Scrutiny of pharmaceutical expenditures is increasing as this is the fastest growing cost component in ambulatory care, with pharmaceutical expenditures now typically the largest cost or equal largest component in this sector across Europe (Ess et al., 2003; Godman et al., 2008a; Simoens, 2008a; Coma et al., 2009; Barry et al., 2010; Sermet et al., 2010). Pharmaceutical expenditure is proportionally higher in middle and lower income countries at between 20 and 60% of total healthcare spending, although from a lower baseline (Cameron et al., 2009). The reasons for increasing expenditures are well known and include demographic changes, the continued launch of new expensive medicines, rising patient expectations and stricter clinical targets (Gumbs et al., 2007; Garattini et al., 2008; Lee and Emanuel, 2008; Barry et al., 2010). New biological drugs marketed at appreciably higher acquisition costs than previous standards provide additional impetus to this growth in expenditure enhancing the scrutiny (Caroll, 2005; Barrett et al., 2006; Lee and Emanuel, 2008).

This unsustainable growth has resulted in increasing urgency among governments, health authorities and health insurance companies to introduce reforms to improve prescribing efficiency for both new and existing drugs (Traulsen and Almarsdóttir, 2005; Toth, 2010). Supply side reforms for existing drugs include compulsory price cuts, measures to lower generic prices, reference pricing in a class (Anatomical Therapeutic Classification Level 4 – World Health Organization [WHO], 2009) including voluntary reference pricing, as well as delisting products from the reimbursement list when they are considered no longer to be cost–effective versus current standards (Godman et al., 2008a,b, 2009a,b,c, 2010a; Simoens, 2008b; Teixeira and Vieira, 2008; Wettermark et al., 2008; Coma et al., 2009; Elshaug et al., 2009; McGinn et al., 2010; Sermet et al., 2010). Demand-side reforms and initiatives for existing drugs include measures to enhance the prescribing and dispensing of generics. This includes academic detailing and prescribing guidance incorporating electronic prescribing support systems, prescribing targets, financial incentives including incentives to enhance substitution in pharmacies, mandatory substitution unless prohibited by government agencies and administrative barriers to relegate the prescribing of patent protected products in a class or related classes to second line (Tilson et al., 2005; Gumbs et al., 2007; Hyde, 2007; Sakshaug et al., 2007; Sjöborg et al., 2007; Gouya et al., 2008; Simoens, 2008b,c; Godman et al., 2009a,c, 2010b; Wettermark et al., 2009a,b, 2010a; Krska and Godman, 2010; Martikainen et al., 2010; McGinn et al., 2010; Sermet et al., 2010). A number of these strategies are aimed at counter-acting the commercial activities of pharmaceutical companies, who have typically been the principal source of information among physicians for new drugs (Jones et al., 2001; Prosser et al., 2003; Szecseny, 2003; Watkins et al., 2003; Pegler and Underhill, 2005). This together with the complex nature of prescribing helps explain why pharmaceutical companies in the UK currently invest over £850 mn/year in marketing activities, with similar experiences across Europe (Beishon et al., 2007; Godman et al., 2008b).

Alongside this, governments, health authorities and health insurance agencies have instigated a range of measures to address physician and patient concerns with the effectiveness and/or side-effects of generics to release valuable resources. This urgency has increased with estimated global sales of products of $US50 bn to $US100 bn of products likely to lose their patents between 2008 and 2013 (Frank, 2007; Jack, 2008). The initiatives are similar and include defined criteria for granting substitutability status for generics, publishing lists of substitutable and non-substitutable products, not reimbursing generics where there are concerns with their quality, physician and patient education, encouraging International non-proprietary name (INN) prescribing as well as incentivizing pharmacists to talk with patients when substituting to allay any fears (Allenet and Barry, 2003; Valles et al., 2003; Kjoenniksen et al., 2006; Kopp and Vandevelde, 2006; Godman et al., 2008a, 2009a; Teixeira and Vieira, 2008; Versantvoort et al., 2008; Duerden and Hughes, 2010; Sermet et al., 2010). These concerns though generally only apply in a limited number of situations (Valles et al., 2003; Kjoenniksen et al., 2006; Heikkilä et al., 2007; Shrank et al., 2009). As a result, there should be considerable opportunities for European countries to further enhance their prescribing efficiency without compromising care. This should be welcomed as further reforms are essential to maintain comprehensive and equitable healthcare throughout Europe as we are already seeing European countries experiencing difficulties with funding new premium priced ambulatory care drugs even when these are considered cost–effective. Current activities to help fund new innovative drugs include (Cooke et al., 2005; DoH, 2006, 2008; Godman et al., 2009a; Krska and Godman, 2010; Wettermark et al., 2010b):

