Revaluation of patient care needed as challenges in medical tech evaluation persist: experts < Special < Article
Korea’s healthcare system is at a critical juncture as it confronts a severe shortage of medical staff and the urgent need to redefine the value of patient care.
With hospitals racing to integrate artificial intelligence (AI) technologies, a pivotal debate arises: How can the life-saving potential of these innovations be accurately measured against traditional pricing models, which often fall short?
“At this stage of healthcare evolution, establishing methods to assess corporate and technological value is essential,” said Kim Kwang-joon, CEO of AITRICS, a Korean provider of AI-based medical services.
Speaking at the K-Healthcare Business Forum, co-hosted by Korea Biomedical Review, The Korean Doctors’ Weekly, and the Korea Health Industry Development Institute on Wednesday, he emphasized the need for robust frameworks to evaluate both the medical and economic value of AI technologies.
Valuation standards are evolving in Korea and the U.S., where the national health insurance system often struggles to adapt to new medical technologies. “Healthcare products must show tangible benefits, like improved patient survival rates, to justify their costs,” Kim said.
Currently, Korea’s healthcare system operates on a fee-for-service model, which incentivizes volume over quality—more procedures lead to more revenue, often compromising patient outcomes. In contrast, the U.S. blends fee-for-service with bundled payments. Kim noted that discussions are underway for a potential shift toward value-based payment systems, with a focus on developing tools for value-based assessments, despite the complexities involved.
This transition highlights a significant shift: while fee-for-service models focus on the volume of procedures, value-based models assess tangible benefits delivered to patients, including improved survival rates. “This shift necessitates a reevaluation of the economic implications of saving lives, including the societal costs of preventable deaths,” Kim said.
In Korea, the Health Insurance Review and Assessment Service (HIRA) is crafting guidelines to assess the effectiveness and value of medical AI solutions across various categories, including imaging, pathology, biometrics, and wearable devices.
Similarly, in the U.S., the Centers for Medicare & Medicaid Services (CMS) has introduced the New Technology Add-on Payment (NTAP) system to structure medical AI evaluations and pricing. Kim remarked that “the potential for adopting similar approaches in Korea is clear,” especially as the evaluation of innovative medical technologies aligns with the assessment of new AI solutions.
AITRICS developed AITRICS-VC (VitalCare), an AI solution that predicts patient deterioration. After receiving approval from the Korean Ministry of Food and Drug Safety in 2022 and 510(k) clearance from the FDA in July, the solution is now used in about 50 hospitals. However, Kim highlighted the challenge of valuing these predictive capabilities due to a lack of standardized evaluation metrics. “The economic benefits of this time-saving capability can vary significantly depending on local labor costs, highlighting the need for market-specific evaluation criteria,” he said.
“As the medical AI landscape evolves and Korea faces a shortage of healthcare professionals, establishing clear and consistent standards for evaluating and pricing these technologies will be essential for their successful integration into healthcare systems.”
Lunit faces hurdles in clinical trials and insurance entry due to unclear evaluation criteria
Lunit, a Korean developer of AI solutions that detect early-stage cancer, is also confronting an urgent question about its future.
“Given that our pipeline isn’t particularly extensive, I’m left wondering what additional clinical trials we need to conduct to receive formal registration in Korea or to enter public health insurance systems in the U.K. or the U.S.,” said Park Sun-young, Lunit’s market access and strategy team lead. “What criteria should we follow, and can we stop here?”
Lunit’s flagship product, Insight CXR, is designed for detecting pulmonary diseases and is currently sold in Korea under provisional approval. Another key product, Insight MMG, helps interpret mammograms for breast cancer. Both products have non-reimbursed approval in Korea, but the company is looking to secure broader international market entry by generating stronger clinical evidence.
Park said that Lunit must consider whether it is generating various forms of evidence from the perspectives of patients, healthcare providers, and insurers, determining the appropriate research levels, clinical environments, and populations for its studies. “It’s currently unclear what the ultimate market value and cost will be, so we want to ensure we start with solid evidence from what we are doing well,” she explained.
Currently, there is no clear global criteria for assessing innovative medical technologies. “We must demonstrate diagnostic accuracy from the patient’s standpoint while showing treatment outcomes. While the efficiency of healthcare institutions is essential, we also need to quantify cost-effectiveness and the economic impact of our products using familiar metrics,” Park said.
