Taiwan Healthcare and National Health Insurance: A World No. 1 with 99.6% Sustained by 12.61%

30-second overview: National Health Insurance began on March 1, 1995. Yeh Ching-chuan later admitted that "the preparation time was actually only three days"¹. Thirty years later, it covers 99.6% of the eligible population², has long kept administrative costs below 2%³, and has ranked first in the world on Numbeo's Health Care Index for six consecutive years⁴. Yet the system is also sustained by three things: a 5.17% premium rate, a ten-year-high nursing turnover rate of 12.61%⁵, and the 30-year political taboo of "total household income" that no one has dared touch. This article begins with the last-six-days scramble before the 1995 launch and moves to the 2026 three-shift nurse-to-patient ratio, unpacking three pieces of conventional wisdom: "NHI is cheap," "the five major specialties are hollowed out because point values are low," and "Taiwan's IC card was the world's first." It shows how a successful institutional design locked in the political path that now blocks institutional renewal.

"The Preparation Time Was Actually Only Three Days"

At midnight on March 1, 1995, National Health Insurance formally began. The external version was "world-record speed": coverage of 96% of the population, with one card accepted across 18,000 contracted medical institutions throughout Taiwan⁶. The internal version came from Yeh Ching-chuan, the first general manager of the Bureau of National Health Insurance, recalling the launch in 2021: "the preparation time was actually only three days," "at that time even the NHI cards had not been printed," "the information system also had no landing point," and everything was "fire at our backs"¹. More than 20 million NHI cards were rushed into production and completed only one month after launch. On March 1, people entered consultation rooms with their national ID cards.

If the timeline is pulled back, the system had in fact been under development for 27 years. In 1968, the Ministry of the Interior first drafted a national health insurance plan. In 1986, the Executive Yuan set 2000 as the target year for launch. In 1988, the Council for Economic Planning and Development established a dedicated planning task force⁷. On February 28, 1989, Premier Yu Kuo-hua announced in the Legislative Yuan that the launch would be moved up to 1995. In July 1990, planning work was transferred from the CEPD to the Department of Health. After Hau Pei-tsun took office, he instructed that the launch be moved up again to 1994. In February 1991, the Department of Health formed the National Health Insurance Planning Task Force. In December 1993, the preparatory office of the Bureau of National Health Insurance was formally established, with Yeh Ching-chuan as its director⁷.

The legislative phase was not calm. On July 16, 1994, Democratic Progressive Party legislator Shen Fu-hsiung went on hunger strike to protest the NHI Act⁸. On July 19, the National Health Insurance Act passed its third reading. On November 12, labor groups took to the streets in the annual Autumn Struggle, chanting "three no-insurances"⁷. On December 23, 1994, the Organic Act of the Bureau of National Health Insurance was passed. On January 1, 1995, the Bureau of National Health Insurance was established. On February 25, 1995, Premier Lien Chan ordered that the system "launch on schedule on March 1." Yeh's remark that "the preparation time was only three days" was calculated from that point¹.

The system's skeleton was selected in 1989 by a task force led by Harvard public health scholar William Hsiao, which compared models from the United States, Britain, Germany, France, Canada, and Japan before choosing a Canadian-style "single-payer" model⁹. Hsiao's reason for choosing the Canadian model was brief: "the quality of services was very high"¹⁰. Scholars later broke single-payer into five attributes: placing the whole population into the same risk pool, a uniform benefits package, adequate healthcare resources, a single purchaser setting the rules of the game, and a global budget to control medical inflation¹¹. These five points are the structural reason Taiwan's NHI could still sustain the world's highest coverage rate 30 years later. They are also the basic binding constraint behind every reform over those 30 years: touch any one of them, and you touch all of them.

The satisfaction figures from the first year are often covered over by myth. In 1995, coverage surged to 96%, but polls during the same period showed satisfaction of only 39% and dissatisfaction of 47%¹². "Universal support" is a retrospective narrative. At the time, Taiwanese society had doubts about this new bill that had suddenly fallen from above.