It is recognized it is difficult for countries to learn from each other in view of different circumstances and starting points, with one approach unlikely to fit all countries. In addition, prescribing behavior is complex (Grol and Grimshaw, 2003; Prosser et al., 2003; Prosser and Walley, 2005; Wettermark et al., 2009b). Having said this, there are examples of European countries learning from each other when considering new health reforms (Toth, 2010). In addition, the plethora of different measures introduced across Europe to enhance prescribing efficiency should stimulate debates within countries on additional reforms and initiatives that could be introduced. Coupled with this, cross national comparisons of drug utilization and expenditure also help identify possible additional reforms that countries could introduce through analytical studies linking datasets from different countries and regions and matching changes in utilization and expenditure with health policy initiatives.

The objectives of this paper are to assess the influence of the many supply and demand-side reforms and initiatives introduced across Europe to enhance prescribing efficiency in ambulatory care once a decision has been made to prescribe a particular class of drug. Subsequently utilize the findings to suggest potential future initiatives that countries could consider to further enhance their prescribing efficiency given continued resource pressures. This though acknowledging the complexities involved.

This is a cross national retrospective observational study involving the analysis of reimbursed utilization and expenditure on a yearly basis for the Proton Pump Inhibitors (PPIs) and HMG CoA reductase inhibitors (statins) among European countries.

Nineteen European countries and regions took part in this study. These were Austria (AT), Croatia (HR), Estonia (EE), France (FR), Finland (FI), Germany (DE), Italy (IT), Lithuania (LT), Norway (NO), Portugal (PT), Poland (PO), Republic of Ireland (IE), Serbia (RS), Slovenia (SI), Spain (ES – only Catalonia), Sweden (SE), Turkey (TR), and the United Kingdom (GB-Eng – England and GB-Scot – Scotland). The countries reflect differences in geography, epidemiology, financing of healthcare, available resources for healthcare as well as different approaches to the pricing of generics, originators, and single sourced products (Table 1). They also reflect appreciable differences in the nature and extent of reforms and initiatives introduced to enhance the prescribing of generics. As a result, provide a number of exemplar initiatives and countries (Table A1 of Appendix).

The following definitions have been used to classify the different pricing approaches for generics across Europe, which build on previous publications (Godman et al., 2010a,c):

We acknowledge though that in some countries only branded generics are available. However, to reduce confusion only the term “generics” will be used throughout the paper.

Only administrative databases were used to ensure standardization across countries. These included (100% coverage of the population unless stated):

As discussed, two classes were chosen for in-depth analysis of ambulatory care prescribing efficiency. These were the PPIs – Anatomical Therapeutic Chemical (ATC) A02BC, and the HMG CoA reductase inhibitors (statins) – ATC group C10AA (WHO, 2009). These two classes were chosen as (AFSSAPS, 2005; MeReC Extra, 2006; National Institute for Health and Clinical Excellence, 2006; Wessling and Lundin, 2006; Godman et al., 2008b, 2009a,b; Eriksson and Lundin, 2009; Martikainen et al., 2010; McGinn et al., 2010):

Utilization rates for the different molecules in each class were computed using Defined Daily Doses (DDDs), with utilization patterns in 2007 generally compared with 2001. These dates were chosen as typically both generic simvastatin and generic omeprazole became available and were reimbursed during this time period among Western European countries (Table 2). Simvastatin was the first major statin to become available as a generic in Europe with generally no or limited utilization of lovastatin. Omeprazole was the first PPI to become available as a generic. Both events resulted in demand-side initiatives to try and enhance the prescribing of generics ahead of more expensive patent protected products to improve prescribing efficiency (Table A1 of Appendix).

The concepts of ATC classification and DDDs were developed to facilitate cross country comparisons in drug utilization especially where there are differences in pack sizes and available tablet strengths (Bergman et al., 1979; Rønning et al., 2000; Rønning, 2002). DDDs are now internationally accepted for comparing drug utilization patterns across countries (Birkett, 2002; WHO, 2003; Walley et al., 2004a; Vlahović-Palcevski et al., 2010). The ATC index from 2010 was used in this study in line with WHO recommendations (Rønning et al., 2000).