A recent study in Sweden highlighted that using Lunit’s software reduced follow-up cases and improved detection rates, leading to its adoption in the national breast cancer screening program. The U.K. Prime Minister also recently announced progress in replacing one human reader with AI in national screenings.
Lunit’s efforts have led to strong international growth, with over 80 percent of its revenue now coming from overseas markets. However, Park expressed concern about the challenges AI healthcare companies face in navigating evaluation pathways, particularly in Korea. “Even with solid medical evidence, it’s difficult to communicate this to medical institutions when reimbursement options are unclear,” she explained, noting the lack of clear evaluation tracks complicates market entry.
Without reimbursement or coverage recommendations, medical institutions institutions are hesitant to adopt Lunit’s AI solutions. As a result, the company is working to establish reimbursement pathways and engaging health ministries abroad, often fielding questions about the situation in Korea.
The classification of Lunit’s products—whether they are new medical technologies or enhancements of existing ones—remains a complex issue. Some AI tools replicate tasks previously done by healthcare professionals, while others assist with basic detection, complicating the evaluation process.
Although progress has been made with government and academic support, uncertainties around the purpose and methods of use remain. Unlike overseas markets, where non-reimbursable products are sold under various payer systems, Lunit operates under non-reimbursable conditions in Korea. “The healthcare sector in Korea, tied closely to social security systems, presents a difficult environment for AI products,” said Park.
She added that despite years of building credibility, there “remains uncertainty about how long it will take to achieve reimbursement status or what pricing might look like,” stressing that as its software evolves in a competitive global market, entry procedures need to adapt. “Clear evaluation criteria, especially for reimbursement adequacy set by HIRA, are critical,” Park noted.
Heuron aims to bridge stroke care communication gaps amid regulatory hurdles
Heuron, a Korean medical AI company specializing in neurological disease, is on a mission to revolutionize stroke care and tackle neurodegenerative diseases by addressing a critical communication gap between hospitals and physicians.
“In healthcare, the market formation demands thorough health verification and an understanding of customer demographics,” said Park Chan-ik, vice president of Heuron, as he emphasized the lengthy product lifecycle in healthcare, which encompasses development, clinical trials, and regulatory approval.
To achieve success, Park said Heuron must clearly define its target customers—be they hospitals, insurers, or patients—and establish effective channels for delivering its value. He stressed that clinical validation is paramount, with robust evidence essential for securing insurance coverage. “Aligning our business model with regulatory and public relations frameworks will be vital for translating value in the medical field as we navigate these complexities,” he explained.
Understanding the regulatory landscape is crucial, especially as the fee-for-service model continues to dominate healthcare in Korea and worldwide. While pilot projects for bundled payment systems exist, the straightforward input-output measurement of fee-for-service remains preferred. Bundled payments focus on comprehensive disease management and evaluating value through outcomes and costs, but measuring these aspects poses significant challenges. Ongoing debates among educators and policymakers highlight the difficulties in implementing effective outcome assessments.
Heuron is honing in on stroke management with its brain disorder analysis solutions, which are designed to automatically detect potential intracranial hemorrhage cases using non-contrast CT images. Park highlighted the importance of timely intervention in preventing severe disabilities. “The concept of the ‘golden hour’ underscores the critical need for rapid treatment, supported by extensive research validating its impact on outcomes,” he stated.
“Demonstrating that our solution reduces treatment time is key to improving patient outcomes, giving us a competitive edge in the AI healthcare landscape,” Park added. “Our alerts could potentially cut treatment times by over an hour, leading to more than a threefold increase in patient outcomes, according to the studies we are validating.”
However, issuing alerts within hospitals alone is not enough; patients often arrive at tertiary hospitals in critical condition, and even a two- or three-minute reduction in treatment time can significantly affect outcomes. To expedite this process, Heuron has created an app for primary and secondary hospitals to enable quick assessments and seamless transfers of stroke patients.
An initial prospective study at a single institution showed an impressive reduction of approximately 11 minutes from CT scan to treatment and a total intervention time decrease of 45 minutes. Park asserted that these results bolster Heuron’s push for recognition and discussions around insurance reimbursement.
Despite being designated as an innovative service, reimbursement codes limit Heuron’s fees to those set for radiological interpretations, highlighting the challenges of the fee-for-service model. “We are actively exploring ways to break free from this pricing structure and persuade insurers to adopt a more favorable view,” Park said.
link