📝 Curator's note: "Single-payer" is translated in Chinese as "single insurer," but those four characters locked in 30 years of political possibility. Putting everyone into the same pool means any decision to change premium rates, benefits, copayments, or the method for calculating total household income has no retreat route of "pilot first," "trial site first," or "separate track first." Either everyone changes together, or nothing changes. This was written into the system's DNA in 1989, when William Hsiao chose the Canadian model. Thirty years later, it is both NHI's greatest advantage and its largest political brake.

The 2017 version of the NHI IC card. Taiwan fully replaced its cards with IC cards in 2004 and is often said to have been "the world's first," but Slovenia completed nationwide deployment as early as September 2000. Taiwan was the second case, the largest in scale, and the deepest in integration.

Lee Yu-chun's Four Characters

The phrase "five major specialties are hollowed out" has become almost synonymous with Taiwan's healthcare system in the 2020s: internal medicine, surgery, obstetrics and gynecology, pediatrics, and emergency medicine cannot recruit enough residents. But the person who coined the original phrase was not Yeh Ching-chuan, not the medical associations, and not the Ministry of Health and Welfare. It was Lee Yu-chun, a public health professor at National Yang Ming University, who in the summer of 2009 coined "four major specialties are hollowed out" for internal medicine, surgery, obstetrics and gynecology, and pediatrics¹³. Only later did the media add emergency medicine, turning it into five.

The timeline is this. In 2018, Professor Chiang Tung-liang recalled: "In the summer of 2009, Control Yuan member Huang Huang-hsiung began conducting a comprehensive examination of NHI. By 2011, as the comprehensive examination report was released, the issue of the four major specialties being hollowed out suddenly became urgent"¹³. Yeh Ching-chuan also wrote the counter-version in 2018: "What the medical community calls the five major specialties being hollowed out probably first began at a seminar, when a hospital superintendent reported that the hospital lacked physicians in internal medicine, surgery, obstetrics and gynecology, and pediatrics, and called this the 'four major specialties being hollowed out.' Afterward, people simply repeated it, emergency medicine joined the battlefield, and with the media's forced interpretations and overgeneralization, it became the 'five major specialties being hollowed out'"¹⁴.

The Control Yuan corrective measure of July 17, 2012 was the first time the phrase was formalized in an official document. Control Yuan members Huang Huang-hsiung, Shen Mei-chen, and Liu Hsing-shan wrote bluntly that the four major specialties had "little money, much work, and proximity to prison," that "medical disputes and the payment system have in effect become the two tightening hoops strangling the development of the five major specialties," and that "most hospitals lack resident manpower in the four major specialties, with some residents working more than 300 hours a month; in some hospitals with no residents at all, attending physicians in the four major specialties must also cover shifts, and with on-call duties added, monthly working hours sometimes reach as high as 400 hours"¹⁵.

Yeh has always had reservations about the "five major specialties" claim. In his 2018 article, he wrote another sentence: "It is not that all five major specialties are short of people; the overall number of physicians is sufficient"¹⁴. This counterargument is almost never heard in the mainstream narrative, but it matters for understanding the later inference that "low point values equal hollowed-out specialties." If the overall number of doctors is sufficient, then "not being able to recruit" must be examined through distribution, work environment, medical liability burden, and shift systems. A problem that can be explained by one word, "money," cannot be something that has gone unresolved for 30 years.

5.17% Appeared Twice

The political history of NHI premium rates can almost be told with one table. At launch in 1995, the rate was 4.25%. In September 2002, it was raised to 4.55%. In April 2010, during Yang Chih-liang's tenure, it was raised to 5.17%. That increase was so controversial that after the second-generation NHI bill passed its third reading on January 26, 2011, Yang resigned immediately¹⁶. In January 2013, second-generation NHI began, the main rate fell back to 4.91%, and a 2% "supplementary premium" was added. From 2016 to 2020, the main rate was reduced to 4.69%, with a 1.91% supplementary premium. In January 2021, under Chen Shih-chung, the main rate was again raised to 5.17%, while the supplementary premium also rose to 2.11%¹⁷.

