NHIA policies over the last two decades largely targeted prescription drugs and behind-the-counter drugs.
Policies governing the scope of drug coverage
According to Article 51 in the National Health Insurance Act (amended in 2011) , the NHIA does not reimburse the following: (1) medicines that are approved by the Taiwan Food and Drug Administration (Taiwan FDA) but are not used for disease treatment, such as contraceptive, hair tonic, dark spots detergent, smoking cessation patches; (2) some vaccines (e.g., quadrivalent Human papillomavirus types vaccine); (3) over-the-counter drugs and non-prescription drugs which should be used under the guidance of a physician or pharmacist (also called the behind-the-counter drugs); (4) drugs for human-subject clinical trials; (5) drugs which are deemed by the National Health Insurance as not essential for medical treatment (e.g. dentures, artificial eyes, spectacles, hearing aids, wheelchairs, canes, and other treatment equipment) or not cost-effective; (6) drugs which do not conform to the indication that stipulated in the approved indication for licensing and/or the “Reimbursement Restriction” enacted by the National Health Insurance. However, in special cases, an application for prior authorization can be made to the National Health Insurance, and the drug will be reimbursed if authorization is given; and (7) any other drug which the NHIA publicly announces that it will not be reimbursed .
Behind-the-counter drugs were covered by the former Civil Servant Insurance and Labor Insurance and in the early years of the National Health Insurance. The NHIA reviewed and reduced the scope of coverage of behind-the-counter drugs over time . Some behind-the-counter drugs were delisted to meet the priorities of the National Health Insurance, in accordance with Article 51 of the National Health Insurance Act, and more generally, to establish patient expectations and the culture of rational use of medicines and basic self-healthcare. This change also intended to reduce costs of the National Health Insurance as well as making better use of its resources for treatment of major diseases.
This delisting of behind-the-counter drugs also benefited consumers who needed such products for treatment of minor illnesses (e.g., headache, cold). They can save time and related expenses when they choose to purchase medicines at pharmacies instead of visiting physicians at primary care clinics or hospitals. For example, a patient’s out-of-pocket costs are less if s/he chooses to purchase medicines for headache from pharmacies (only US$3) compared with visiting a physician which requires physician visit copays (US$5).
In total, the National Health Insurance delisted 176 (e.g., some antacids) and 240 (e.g., some vitamins, electrolytes) behind-the-counter products in 2005 and 2006 respectively. However, there was considerable resistance from both physicians and patients; thus, no further behind-the-counter drugs were delisted. Currently, there are still around 1400 behind-the-counter products reimbursed by NHIA (e.g., gastrointestinal drugs, antihistamines, antitussive agents) . However, delisting in the future is likely under the pressure to contain costs.
The pricing system for pharmaceuticals under the drug reimbursement scheme
The pricing system under the drug reimbursement scheme of the National Health Insurance can be divided into four phases over time: (1) internal audit price (1995/3-1997/3); (2) uniform pricing (1997/4-1999/2); (3) Pharmaceutical benefit scheme (1999/3-2012/12); and (4) Pharmaceutical benefits and reimbursement schedule (after 2013/1), which are described as below and summarized in Figure 2.
Prior to the implementation of the National Health Insurance in 1995 in Taiwan, about 50% of the population was insured under the Civil Servant Insurance, Labor Insurance, and Farmer’s Health Insurance. At the time, pharmaceutical companies were allowed free pricing, and they were subject to hospitals’ pharmaceutical tender and negotiation to determine the price of drugs in hospitals. Hospitals would bill the insurers. The insurers would then reimburse individual hospitals by an approach known as “transaction cost-plus”. For drug reimbursements, the joint bid price would be paid to public hospitals while to this price plus 10-20% would be paid to private hospitals. Profits were usually used to pay for drug warehouse management, dispensing and other expenses. High-level hospitals tended to use more expensive, brand-name drugs or imported drugs because of profits from pharmaceutical sales. At the time of public bidding in public hospitals, manufacturers were reluctant to cut prices, resulting in high tender prices. Prescription drugs were paid out-of-pocket in primary care settings because most patients were not insured under the Government Employee’s Insurance, Laborer Insurance, or Farmer’s Health Insurance. As a result, patients were sensitive to drug prices; many would choose domestic, generic drugs over the more expensive, brand-name drugs or imported drugs.
At the early stage of the National Health Insurance (1995–1997), the NHIA released the “National Health Insurance Drug Items Table” that listed products being covered, and drugs were reimbursed through the “internal audit pricing” approach. However, the internal audit pricing system was unclear, and the price of imported generic drugs was relatively high due to a lack of international drug price comparison. These led to substantial variations in drug prices between domestic drugs and drugs by international manufacturers . Hospitals generally adopted the “fee for service” approach for drug reimbursements while primary care clinics adopted the “fixed fees by days of supply” approach (e.g., one-day supply of any medications received a reimbursement of NT$35, regardless therapeutic indication and actual procurement price; two-day supply was NT$70; and three-day supply was NT$100). The following ‘consequences’ were observed: (1) primary care clinics tended to reduce drug costs and were reluctant to release prescriptions to patients so patients had to obtain drugs at the clinics (resulting in the phenomenon of “next-door pharmacy”, which had negative impacts on the separation of drug prescribing and dispensing); (2) some patients were transferred to higher level hospitals in order to obtain drugs of higher prices and/or for longer supplies; and (3) use of drugs of higher prices in primary care clinics was subsequently reduced [13,14].