Demand-side measures, i.e., initiatives and reforms to influence subsequent prescribing or dispensing of generics, have been collated under the “4 Es,” i.e., education, engineering, economics and enforcement. This approach has been used in other settings and successfully adapted to healthcare to provide a concise and easily understandable methodology to compare and contrast the complexity and multiplicity of demand-side measures implemented within and between countries (Coma et al., 2009; Godman et al., 2009a,c; Wettermark et al., 2009b; McGinn et al., 2010). Examples of the “4 Es” include:

Reimbursed expenditures from 2001 to 2007 were typically captured for each class to assess the influence of recent reforms on overall expenditure from a health authority or health insurance perspective. The only exceptions were Austria, Germany and Norway where there are difficulties with disassociating co-payments from total expenditure. However, this typically represents only a small proportion of overall expenditure in these three countries. Expenditure data was collected in local currency.

Reimbursed expenditures, as opposed to total expenditures, were chosen for the analysis as this is the actual expenditure incurred by health authorities or health insurance agencies reflecting the focus of the paper. Reimbursed expenditures in 2007 was subsequently converted to /1000 inhabitants/year to compare expenditures across countries adjusted for population sizes. This includes currency conversions where pertinent to standardize the approach. This was based on established rates for the country; alternatively an average for the year from national banks (Godman and Wettermark, 2009a,b). 2007 was chosen for this calculation as this was the latest year for comprehensive data from all countries. Again, expenditure/1000 inhabitants/year is the internationally accepted standard approach for comparing expenditures across countries. Exchange rates used were 1 = 0.734GB£, LTL3.453, 8.219NOK, 3.783PLN, 79.24RSD and 9.25SEK (2007).

There has been no allowance for inflation in the analysis in order to directly compare the impact of different policies over time. In addition, health authorities and health insurance agencies typically refer to pre-patent loss prices when establishing reimbursed prices for generics especially for prescriptive pricing or mixed approaches to the pricing of generics (Godman et al., 2010a,c). It is acknowledged though that savings will be greater if inflation is factored in.

The data sets collected to compare prescribing efficiency for the PPIs and statins among the European countries included:

Two principal analyses were undertaken for both the PPIs and statins to assess overall efficiency, with criteria subsequently broken down into three categories. These are summarized in Table 3. The three cut-off points for assessing efficiency were chosen intuitively; however, tested among the co-authors for internal validity.

In view of the limited number of peer-reviewed publications documenting current reforms for the pricing of generics, as well as current demand-side reforms and their impact especially for the PPIs and statins among the 19 European countries and regions outside of those from a number of the co-authors, details of these were typically provided by the co-authors. This method was also chosen to add robustness and standardization to the documentation since many of the co-authors are involved with either implementing or suggesting additional reforms in their country or region. This especially as there have been concerns with the accuracy of some of the health policy information contained within some of the web based publications (Blaszczyk et al., 2007).

No attempt has been made to analyze the appropriateness of prescribing of either the PPIs or statins. This is due to a lack of access to patient databases to determine the indication and/or doses prescribed. In addition, the main emphasis of this paper is regarding prescribing efficiency once a decision has been made by the physician to prescribe either a PPI or statin. These issues though have been discussed in individual country publications (Coma et al., 2009; Godman et al., 2009a,b,c; McGinn et al., 2010).

No impact analyses have been undertaken as typically multiple supply and demand-side initiatives were instigated in each country during the study period and the datasets generally covered the whole population. In addition, the intensity of different initiatives may vary over time and between different regions further hindering the usefulness of such analyses. This is reflected in the discussion. No regression lines have been added to Figures 3 and 4 as each point represents a different country subject to different supply and demand-side reforms (Table A1 of Appendix).

Table A1 in the appendix documents the main pricing reforms for generics during the study period among the 19 European countries and regions. Table A1 also documents the nature and intensity of the demand-side reforms introduced to enhance prescribing efficiency principally for the PPIs and statins collated under the “4 Es.” Any co-payments for the product and/or indication, in addition to the standard co-payment for the package, are also included in Table A1. This recognizes that some European countries use this “economic” approach to influence utilization.

Figure 1 demonstrated the influence of the various supply and demand-side measures (Table A1 of Appendix) on PPI prescribing efficiency among the different European countries and regions as measured by the rate of change in utilization (DDDs) versus reimbursed expenditure principally between 2001 and 2007. The countries have been broken down by:

Those showing considerable efficiency, in addition to general efficiency, i.e., below the line drawn, are highlighted using the definitions in Table 3.