5.17% appeared twice. The first time was in 2010, after Yang pushed it through and resigned. The second was in 2021, under Chen. Eleven years passed between them, and the system circled back to the same number. This equilibrium is the structure created by the double bind of "single-payer plus politically sensitive premium rates": every adjustment carries political cost; whenever that political cost reaches its limit, the rate is pushed back down; then when the finances can no longer hold, it is raised again.

The second-generation NHI's eventual choice of a "supplementary premium" was itself a compromise forced into a corner. The original version was "total household income": calculating premiums based on a household's total annual income, theoretically the fairest model. But the 2010 legislative process overturned it. After the bill passed its third reading in 2011, Yang stated directly: "The supplementary premium is a mistake." "From beginning to end, second-generation NHI was supposed to collect premiums based on total household income, but first the Ministry of Finance said total household income could not be done, and later legislators on the Finance Committee overturned the decision of the Social Welfare and Environmental Hygiene Committee, taking only three days to propose the idea of a supplementary premium"¹⁸.

The supplementary premium that began in 2013 therefore targeted six categories of income and calculated them per transaction: high bonuses, part-time wages, professional practice income, dividends, interest, and rent¹⁹. From day one, this design had two loopholes: per-transaction calculation made it easy to split payments to avoid the levy, and income outside the six categories, such as overseas income and crypto assets, was not captured. Fifteen years later, in November 2025, the National Health Insurance Administration planned to move from per-transaction collection to an "annual settlement system": cumulative income over NT$20,000 within a year would be levied, the withholding ceiling would rise from NT$10 million to NT$50 million, an estimated 4.8 million people would be affected, and NT$10 billion to NT$20 billion would be injected into the system²⁰. But in November 2025, the Executive Yuan abruptly halted the reform, saying planning would be "temporarily suspended"²⁰, triggering anger among dividend-stock investors and retirees.

"Total household income" is the phrase no one has dared touch for 30 years. William Hsiao's 1989 single-payer prototype already pointed toward an integrated-income-tax style of premium calculation. Yang Chih-liang pushed it in 2010 and stepped down immediately. In 2025, even the moderate version of "annual settlement" was braked by the Executive Yuan. Technically, it can be done. Politically, it cannot. Any reform that puts different sources of income onto the same basis will find opponents in the Legislative Yuan.

⚠️ Controversial view: The supplementary premium has long been criticized by the left as a "regressive tax." Salaried workers are compulsorily deducted every month, while capital-income earners are touched only when they receive dividends or rent, and a withholding ceiling applies. But the 2010 legislative process was exactly the opposite: left-leaning labor groups opposed "total household income" because they feared working households would be aggregated and burdened more heavily, while the right-leaning Ministry of Finance opposed "total household income" because it feared insufficient administrative capacity. The result was that the Social Welfare and Environmental Hygiene Committee's proposal was overturned by the Finance Committee in three days¹⁸. Who exactly is the "fat sheep that slipped through the net" has gone through three rounds of narrative change over 30 years, and there is still no consensus.

Official documentary by the National Health Insurance Administration, Ministry of Health and Welfare: 30 years from "the preparation time was only three days" to a 23M-person shared single-payer system.

39.8% Are Using It, 54.8% Are Paying

In 2020, National Statistics Bulletin No. 231 from the Directorate-General of Budget, Accounting and Statistics published a precise figure that is rarely cited: people aged 65 and older used 39.8% of NHI resources; people aged 15 to 64 used 54.8%; and people aged 0 to 14 used 5.4%²¹. Compared with 2015, the share for people 65 and older had risen by 4.2 percentage points. Annual per-capita NHI spending for people 65 and older exceeded NT$70,000. For those under 65, it was about NT$10,000, a sevenfold gap²¹.

That sevenfold gap, multiplied by Taiwan's accelerating aging curve, is NHI's largest financial clock of the past 30 years. Yeh Ching-chuan put it more bluntly in 2020: "The current system can last at most another five years"²². Five years from 2020 is 2025. In November 2025, the NHI Committee approved a 2026 global budget of NT$988.335 billion, growing 5.5%, while keeping the safety reserve at two months and maintaining the current 5.17% premium rate without adjustment²³. On the surface it held. But once unpacked, the way it held was precarious.

