International Variation in Drug Usage

An Exploratory Analysis of the "Causes" of Variation

by Ellen Nolte, Jennie Corbett

This Article

RAND Health Quarterly, 2015; 4(4):1


This article explores the range of possible causes that might explain observed international variations in the usage of medicines for selected disease areas: dementia, osteoporosis, cancer, diabetes and hepatitis C. Commissioned by the UK Department of Health, through its Policy Research Programme, it complements a quantitative analysis of medicines uptake carried out by the Office for Health Economics (OHE) of medicines uptake across 16 classes of medicines in 13 high-income countries in 2012/13. Both studies build on an earlier study led by Professor Sir Mike Richards (UK) into the extent and causes of international variations in drug usage, published in 2010. Drawing on a rapid evidence assessment, we explore, for each of the five disease areas, epidemiological factors such as the disease burden and aspects of health system and service organisation that were shown to have a direct or indirect impact on drug usage, such as reimbursement mechanisms, access to diagnosis and treatment more broadly. We also provide a summary overview of key features of the health systems and of the principles of drug assessment or approval processes across the countries included in the OHE analysis. We find that a range of factors are likely to play a role in explaining international variation in medicines use, but their relative importance will vary depending on the disease area in question and the system context. Any given level of use of a given medicine in one country is likely determined by a set of factors the combination and the relative weight of which will be different in another country.

For more information, see RAND RR-899-DH at

Full Text

A 2010 report into the extent and causes of international variations in drug usage (the Richards report) explored levels of medicines uptake for 14 categories of drugs in 14 high-income countries during 2008/09. There was a perception among stakeholders at that time that usage of new medicines was low in the UK when compared with other countries. The study showed that while some countries emerged as generally high or low users, there was no uniform pattern across disease areas and categories of drugs. For example, the US ranked first with regard to drug usage for all therapy areas combined, followed by Spain and France, but levels of usage were not consistently high across all disease areas. Generally lower than average levels of usage were observed for Norway and Sweden, and New Zealand had the lowest ranking in nine out of the 14 disease areas. The UK ranked eighth out of the 14 countries but usage patterns varied.

The Office for Health Economics (OHE) (the OHE report) updated the quantitative analyses of medicines uptake to 2012/13, employing the same method, and analysing the same 14 classes of medicines (plus two additional categories) and the same group of comparator countries (excluding Denmark for which up to date data were not available). This found that the US, Spain, and France continued to have the highest medicines usage among all 13 countries, although the order had reversed, while Norway, Sweden, and New Zealand continued to rank lowest. The UK ranked ninth across all of the medicines studied, but as before usage patterns varied across drug categories and disease areas. Specifically, in 2012/13 UK usage per person was above the international average for cancer medicines launched more than ten years ago (as at March 2013) and medicines for the treatment of osteoporosis, respiratory distress syndrome, wet age-related macular degeneration and statins. But it was below the international average for 11 out of 16 categories.

It is important to recognise that there is uncertainty about the optimum level of drug usage in different disease areas and the extent to which high or low usage point to inappropriate use. The appropriate level of usage may vary because of different factors at work in different system contexts. For example, for some disease areas, high drug usage might reflect overuse as a consequence of weaknesses in disease prevention whereas low usage would point to effective and timely treatment. For other disease areas, low usage might point to failure to meet patients' needs while high usage could indicate optimal treatment.

Thus, there are a many possible reasons that might explain variation in medicines usage across countries. The 2010 Richards report noted that causes of international variations in drug usage were complex, and it highlighted a number of themes that were thought to influence the level of usage in the UK compared with other countries. These included health technology assessment processes and outcomes; service planning, organisation and direction as important enablers or barriers to usage; and clinical culture and attitudes towards treatment. Evidence for the impact of level of spending on health was not thought to be a strong determinant as countries that spent the most on health did not always have the highest levels of usage, while those with low health expenditure could be high users of drugs.

This study builds on the analyses carried in the context of the 2010 Richards report, seeking to better understand the range of causes that may explain international variation in drug usage as observed in the 2012/13 update of the Richards report.

Specifically, we sought to provide:

  • a summary overview of key features of the health systems in the 13 countries included in the 2014 update of the report
  • a summary overview of the principles of drug assessment or approval processes in the 13 countries covered in the 2014 update of the report
  • an exploratory analysis of the “causes” of international variation in medicines usage in five selected areas: dementia, osteoporosis, cancer, diabetes and hepatitis C.

