Reference pricing is not a form of price regulation; it is a means of limiting expenditure on the reimbursement of drugs by making use of the existence of equivalent drugs on the national market and setting a reimbursement tariff (called reference price) for groups of drugs which are considered to be “interchangeable”. In many therapeutic fields one encounters a series of alternative marketed products which, while not identical, essentially serve the same purpose. Nevertheless, the prices of the drugs in the “interchangeable” group may vary very greatly, with the newest products often being considerably more expensive than the others. Rather than agreeing that public funds should cover the costs irrespective of which product in the group a physician prescribes, a reference price system sets an acceptable price level. This may be the mean of the various prices, or may reflect the price of one of the lowest-cost items in the class or an average of various low prices; alternative it may be the price of the product considered to be the most cost-effective in its category. If the physician wishes to prescribe a more expensive drug within the group, he may be called upon to provide special justification or the patient may be required to cover the extra costs.
Reference pricing as defined above, and as discussed further in this chapter, is based purely on a comparison of prices in the home country. However, there is also an alternative type of reference pricing which can be applied, in which the prices charged for drugs in other countries are also taken into consideration. Where a country has a system requiring official approval for drug prices, a comparison with prices in other countries is often used in determining the price which can be charged nationally. In that situation, lower prices encountered in other countries can be also used as references for setting reimbursement prices.
2. Use of national reference pricing
2.1. Development of national systems
The use of national reference pricing has been the subject of much debate within governmental institutions and with major stakeholders such as industry, the medical profession and patient groups. In Europe, the system has been adopted in countries such as Germany in 1989, Denmark in 1991, The Netherlands in 1993, Italy in 1995, and Spain in 1999. It is not used in large markets such as France and the UK, though in the latter country the setting of acceptable prices for generics follows a similar approach. Reference pricing has also been widely adopted in Central and Eastern European countries (CCEE). Historically, the use of national reference pricing in Europe first attracted countries with high drug prices, a large generic competitive market and large price differences between the various versions of multi-source drugs - a set of factors which rendered this approach both necessary and feasible. In recent years countries with lower drug prices have also adopted the approach (e.g., Spain and CCEE). Outside Europe, reference price approaches have been applied in British Columbia (Canada) since 1995, and in New Zealand. The possibility of adopting the system was also discussed in Japan in 1998.
In countries, such as some of the member states of the European Community, where social health insurance funds are the largest purchasers of drugs, national reference pricing can have a considerable effect on corporate strategies for R&D and on marketing policies.
Setting a National Reference price system involves four main decisions:
- defining the number and scope of each class of “interchangeable” drugs for which a reference price is to be set;
- determining the way in which the reference reimbursement level is to be calculated for each individual class of drugs;
- establishing a procedure for defining the classes and setting acceptable reimbursement levels;
- setting mechanisms to permit exceptions where these are justified.
2.2. Definition and scope of drug classes
In order to implement such a system, one must define the various classes within which drugs are considered “interchangeable”, and one must define the market segment within which the system is to be implemented. Several options exist: one can for example limit the system to certain drug categories, usually those representing a major share of a drug budget; one can apply different criteria to the various classes in order to decide on the degree of interchangeability of the drugs within each; and one can choose to introduce the method gradually, experimentally or incrementally, perhaps in order to arrive ultimately at a comprehensive reference system.
Criteria used to define drug classes vary country by country and according to the market segment which is considered for reference pricing. A class may simply comprise drugs having the same active ingredient (as is the case of Germany for level I drugs), or it may extend to drugs with ingredients which are pharmacologically and therapeutically closely similar (Germany for level II drugs and British Columbia); in the broadest approach, a class may comprise all drugs with a similar therapeutic indication (The Netherlands).
The following are some country examples on the definition and scope of drug classes for a national reference price system.
Germany was the first country in Western Europe to introduce a reference price system, initially based on classes of drugs having the same active ingredients, and later moving to classes with therapeutic comparability. The system was designed to apply both to research-based and generic drugs, which were to be included three years after patent expiration. The length of this period was judged sufficient to ensure that generic competition would allow a fair representation of market prices.
Three types of drug classes were then defined:
1. Drugs having identical active ingredients (Level I).
2. Drugs having therapeutically comparable active ingredients (especially drug containing similar chemical ingredients) (Level II).