The first financial framing that needs correction is the common claim that "NHI is about to go bankrupt." NHI is a pay-as-you-go system. There is no concept of "bankruptcy." The precise issue is whether the "safety reserve" falls below the statutory floor²⁴. In June 2025, the NHIA estimated that the safety reserve was "0.96 months." By November, it had become "two months." The difference came from NT$3 billion in government subsidies plus NT$18.1 billion shifted in from the official budget²³. On November 19, 2025, NHI Committee executive secretary Chou Shu-wan gave the technical version: "For 2026, the NHI global budget safety reserve is two months and has not fallen below the level, so the insurance premium rate will remain at the current 5.17% without adjustment"²³.

The more hidden way the system is being sustained is the policy of "guaranteeing a point value of NT$0.95." Taiwan's NHI settles payments through a "global budget plus floating point value" system: the points reported by medical institutions are divided against a fixed total budget and converted into New Taiwan dollars. In recent years, counties and cities where point values fell to NT$0.7 or NT$0.8 have often been criticized as places where providers "lose 30 cents on every dollar of work"²⁵. To reassure the medical community, the government promised beginning in 2024 to guarantee point values at NT$0.95. But The Reporter disclosed that this policy was estimated to add NT$70 billion in spending, with the Executive Yuan budgeting NT$33.6 billion in subsidies, including NT$18.1 billion shifted from the official budget and NT$15.5 billion from an enhancement fund²⁵. Cheng Shou-hsia, dean of National Taiwan University's College of Public Health and the first director-general after the BNHI was reorganized, put it in one sentence: "Guaranteeing point values is equivalent to abolishing the global budget." It is "arithmetic" and "does not solve the substantive problem"²⁵.

Copayment reform began on July 1, 2023: the ceiling on outpatient drug copayments at medical centers and regional hospitals was raised to NT$300; emergency copayments were set at NT$750 for medical centers and NT$400 for regional hospitals²⁶. After the new system began, 8.45 million people were affected, and annual revenue increased by NT$3.273 billion. But Public Television Service's assessment one year later was ironic: annual outpatient visits reached 14.3 per person, a four-year high; emergency visits per person increased by 0.1 compared with the previous year²⁶. NHIA Director-General Shih Chung-liang emphasized increased primary-care chronic-disease treatment plus an added NT$5.27 billion injection, but Taiwan Healthcare Reform Foundation CEO Lin Ya-hui's criticism captured the key point: "NHI is not commercial insurance; it emphasizes shared risk in times of hardship and payment according to capacity"²⁶. Treating copayments as a "demand-suppression tool" runs into backlash from the spirit of the NHI Act itself.

Low Point Values Are a Diagnosis of the Phenomenon, Not the Fundamental Solution

The mainstream narrative attributes the hollowing-out of the five major specialties to one variable: NHI reimbursement is too low. The extended inference is: "point values are too low → physician salaries are low → no one can be recruited." In medical politics from 2024 to 2026, this inference has almost become consensus. But academia and the NHIA itself have five counterarguments that, taken together, dismantle this single-cause explanation.

The first is Cheng Shou-hsia's "abolish the global budget" argument: "Guaranteeing point values is equivalent to abolishing the global budget" and is merely "arithmetic" that "does not solve the substantive problem"²⁵. The original intent of NHI's global budget system was to act as a brake on medical inflation. Guaranteeing point values at 0.95 is equivalent to loosening that brake. Money is added, but medical utilization rises along with it; the next year, point values fall again, turning the policy into a perpetual subsidy machine.

The second is Kao Jo-hsiang's argument that "the trickle-down effect simply does not exist"²⁵. Kao, secretary-general of the Taiwan Medical Workers' Union, analyzed the financial reports of 13 private hospitals and found only a low correlation between personnel-expense ratios and profit rates. Hospitals spend surplus on construction, equipment, and expansion; it does not automatically trickle down to front-line healthcare workers' wages. In other words, the transmission chain of "government raises NHI point values → hospitals earn more money → physicians and nurses receive raises" is not supported by the financial reports of those 13 hospitals.