During 1997–1999, the NHIA invited the pharmaceutical industry to engage in the development of pricing guidance for drugs covered under the NHI (“National Health Insurance drug pricing principles”). The goals were to lower drug prices, control the growth of drug prices, reduce the prices of brand-name drugs and generic drugs, encourage the use of generic drugs, and protect the domestic generic drug market. The pricing guidance governed drug pricing by a drug classification system: (1) new drugs: the NHIA invited the medical and pharmaceutical experts to engage in the review and approval process; (2) the compound and special specification drugs: paid the same minimum price as other drugs of the same composition; (3) brand-name drugs: brand-name drugs were subdivided into two categories: ones that have no bioavailability/bioequivalence (BA/BE) generic drugs as alternatives in the market, and others that have BA/BE generic drugs as alternatives. Drugs of the former category were priced according to the international drugs with average market prices; while the price of the latter category must not exceed 85% of the average market price of international drugs; (4) the price of BA/BE generic drugs must not exceed the price of brand-name drugs; and (5) the price of non-BA/BE generic drugs must not exceed 80% of the price of brand-name drugs. From then on, drugs covered by the National Health Insurance were uniformly priced.
During 1999–2012, the NHIA attempted to address the problem of differences between drug procurement and reimbursement prices. To determine the market drug price difference, the NHIA required hospitals and manufacturers to provide the actual transaction prices and trading volumes. However, many hospitals and manufacturers resisted such investigation or supplied false declarations about prices, and the drug price gap remained a serious problem. The NHIA therefore announced the “Pharmaceutical Benefits Scheme” in March 1999 to govern the listing of drugs, the pricing of pharmaceuticals, and the adjustment of drug reimbursement prices. It also set the target to reduce the gap in drug prices to less than 15% within five years. Moreover, NHIA announced in April 1999 that drug price surveys for NHI reimbursed drugs were to be conducted every 1–2 years. Since then, the NHIA has implemented eight drug price surveys; new drug reimbursement prices were announced and implemented respectively on April 1st, 2000, April 1st, 2001, March 1st, 2003, September 1st, 2005, November 1st, 2006, October 1st, 2009, November 1st, 2011 and May 1st, 2014 (Table 1) . Adjustments of drug prices are discussed in detail in the next section.
The “Pharmaceutical Benefit Scheme” was replaced by “Pharmaceutical benefits and reimbursement schedule” on January 1st, 2013. Compared with “Pharmaceutical Benefit Scheme”, the present new scheme emphasizes stakeholder engagement (including the insurer and the relevant authorities, experts and scholars, the insured, employers, health care service providers, etc.) to discuss and design the listing of drugs and reimbursement prices for specific products. Further, health technology assessment (considering human health, medical ethics, cost-effectiveness of products, and financial sustainability of the NHI) was required for new drugs prior to listing in the National Health Insurance under this new phase. In 2007, the Center for Drug Evaluation (CDE) was created to conduct health technology assessment (HTA), that is, assessment of comparative efficacy, cost-effectiveness, and budget impact of new drugs. The CDE only provides HTA reports to the NHIA, it is not involved in pricing. The NHIA considers HTA evidence as part of the information used for listing and reimbursement decisions .
The adjustment of drug reimbursement prices
To ensure reasonable drug prices and close the gap between procurement and insurer reimbursement prices for prescription drugs, Taiwan made multiple efforts. Because institutions procure large quantities of medicines, procurement prices are typically lower than the amount reimbursed by the NHIA and the differences constitutes a profit for hospitals . To assess procurement prices, the NHIA conducted surveys to obtain drug wholesale prices from pharmaceutical companies and procurement prices from hospitals since 1999 . Reimbursements were adjusted if there was a difference of 30% or more between the average procurement price and the NHI reimbursed price. Prices were subsequently monitored and adjusted every two years for patented products, for products whose patent right has expired for more than five years, and for products that have no patent right. These drugs are further divided into the following two categories: (1) drugs from original R&D pharmaceutical companies, drugs of which the process of pharmaceutical form complies with “The Pharmaceutical Inspection Convention and Pharmaceutical Inspection Co-operation Scheme - Good manufacturing practice (PIC/S GMP)” requirements, BA/BE generic drugs, drugs approved to by the US FDA and/or The European Medicines Agency for marketing, controlled items of BE generic drugs; (2) common generic drugs which do not fall into the first category .
Studies have been conducted to examine effects of drug reimbursement price reductions in Taiwan. Lee et al.  assessed the effects of six drug price policies and found that they reduced pharmaceutical expenditures, especially for outpatient medications and for hospitals (compared with clinics) . Chen et al.  showed that reimbursement price reductions for targeted cardiovascular medications reduced the daily medical use and expenditures, but did not affect non-targeted products . Chu et al.  studied price reductions for anti-hypertensive drugs. They suggested that reimbursement price adjustments may have created an incentive for physicians to prescribe drugs with higher profit margins, and to increase prescription duration or the number of drug items per prescription . Hsiao et al.  did not find strong associations between reimbursement price adjustments and drug utilization and expenditures during 2001–2004. Chu et al.  studied effects of reimbursement price adjustments on outpatient hypertension treatments among the elderly. They found that the average cost per prescription increased slightly, and that physicians tended to prescribe drugs whose prices were not reduced instead of those subject to price reductions. Findings by Hsu et al.  indicated that prescribing shifted from targeted to non-targeted products . Overall, these studies suggest shifts in use from targeted to non-targeted products to maintain profits from drug price gaps but whether they reduced pharmaceutical expenditures is unclear.