In both Lithuania and Poland, there was approximately a twofold difference in the rate of increase in utilization (DDD basis) versus the rate of increase in reimbursed expenditure for the PPIs between 2001 and 2007, e.g., in Lithuania utilization increased 10.8-fold between 2001 and 2007 and Poland over 150-fold between 2002 and 2007. This appreciable increase in utilization following reimbursement, which was considerably greater than seen in the other European countries, led to their exclusion from Figure 1. Serbia was also excluded from Figure 1 with comprehensive data only recently becoming available, and after the availability of generic PPIs.

The various demand and supply side reforms instigated among the European countries and regions similarly influenced prescribing efficiency for the statins (Figure 2). The same categorization for efficiency has been used (Table 3), and again countries have been broken down into geography and approaches to the prescribing of generics.

In Poland, there was over a 140-fold increase in statin utilization between 2002 and 2007 following their reimbursement, with overall a 4.5-fold difference between the rate of increase in utilization versus the rate of increase in reimbursed expenditure between 2001 and 2007. This efficiency gain was helped by the instigation of reference pricing for the statins. Again, Serbia was excluded from Figure 2 with comprehensive data only recently becoming available.

The differences seen in prescribing efficiency for the PPIs (Figure 1) translate into considerable differences in overall expenditure adjusted for the differences in population sizes, i.e., expenditure expressed in /1000 inhabitants/year and utilization by DDDs/1000 inhabitants/day (DDD/TID) by 2007 (Figure 3). The differences in geography and approaches to the prescribing of generics have again been highlighted. Expenditure figures for the PPIs will be affected by whether there are high patient co-payment levels (Table A1 of Appendix).

The differences seen in the rates of prescribing efficiency for the statins between 2001 and 2007 among European countries (Figure 2) are again reflected in considerable differences in overall expenditure in 2007 adjusted for population sizes (Figure 4). The differences in geography and approaches to the prescribing of generics have again been highlighted, with overall expenditures again affected by whether there are high co-payment levels for the statins (Table A1 of Appendix).

Additional reforms are essential across Europe to continue funding increased volumes and new drugs without prohibitive increases in either taxes or health insurance premiums. As such, we consider the findings from this study help provide future direction to health authorities and health insurance agencies as they seek to instigate additional measures. This is despite the limitations of the study design, which are discussed later.

General findings from the study include more limited utilization, and hence expenditure, of the PPIs and statins among Central and Eastern European countries compared with Western European countries (Figures 3 and 4). This is principally due to prescribing restrictions and higher patient co-payments in these countries (Table A1 of Appendix). This endorses the need to document ongoing reforms when comparing utilization rates across countries otherwise there could be concerns with the accuracy of the data provided. Table A1 of Appendix also demonstrates that the “4 Es” provides a methodology for health authorities and health insurance agencies to comprehensively categorize their current demand-side initiatives ready for comparisons.

More specific findings include the fact that both supply and demand-side reforms are essential to maximize prescribing efficiency. The findings also demonstrate that the influence of the reforms appears to be additive, with “enforcement” having appreciable influence on subsequent utilization patterns.

Prescribing efficiency in Norway for the PPIs (Figure 1) is enhanced by its aggressive prescriptive pricing policy for generics, overcoming to some extent more limited demands side measures for the PPIs compared with Sweden and the UK (Table A1 of Appendix). The various pricing policies for generics in Austria, France and Portugal also helped improve prescribing efficiency with for instance the PPIs despite limited demand-side measures in these countries for this class (Table A1 of Appendix). Overall though just concentrating on one aspect of reforms, i.e., either supply or demand-side measures but not both, will not help health authorities or health insurance agencies fully realize potential efficiency gains from the availability of generics. This is illustrated when comparing for instance prescribing efficiency for the statins in Sweden and the UK (England and Scotland) versus Germany (Figures 2 and 4). In Germany in 2007, there was very limited utilization of atorvastatin following the introduction of reference pricing for the class in 2003 at just 2% of overall statin utilization (Godman et al., 2009b) with rosuvastatin not available. This compares with 21 and 33% respectively on a DDD basis for the appreciably more expensive atorvastatin and rosuvastatin in Sweden and England in 2007 (Godman et al., 2010c). However, comparative expenditure appears similar or greater in Germany due to higher expenditure/DDD for simvastatin (Figure 4). There are also differences in prescribing efficiency between Croatia and Finland even though there are high patient co-payment levels in both countries. Prescribing efficiency has been enhanced in Finland by active reforms to lower generic prices, e.g., a 3-month's course of simvastatin during early 2006 was 17 versus 127 in 2002 before the introduction of generic substitution (Martikainen et al., 2010), as well as measures restricting the prescribing of atorvastatin and rosuvastatin (Table A1 of Appendix).