The third counterargument comes from Lee Yu-chun herself. In 2024, the person who coined the phrase "the five major specialties are hollowed out" told The Reporter: "Point value is a diagnosis of the phenomenon, not the fundamental solution"²⁵. The link is almost ironic: she coined the term that sounded the alarm, and 15 years later she came out to say, "It is not a money problem."

The fourth is Shih Chung-liang's position: "We hope the target point value can reach 0.95, but not through a guarantee, otherwise medical services will be distorted," and "physician salaries come from hospital regulation and allocation; operators bear considerable responsibility"²⁵. The NHIA director-general personally put the ball back in hospital managers' court: the point value provided by the government is an overall budget. How it is distributed to physicians, nurses, and technicians is a hospital management issue, not an NHI issue.

The fifth is Yeh Ching-chuan's 2018 statement that "the overall number of physicians is sufficient"¹⁴. If the total number is enough, then "not being able to recruit" is a distribution problem, not a supply problem. What has to be solved is why emergency physicians would rather go to self-pay clinics, and why obstetricians and gynecologists would rather switch to medical aesthetics, not why overall physician pay is low.

Put together, these five counterarguments lead to an uncomfortable conclusion: the framing "low point values equal the hollowing-out of the five major specialties" is simple, politically correct, and easy to mobilize around, but it cannot withstand close inspection. The real solutions involve medical dispute handling, caps on resident working hours, internal hospital distribution, specialty recruitment systems, and recognition of overseas licenses. Every one of these is ten times harder than "raise the premium rate" or "raise the point value."

💡 Did you know? People in Taiwan average 18 outpatient visits per year, more than twice the OECD average²⁷. The same cold symptoms may lead someone to visit three clinics and receive three prescriptions. This is what scholars have long discussed as "doctor shopping." One cost of cheap NHI is the structural amplification of medical utilization, and that amplified utilization in turn pushes point values down. The chain of "cheap → more use → point values fall → healthcare workers' pay cannot hold up" feeds back and accelerates itself. It is not just a linear transmission.

A 12.61% Turnover Rate, a 60% Practice Rate

The precise way to read Taiwan's nursing shortage is this: a nursing license is not the same as nursing practice. In 2023, Taiwan had about 310,000 licensed nurses, but the practice rate was only 60%, with about 186,000 actually in clinical work²⁸. One-third of newly hired nurses left within three months²⁸. The annual turnover rates, based on hospital reports, were 11.12% in 2019, 10.13% in 2021, 11.73% in 2022, and 12.61% in 2023, a ten-year high. The vacancy rate that same year was 9.05%, also a ten-year high²⁹.

The population-aging curve magnifies the problem. The average age of practicing nurses was 36.33 in 2019 and broke 39 by June 2024²⁸. The share under 40 fell from more than 70% in 2017 to below 50% in 2023. The outflow pattern, according to Han Hsing-wen of National Taipei University of Business, who analyzed the financial reports of 13 private hospitals, was 30% to NHI clinics, 25% to self-pay clinics, 10% to long-term care, 15% to other industries, and 20% leaving the workforce, going abroad, or retiring²⁸.

Global demand is expanding at the same time. By 2030, the world will be short 4.5 million nurses. Taiwan had 186,000 practicing nurses in 2023, while projected demand in 2030 is 241,000 to 260,000, leaving a gap of 55,000 to 74,000²⁸. Registered nurses in Australia earn NT$1.44 million to NT$2.2 million a year, two to three times the level in Taiwan. Lin Hsiu-chu, deputy chair of the National Union of Nurses' Associations, said an Australian scholar told her in person: "Sorry, but we already began recruiting Taiwan nurses aggressively several years ago"²⁸.

