Regional doctors should want to stay after mandatory service ends

Kim Hyun-cheol

The author is a professor at Yonsei University College of Medicine and director of the Institute for Population and Talent.

The government’s logic is simple. Doctor shortages are severe in essential care and outside major cities, so train more doctors and require them to work in underserved areas.

Starting in 2027, medical school enrollment will increase by 3,342 over five years, with all additional places reserved for “regional doctors.” In return for tuition and housing support, they must serve 10 years in designated regions or lose their licenses. New public and regional medical schools will admit 100 students each from 2030.

The first regional doctors will be licensed around 2033, so the policy cannot truly be judged until the 2040s. But is it being designed for health care then? The policy rests on three misconceptions.

The first is that sending doctors to regions will work, and if not, compulsion will.

It recruited 90 doctors by offering an additional 4 million won ($3000) monthly, but 59 were already working locally and merely changed contracts.

Money is not the only problem. Spouses lack jobs, and suitable schools for children are scarce. Doctors treating high-risk patients without adequate anesthesiology, nursing or emergency staff also face greater legal risks.

Is mandatory service the answer? Japan’s Jichi Medical University requires nine years of service and achieves 98 percent compliance, while about 70 percent remain in their home regions afterward. It recruits local students, trains them locally and designs career paths beyond mandatory service.

By contrast, Korea offers compulsion without incentives to stay. Compulsion can ensure attendance, not commitment.

There is a deeper problem. In vulnerable regions, the medical need is not shrinking, but market size is. Sending doctors everywhere is ultimately unsustainable.

The second misconception is that hospitals and clinics can simply be spread nationwide.

Hospitals have strong economies of scale. Complex care around the clock requires several specialists sharing duty, plus nurses, imaging equipment, transfusion systems and intensive care.

Yet everything cannot be centralized because distance matters for emergencies and childbirth. Medical service areas should therefore be organized by time sensitivity and economies of scale, not simply disease categories.

The government is reorganizing care into three tiers: five major regional blocs plus three special regions, 70 intermediate areas and municipal-level local areas.

But while emergency and trauma care can be organized at the provincial level, rare diseases and cancer surgery cannot be completed within eight regions. Would patients accept requiring cancer patients in North Jeolla to receive treatment only there?

The third misconception is putting people before technology.

AI is already changing medicine. In a Bangladesh experiment, autonomous AI identified patients who did not need specialist referrals, increasing ophthalmologists’ hourly capacity by about 40 percent. In a randomized Swedish trial of 100,000 breast cancer screenings, AI assistance increased cancer detection by 29 percent while reducing reading workloads by 44 percent.

Utah has introduced a regulatory sandbox allowing AI to renew prescriptions for chronic diseases. In one randomized trial, doctors scored 74 on diagnostic reasoning and doctors working with AI scored 76, while AI alone scored 92.

Doctors remain necessary to decide where AI should be used and supervise it. But restricting tasks AI can safely perform merely because of licensing is not a safety regulation. It is an entry barrier.

Hong Kong faces an aging population and doctor shortages. Its answer is the Hong Kong University of Science and Technology’s third medical school, which will integrate AI and engineering with clinical medicine. Korea, meanwhile, selected Sunchon National University for a regional medical school after two universities in South Jeolla competed, separated by just 1.3 points. Should underserved regions receive doctors or technology?

Patients outside major cities do not expect every disease to be treated near home. They want timely emergency treatment for heart attacks, rapid referral to the best hospital for cancer and convenient local management of diabetes and high blood pressure.

Emergency care, trauma and childbirth should remain organized at the provincial level. Cancer, rare diseases and complex surgery, where experience and scale save lives, should be concentrated in two or three national hubs. Local primary care should center on telemedicine and AI.

People will still be needed, but Korea should make good doctors want to put down roots rather than force them. That requires better living conditions for families, career paths and legal protection.

Instead of creating doctors who count the days until mandatory service ends, Korea must create conditions that give them no reason to leave afterward. That requires a health system designed around what patients need, rather than where doctors can be assigned.

New medical schools should follow the same principle. Rather than distributing one to each region, Korea should build schools preparing for health care in 2040, transformed by AI and engineering. Korea needs an engineering-based medical school for the future more than another regional medical school.