3. Drugs with a comparable pharmacological profile (especially fixed combinations).
Level I drugs were grouped easily, but decisions on level II and III were more complex and only a selection of these drugs were clustered within the reference price system.
Spain has introduced very restrictive reference pricing, limited to level I drug classes in the German definition, involving 50 drugs with identical ingredients .
British Columbia (Canada), under its Pharmacare programme, introduced reference pricing in five drug classes only, chosen primarily because of the level of expenditure which they represent in the drug budget: NSAIDs, histamine-2 receptor antagonists, oral nitrate, ACE inhibitors and dihydropyridine calcium-channel blockers (BC Pharmacy Association).
The Netherlands uses for its reference price system (known as the GVS), the principle of therapeutic substitution [3,7]. The government identifies drug classes having the same therapeutic indication, mainly by using the ATC classification. This classification adopted by the Nordic Council and used by the World Health Organization, has been described earlier in this volume. It is considered an adequate basis for therapeutic classification for the purpose of reimbursement. The Dutch government also introduced some modifications to take into account therapeutically relevant side-effects which are not reflected in the ATC classification, but which in some cases provide a reason for classifying certain drugs into separate clusters.
2.3. Calculation of an acceptable reimbursement level
Techniques used to set the level of reimbursement prices also vary considerably from country to country. As a rule, countries calculate their reimbursement price for a particular drug class by using a (weighted) average of the prices of drugs in the group as sold on the domestic market. However, in countries where generic competition is substantial and leads to large price differences among products, the health funds or governmental agencies/commissions taking the decisions give a greater importance to the cheapest generic prices. With the growth of economic evaluation studies and investigations into cost effectiveness, some policy makers also introduced other criteria to calculate an acceptable reimbursement level based on the most cost effective therapy available (e.g., British Columbia).
The following are examples of the manner in which different countries have calculated an acceptable reimbursement level:
In Germany, the reference price is fixed in such a way that within each drug group a sufficient number of drugs remains reimbursable, thus ensuring that there is a degree of choice and sufficient availability of drugs falling within the selected reference price level, to supply the national market [1,8]. Within a comparable group, the reference price usually tends to be close to some of the lowest pharmacy prices. A model of the structure of the pharmaceutical market for each drug class is then constructed to take into account prices, pack sizes and dosages, and to estimate - through a regression technique - the prices for different dosages and pack sizes. However, there is not necessarily any justification for a linear relationship between various prices and dosage levels .
In British Columbia, the authorities have sought to take as a standard the reference drug that is most cost-effective within its class, based their choice on the scientific evidence as accepted by the national regulatory body. The figures for cost-effectiveness take into account the cost structure of the Canadian health system, but not necessarily the health care structures specific to the British Columbia province.
In The Netherlands, for each drug group recognized within the GVS, a reimbursement price is set per defined daily dose (DDD), in order to be able to compare drug prices within a reimbursement group. Instead of the treatment cost per day, the cost of an entire course of treatment is considered .
2.4. Use of international comparisons
Usually the price information used to set the reimbursement level reflects the price structure within the country that is implementing the system. It is therefore influenced by the results of the various tradeoffs in domestic price negotiations that have led to the prevailing national price structure within each drug class. Some countries may however, as noted above, opt to rely in part on international sources of price information in setting their reimbursement reference price levels. The Netherlands for instance, uses price comparisons with other European countries, such as France, Germany, the UK and Belgium, to determine maximum prices which then in turn affect the reimbursement price. The selection of countries in such a case is based on the criterion of similar identical close purchasing power.
2.5. The decision making process
The definition of drug classes and the setting of an acceptable reimbursement level is a highly political process, in particular since it may have major economic implications for industry. Most government agencies have set up committees drawn only from the Sick Funds or with participation of the Sick Funds, the medical professions or academic experts. In Germany for instance, the Federal Commission of Physicians and the health insurance funds specify the pharmaceuticals for which reference prices are to be determined . The Board of the Health Insurance Funds sets the reference prices. In British Columbia, an independent advisory committee is in charge of thorough reviews of the published scientific evidence on cost-effective therapies.