Closed beds are the most visible indicator. Among Taiwan's 22 medical centers, NHI-declared beds account for 92% of total beds. After the pandemic, the conservative estimate is that more than 30% of beds were closed²⁸. Kao Ching-chiu, deputy superintendent of Wanfang Hospital and former chair of the National Union of Nurses' Associations, gave the most direct version: "This is also the most severe loss in 30 years." "Which hospital would proactively report to its local health bureau? Hospitals do not talk about closing beds; they only say they are 'reducing load'"²⁸.

The political tug-of-war over closed beds exploded in June 2024. On June 2024, Tsai Shu-feng, director of the Ministry of Health and Welfare's Department of Nursing and Health Care, publicly asked: "Which hospitals are closing beds?"²⁸. The remark was understood as "the Ministry of Health and Welfare does not acknowledge bed closures." On June 5, Minister Chiu Tai-yuan apologized for it at the Legislative Yuan³⁰. In the same month, Chao Lin-yu, director of the spine surgery department at Chiayi Christian Hospital, gave The Reporter the front-line version: "If there are no operations to perform, do surgeons still have a career?" "Because we lack nurse anesthetists, surgeons are restricted to only one surgery day per week"²⁸. Lo Hsiang-yun, director of emergency medicine at Linkou Chang Gung Memorial Hospital, gave a line to patients' families that became one of the most quoted statements about 30 years of nursing shortage: "You can lose your temper at me; you can be rude to me. But not to the nurses!"²⁸

A mobile nursing station at Chiayi Christian Hospital in 2007. The three-shift nurse-to-patient ratio reform, moved up to begin in 2026, is the largest structural adjustment to nursing working conditions in 30 years.

The three-shift nurse-to-patient ratio reform: on March 1, 2024, the Ministry of Health and Welfare announced an administrative order, with ratios of 1:6/1:7/1:10 for day shifts, 1:9/1:11/1:13 for evening shifts, and 1:11/1:13/1:15 for night shifts. On May 8, 2026, the Legislative Yuan passed amendments to the Medical Care Act on third reading. On May 12, 2026, Lai Ching-te announced that phased implementation would be moved up to May 20, 2027³¹. Penalties are NT$250,000 to NT$500,000 for district hospitals, NT$500,000 to NT$1 million for regional hospitals, and NT$1 million to NT$2 million for medical centers³¹. This is the largest structural adjustment to nursing working conditions in 30 years, but academics generally assess that "improved working hours does not equal automatic return." Australia and Canada are still offering what they offer. Without synchronized changes to Taiwan's healthcare wage structure, the three-shift nurse-to-patient ratio can slow the acceleration of outflow, but may not reverse it.

The Blind Spots of 99.6%

The figure "99.6% coverage" is often abbreviated as "universal." More precisely, it means "99.6% of the eligible population"². Outside the definition of eligibility are three groups: missing migrant workers, unregistered infants, and overseas citizens who suspended coverage.

After missing migrant workers lose NHI eligibility, emergency medical care often leaves them with unpaid bills, a long-unresolved hole in Taiwan's migrant-worker medical policy³². A January 2025 policy report by the Global Taiwan Institute noted that the stop-and-resume mechanism for overseas citizens had long been treated as exploitation: people living abroad did not pay premiums and resumed coverage only before returning to Taiwan for medical treatment³³. On December 23, 2024, reform finally abolished stop-and-resume coverage. But after abolition, how to calculate the spillover costs, and whether to pursue past costs, remain political forbidden zones.

The National Health Insurance Database problem goes deeper. From the 1995 launch through 2024, Taiwan's NHIRD accumulated 70 billion medical records and 3.4 billion medical images, supporting more than 8,414 peer-reviewed publications³⁴. The NHIA and National Taiwan University Hospital signed an MOU to jointly build "the world's first image-information-based cardiovascular risk database"³⁴. In academia, this is one of Taiwan's greatest sources of soft power.

But on August 12, 2022, Constitutional Court Judgment 111-Hsien-Pan-13 ruled that this massive NHI database was unconstitutional on three levels: lack of an opt-out right, insufficiently clear regulations, and absence of an independent oversight mechanism³⁵. The deadline was three years. On December 2, 2025, the National Health Insurance Data Management Act passed its third reading, granting an opt-out right³⁶. It was the largest remedial lesson in NHI data governance in 30 years.