The additive nature of the demand-side measures is illustrated by greater prescribing efficiency in Catalonia (Spain), Sweden and the UK with their multiple and intensive measures based on education, engineering and economic initiatives (Table A1 of Appendix) compared with France, Portugal, the Republic of Ireland and Turkey (Figures 1 and 2) due generally to more limited demand-side measures in these countries (Table A1 of Appendix); although this is now changing in France (Sermet et al., 2010). These findings concerning the additive influence of demand-side measures endorse the results from previous studies which also showed that multiple interventions appear more successful in changing prescribing behavior than single interventions (Bero et al., 1998; Barton, 2001; Grol and Grimshaw, 2003; Prosser and Walley, 2005).

Introducing prior authorization, or other similar enforcement schemes, also enhances prescribing efficiency, e.g., statins in Austria, Germany and Norway (Figure 2), coupled with reforms to lower generic prices (Table A1 of Appendix). This compares with a more limited influence on prescribing efficiency for the PPIs in Austria with just “education” and “economic” measures in the absence of “enforcement” (Figure 1; Table A1 of Appendix). The improved efficiency seen with introducing “enforcement” measures for the statins in these three countries appears similar to the combination of extensive educational, engineering and economic initiatives for the statins in for instance England and Sweden (Figure 2; Table A1 of Appendix) along with measures to lower generic prices.

“Enforcement” can be also additive and introduced at any stage as seen in Austria where prescribing restrictions for atorvastatin built on existing educational and economic activities (Table A1 of Appendix; Godman et al., 2009c). This was also seen in Sweden where recent restrictions on Angiotensin Receptor Blockers have increased their prescribing second line building on existing educational, engineering, and economic initiatives among the counties (Wettermark et al., 2010a).

As a result of the differences in the nature and extent of demand-side initiatives across these countries, there are appreciable differences in overall prescribing efficiency between France, Ireland and Portugal when compared with Catalonia (Spain), Sweden and the UK (Figures 1 and 2), and when adjusted for population sizes (Figures 3 and 4). Reimbursed expenditure is also high in Ireland in their selected GMS population at over 60,000/1000 inhabitants/year for both the PPIs and statins.

There have been concerns that extensive demand-side measures including prescribing restrictions can alter the quality of subsequent care (Fein, 2010). However, a recent ecological study demonstrated similar surrogate outcomes in patients with hypercholesterolemia whether they were prescribed formulary drugs, i.e., generic simvastatin, or non-formulary drugs including atorvastatin (Norman et al., 2009). Recent studies conducted in the UK have also shown that patients can be successfully switched from atorvastatin to generic simvastatin without compromising care (Usher-Smith et al., 2008). Conserved resources can be re-directed to fund programs to improve compliance as well as fund increased volumes with the growing incidence of cardiovascular diseases. Compliance is a real concern in patients with chronic asymptomatic diseases (Cramer et al., 2008) rather than any minor differences in effectiveness between the statins in clinical trials, which is not seen in practice (Usher-Smith et al., 2008; Norman et al., 2009).

We are also seeing countries learning from each other as resources pressures grow. This builds on earlier examples generally within healthcare (Toth, 2010), with some examples contained in Table 4.

As discussed, we accept there are limitations with the study design, which are summarized in Table 5. However, some of these issues are less important when comparing changes in utilization and/or expenditure as opposed to comparing absolute numbers.

As a result, some of the findings especially regarding expenditures need to be treated with caution.

Never-the-less, we consider the findings will be of interest to health authorities and health insurance agencies as they plan future supply and demand-side measures to further improve their prescribing efficiency. We also believe the findings will be of interest to pharmaceutical companies as they plan for the future, especially as health authorities and health insurance agencies become increasingly proactive to conserve resources for existing products (Moon et al., 2010).

Ongoing initiatives to optimize the managed entry of new drugs will be discussed in future papers particularly as they underscore the notion that the funding of new premium priced products is an important challenge in Europe (Garattini et al., 2008).

The majority of the authors are employed directly by health authorities or health insurance agencies or are advisers to these organizations. No author has any other relevant affiliation or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript, although Morten Andersen has received teaching grants from the Danish Association of Pharmaceutical Industries for providing education on pharmacoepidemiology. No writing assistance was utilized in the production of this manuscript.