2.6. Exception mechanisms
As pointed out already, physicians may not be willing to switch all their patients from the drugs which they have come to prefer to drugs falling within the reference price; they may have various reasons for objecting, one being a belief that another drug is better suited to a particular patient’s needs. Some countries have therefore introduced into the design of their national reference price systems mechanisms to allow for exceptions to be made where these are justified. For instance, in British Columbia physicians can choose not to switch medications for particular patients if side effects or other adverse consequences are expected to result. A physician may present the case to the sick fund, arguing that the patient should be fully refunded, but the patient may ultimately have to pay the difference in order to receive a more expensive drug.
3. The effects of reference price systems
Since the first reference price system was implemented ten years ago, some studies have been performed to evaluate the effects of such a system. However, the assessment is still incomplete. Despite several years of implementation, there are still very few health outcome data related to drug utilization  nor is there as yet much research providing scientific evidence to evaluate the impact of such a system on various health outcomes . A major initiative in this field has been undertaken by British Columbia province, in response to strong opposition to the system from the pharmaceutical industry, British Columbia physicians and some other health care providers. An independent project for the Scientific Evaluation of the Reference Drug program has mandated studies from some leading US and Canadian research academic centers (Harvard University, MacMaster University, University of Washington).
One can however distill some conclusions from readily available facts, since the evolution of drug prices over given periods of time is well-documented. As a rule, pharmaceutical prices in the drug classes affected have adjusted to reference reimbursement price levels in European systems. Very rarely, companies have tried to maintain higher prices, but where they have done so, it has usually only been for a limited period. The scope of reference pricing is however generally limited to particular market segments (notably those with multiple similar products) and it is usually possible to observe that prices have gone up on other market segments. In Germany for example, drug prices in prescribing areas covered by the reference price system were over a given period reduced by 1.5%, while drug prices not covered by the system increased by 4.1% .
Some recent contributions to the field have assessed the impact of reference price systems on patient drug use and the use of other services. One series of epidemiological studies based on quasi-experimental designs, performed by the teams previously referred to, has begun to provide scientific evidence on the health outcomes of reference pricing. For instance, the Schneeweiss, Mclure and Soumerai study on ACE inhibitors  shows a 29% decline in the use of high-priced cost-sharing ACE inhibitors immediately after the policy implementation. After a transition period, the overall post-policy utilization rate of all ACE inhibitors was 11% lower than the projected pre-policy. Some discontinuation of such therapy was observed among low-income groups; however, according to the authors this did not appear to lead to a discontinuation of anti-hypertensive therapy as such. Switching patients to other anti-hypertensive drugs, under the influence of the reference price system, also led to a slight increase in use of physician services. One of the reasons why no study of the impact of the system on drug exposure or health care utilization by patients has so far been carried out lies, especially in Europe, in legal constraints concerning privacy and the protection of individuals. There are for instance no cross-institutional links of administrative databases on drug utilization at the individual level . In the British Columbia program, privacy is secured by encryption of data.
Even if the Canadian initiative and its collaborative work with US centres provide scientific evidence on health outcomes, there will still be little research and few assessments of the actual or potential impact of reference pricing (and consequent drug switching) on the appropriateness of the drug treatment which patients receive. The Thomas and Mann study  on the prescribing of statins in New Zealand may provide some evidence on the adverse effects of certain uses of products within this drug class.
In terms of the impact on health care expenditures, budgetary savings on drugs are generally attained within the classes covered by a reference pricing system (Germany, The Netherlands). As the impact on other services is not usually measured provided, it is not simple to estimate the effect on global health care budgets. In the case of ACE inhibitors in British Columbia, the saving on the drug budget far outweighed the increase in the cost of physician services. Drug budget savings amounted to Can$ 6.7 million and the policy led to a net saving of Can $6 million over 12 months. However, the budget savings achieved with reference pricing are considered to be relatively [4,6].