Read the two things together: 99.6% is an achievement in coverage; the unconstitutional database is a side effect of using that coverage "too thoroughly." One cost of universal enrollment is that everyone's medical records enter the same pool. Thirty years later, when big data was taken back out for research, people discovered: "I never consented." The Constitutional Court judgment was, in essence, returning NHI "data rights" to individuals. For academic research, it was a major shock. For the protection of fundamental rights, it was a correction.

Access has another blind spot that is often overlooked. The Integrated Delivery System program for mountainous areas and outlying islands began in 1999, covering 49 mountainous or outlying-island areas plus 49 non-mountainous outlying islands, and providing fixed-site outpatient clinics, 24-hour emergency care, mobile medical services, home medical care, and referral transport³⁷. NHI covers "enrollment" at 99.6%. It does not cover "accessibility." A rural resident may have an NHI card, but if the nearest hospital is 50 kilometers away, that card has a different practical meaning from the same card in the city. IDS is a remedy, not a system-level solution.

The IC Card and Database Went Too Fast and Too Far

Taiwan fully replaced NHI cards with IC cards in 2004, and this is often claimed as "the world's first." The accurate version is that Slovenia completed nationwide deployment in September 2000, making it the first country to universalize health smart cards. Both HIIS, Slovenia's health insurance institute, and a PubMed-indexed 2001 paper by Hriberšek record this fact³⁸. Taiwan was the world's second case, the largest in scale, and the deepest in integration. The difference was three and a half years.

This correction is not meant to deny Taiwan's achievement, but to correct the narrative habit of "we are the world's first." Once internalized, that habit can become an extended inference: "therefore our data governance is also world-class." Thirty years later, that is exactly how the system ran into the 2022 Constitutional Court judgment.

✦ Closing warning: The NHI IC card and NHIRD standing at the forefront of the world are achievements accumulated over 30 years. But systems that go too fast, too far, and too deep are often tripped by their own achievements. Constitutional Court Judgment 111-Hsien-Pan-13 marked not the failure of the system, but the moment when a mature system had to go back and do remedial work.

The "world No. 1" framing also needs correction in international indices. The Numbeo Health Care Index ranked Taiwan first for six consecutive years in 2024, with a score of 86/100, ahead of South Korea at 82.7, Japan at 79.3, the Netherlands at 78.9, and France at 78.1, covering 4,119 cities and 43,700 respondents⁴. But this is a user perception index, not an institutional metric. Respondents are self-selected, and the score includes subjective ratings from healthcare users. Taiwan's real position in the Bloomberg Health-Care Efficiency Index was 14th in 2018, not the often-cited No. 9. In 2020, it was rearranged into the top four because of Covid prevention performance. By 2024, the index had stopped updating³⁹.

The academic citations are substantive. William Hsiao's 2003 Health Affairs 22(3):77 article, "Does Universal Health Insurance Make Health Care Unaffordable? Lessons From Taiwan," not a Lancet article, is the most-cited English-language academic paper on Taiwan's NHI⁹. Cheng Tsung-han's 2015 Health Affairs article wrote: "Taiwan's NHI stands out as a high-performing single-payer national health insurance system that provides universal health coverage to Taiwan's 23.4 million residents based on egalitarian ethical principles"⁴⁰. Yet Princeton professor Uwe Reinhardt, who died in 2017 and had been one of Taiwan's NHI advisers since 1989, wrote in his own article: "I have not advocated the single-payer model here because our government is too corrupt"⁴¹. One of the designers openly opposed copying the model in the United States.

As for the narrative that "Obama learned from Taiwan," currently available public information does not show an official delegation directly visiting Taiwan to study the system. The accurate version should be: Taiwan is one of the cases most often cited by U.S. single-payer advocates, through a citation chain composed of scholars such as Reinhardt and Cheng and politicians such as Bernie Sanders, rather than "the Obama administration directly learned from Taiwan"⁴¹.

