoses, as well as the opportunity to get a second opinion on patients with complicated cases.
“AI is fun and helpful,” Kang said. “It has endless possibilities, and it offers customized treatment options for patients who may have relapsed cancer or when it seems like there is no cure.”
Will policymakers use PNUH’s medical treatment information?
As early as two years later, PNUH plans to analyze Watson-involved cases to share them with local medical professionals. It will look into how medical staff’s perceptions of AI may have changed, and what the implications of Watson may be for the broader picture of developing healthcare policies.
To provide hope for PNUH patients and further develop the nation’s medical industry, it will analyze patient treatment data. Watson can also be of help for basic research at the Precision Medical Center. The hospital expects to become a leader in the nation’s southeastern region in finding the optimal target treatment for the gene mutation.
Chung said that although there hasn’t been much change aside from doctors reaffirming their decisions through Watson, he hopes PNUH will eventually become a regional hub that helps to spread AI and Watson services to other hospitals.
“Once we have sufficient, medical staff from Seoul may be coming here for research, he said. “As a national university hospital, the Ministry of Health and Welfare and the Health Insurance Review and Assessment Service can use its data. We will be able to play a leading role in making the nation’s healthcare policies.”
The industry will also be watching how high PNUH can fly attached with the wing named Watson.
<© Korea Biomedical Review, All rights reserved.>