Reference pricing approaches have become increasingly popular among European countries, but the principle remains controversial. The main points of controversy are the objections to the system raised from an industry or clinical perspective, the definition of drug classes, the breadth of application of the system and the reimbursement level which will be sufficient to ensure continued availability of drugs (for instance in CCEE countries). In the first drug benefit programs that aimed to use generic competition in order to provide drug coverage at the best available price (e.g., Medicaid MAC program in the USA), the concept of interchangeability between original brands and generic drugs having similar bioequivalence was applied. Reference pricing, however, may use concepts of interchangeability that go beyond generic substitution and may be more questionable from the point of view of industry, the physician or the patient. Issues of heterogeneity within drug classes and limitations on interchangeability have already been extensively discussed in the literature (see ). When applying reference pricing, decision makers have usually addressed such issues by increasing the number of drug classes (smaller groups) or by considering “safeguard mechanisms” or providing for exceptions. However, incremental approaches or experiments pursued by countries such as Spain or Canada aim to search for solutions in particular to the dilemma of balancing the need for innovation with the need to contain health care budgets.
Industry usually opposes reference pricing on the basis that it does not take into account the characteristics of incremental innovation in the drug discovery process as it relates to a drug class. Companies market their recently developed drugs with claims of unique advantages in order to justify a higher price and reimbursement level than the other products with which the authorities have sought to group them. A reference price system may cause R&D resources to be diverted to therapeutic areas not covered by reference systems, as price setting in those categories is basically free.
Clinicians or patients may also question the grouping on drugs: physiological responses from individual patients to “similar” drugs may be different, in term of quality, absorption, indications, secondary effects, mode of preparation, application forms, the frequency of undesirable effects and contraindications. Adequate information for health professionals and for patients can deal with this problem. Payers and health care organizations may also reconsider drug classes and the switching of drugs, since the use of one drug may require greater involvement of other health care resources (e.g., more physician visits) than the use of another; it may also affect the duration and extent of treatment, with direct and indirect consequences for the overall health care cost .
Views on reference pricing within governmental agencies, industry, the medical profession and patient groups are very diverse. Efforts to refine methods of grouping and product classification so as to render them as realistic as possible will therefore continue. Reference pricing is here to stay, and is extensively in use in many major drug markets; as it continues to evolve, its influence will certainly further expand.
 G. Brenner, Drug policy and control of drug expenditures in Germany, report prepared for the concerted action, Network for setting an evaluation team of control of drug expenditure in Europe, EIASM, Brussel, 1995.
 British Columbia Pharmacy Association, Position Statement, reference based pricing, British Columbia Pharmacy Association, Richmond, 1996.
 J.R. Bult and Haaijer-Ruskamp, National report on drug policies in The Netherlands, Report prepared for the concerted action Network for setting an evaluation team of control of drug expenditures in Europe, EIASM, Brussel, 1995.
 P. Danzon and H. Liu, Reference pricing and physician drug budgets: The German experience in controlling pharmaceutical expenditures, working paper, Wharton School, Philadelphia, 1997.
 P. Grootendorst and A. Holbrook, Evaluating the impact of reference-based pricing, Canadian Medical Association Journal (August 1999).
 G. Guiliani, G. Selke and L. Garattini, The German experience in reference pricing, Health Policy 44 (1998), 73-85.
 GVS (Dutch pharmaceutical products reimbursement system) explanatory notes, Dutch State Gazette, April 16, 1991.
 C. Huttin, La régulation des prix du médicament en Allemagne, document interne, Commission Européenne, DGIII, 1992.
 G. Lopez-Casanovas and J. Puig-Junoy, Review of the literature on reference pricing, Health Policy 54(2) (2000), 87-123.
 S. Schneeweiss, O. Schoffski and G. Selke, What is Germany’s of adverse health outcome or substitution?, Health Policy 44 (1998), 253-260.
 S. Schneeweiss, M. Maclure and S.B. Soumerai, Effect of prescription duration on drug utilization trends after drug benefit changes in elderly: The case of reference pricing, J. Am. Geriatrics Soc. (2001) (forthcoming).
 S. Schneeweiss, S.B. Soumerai, R.J. Glynn, M. Maclure, C. Domuth and A.M. Walker, Impacts of reference pricing for ACE inhibitors on antihypertensive therapy, J. Can. Med. Assoc. (2001) (forthcoming).
 M. Thomas and J. Mann, Increased thrombotic vascular events after change of statins, Lancet 352 (1998), 1830-1831.
 J. Zammit-Lucia and R. DasGupta, Reference pricing. The European experience, Health Policy Review, paper N.10, St Mary’s Hospital Medical School, London, 1995.