The private insurance market is another piece of counterevidence. A 2004 survey found that 72.3% of Taiwanese households had purchased private health insurance; in 1993, before NHI began, the figure was 63.9%⁴². The assumption that "NHI locked down the private insurance market" was reversed: nine years after NHI began, household private insurance penetration did not fall, but rose by 8.4 percentage points. Taiwanese people are practical. They know NHI is good, but they also know that needs outside the NHI benefits package, such as hospital-room upgrades, self-pay treatments, overseas emergency medical care, long-term care, and targeted cancer drugs, require private insurance as a supplement. This is also the structural reason why the question "is NHI enough?" has become increasingly common 30 years later.

Three Sentences Put Together

Thirty years later, the small card from midnight on March 1, 1995 is still here, except now it has entered an app and become a QR code. But the same system rests on three things: 99.6% coverage, a 5.17% premium rate, and a 12.61% nursing turnover rate. The first two are sources of pride for Taiwanese people. The third is an account no one has dared settle for 30 years.

Yeh Ching-chuan said in 2020: "The current system can last at most another five years"²². Yang Chih-liang said before resigning after the 2011 third reading: "The supplementary premium is a mistake"¹⁸. In 2024, Lee Yu-chun, the person who coined the original phrase "four major specialties are hollowed out," said: "Point value is a diagnosis of the phenomenon, not the fundamental solution"²⁵. Put together, the three sentences are not exactly pessimistic. They are more like words finally spoken after 30 years of implementation. A world-class NHI system needs world-class investment. But "investment" has never meant only adding money. It points to total household income, internal hospital distribution, resident working hours, the burden of medical disputes, and data governance. Each is ten times harder than adjusting premium rates.

William Hsiao's reason for choosing the Canadian model in 1989 was that "the quality of services was very high"¹⁰. Thirty years later, Taiwan's globally highest coverage rate is sustained by the lowest payment point values and a 12.61% nursing turnover rate. Success is not the opposite of defects. Success itself produced the political path dependence of the defects. Getting out of it will not depend on the next "three days of fire at our backs." It will depend on someone, for the first time in 30 years, being willing to touch the forbidden zones of "total household income," "hospital distribution," and "data opt-out rights," one by one.

The main building of National Taiwan University Hospital. Founded in 1895, it is the largest medical center in Taiwan's NHI system. Thirty years ago on March 1, when National Health Insurance began, this was also one of the places where national ID cards stood in for NHI cards that had not yet been printed.

Further Reading:

• Medical Care Act — The NHI Act governs "benefits," while the Medical Care Act governs "institutions"; together, the two laws form the benefits side and institutional side of Taiwan's dual-track healthcare governance
• A Practitioner's Account of the History of Taiwan's Two Regenerative Medicine Acts — Outside the boundaries of NHI benefits, the two regenerative medicine acts passed in 2024 form another governance track beyond the NHI SOP
• Taiwan's Animal Drug Controversies — NHI covers people, not animals; disputes over pet medication are a comparison case for the NHI system
• Taiwan's Disaster Medicine System — NHI sustains everyday healthcare, while the disaster medicine system sustains exceptional times; together, the two systems construct the ordinary and emergency sides of Taiwan's public medical governance

References

Image Sources

• Hero: NHI Building, ROC-MOHW-NHIA Taipei Division main entrance by Solomon203 / Wikimedia Commons / CC BY-SA 4.0
• Scene-mid 1 (2017 version of the NHI IC card): Official image from the National Health Insurance Administration, Ministry of Health and Welfare / Public Domain (public government information)
• Scene-mid 2 (NHI 30th-anniversary documentary iframe): Official YouTube channel of the National Health Insurance Administration, Ministry of Health and Welfare
• Scene-mid 3 (Chiayi Christian Hospital mobile nursing station, 2007): Chiayi Christian Hospital / used for nonprofit educational purposes
• Closure (National Taiwan University Hospital main building): National Taiwan University Hospital main building / Wikimedia Commons / CC BY-SA 4.0