The five areas were selected following recommendations from the Metrics Oversight Group, and confirmed on review of a draft report of the quantitative analysis presented by the OHE in summer 2014 by the Metrics Oversight Group. For each of the five disease areas we reviewed the published evidence on:

  • epidemiological factors such as the disease burden (incidence or prevalence) and stage of diagnosis of the disease to understand “population need”
  • international variation in drug usage to enable the placing of the quantitative findings of the OHE analysis into the wider context
  • aspects of health system and service organisation that were shown to have a direct or indirect impact on drug usage, in particular reimbursement mechanisms, access to diagnosis and treatment more broadly, and other factors identified in the literature.

We drew on an iterative search of the published and grey literature using the bibliographic database PubMed, alongside Google Scholar and searches of websites of governmental and non-governmental agencies or organisations of documents on general health-related policies in the countries in question. It is important to note that it was beyond the scope of the present study to provide a comprehensive review of all possible aspects that could impact on an observed variation in drug usage across countries. A full understanding of system, service and cultural factors would require a different approach, involving working with decisionmakers and practitioners in each country to assess the specific systemic and cultural aspects that inform decisions in daily practice.

High-Income Countries Vary in the Way They Fund and Organise Their Health Systems But Share Some Common Features

In the majority of the 13 countries reviewed in this study expenditure on health is largely from public sources, mainly taxation and mandatory health insurance. This ranged in 2012 from just under two-thirds in Switzerland to some 85 per cent in Norway and the UK. In the US, just under half of health expenditure is from public sources. The majority of countries also provide (almost) universal coverage, with residence in the given country being the most common basis for entitlement to healthcare. The US has so far been an outlier in that entitlement to publicly funded health services was dependent on certain conditions such as age (Medicare) or income (Medicaid). The 2010 Patient Protection and Affordable Care Act (ACA) seeks gradually to expand healthcare coverage and it is projected that it will reduce the number of uninsured by half by 2022.

All systems offer a basic basket of services, including general practitioner and specialist care, and hospital inpatient and outpatient services. Among the 13 countries reviewed, access to specialist services tends to be directed by referral. Most countries have assigned a gatekeeping role to general practitioners except for France, Germany, Switzerland and the US, which have put voluntary gatekeeping arrangements in place.

All countries reviewed have introduced user charges for prescription drugs under the public system. The level of cost sharing required varies although each system applies uniform rules, with most countries offering some form of mechanism to protect the income of selected population groups.

Most Countries Have Established National Bodies That Advise Government or Are Acting on Its Behalf in Decisions About the Routine Inclusion of New Drugs Under the Publicly Funded System

Decisions are typically informed by formal health technology assessments which may be carried out by the relevant institutions or commissioned externally. Public bodies with a largely advisory or guidance producing role have been established in Australia, Canada, France and the UK, whereas in all other countries reviewed, relevant organisations or agencies have a regulatory function, such as Pharmac in New Zealand, or the Federal Joint Committee in Germany. In some countries the ministry of health has remained the final decisionmaker. In the US there is no single (national) body responsible for appraising new drugs for funding but public payers have established their own systems to undertake such assessments. Cost-effectiveness is an overt criterion in informing recommendations on the inclusion of new drugs under the statutory system in Australia, Canada, New Zealand, Norway, Spain, Sweden, Switzerland and the UK. However, decisions do not depend on the cost-effectiveness as the sole criterion with countries also taking account of factors such as patient and therapeutic benefit, health need, budgetary impact and comparative effectiveness, which may play a more important role than cost-effectiveness.

There Is No Single, Overarching “Cause” Explaining International Variation in Medicines Usage in the Five Disease Areas Explored in This Study

From our analyses it is not immediately obvious that any particular system characteristic such as the level of overall health expenditure, sources of system funding or coverage acts as a strong determinant of levels of medicines usage. The 2010 report Extent and Causes of International variations in Drug Usage suggested that health technology assessment processes and outcomes can have a significant impact on levels of usage. This observation is partly supported by observations from the disease areas reviewed in this report (dementia, osteoporosis, cancer, diabetes and hepatitis C). Policies on the inclusion of new medicines in publicly funded systems are important as they determine whether patients have routine access to a given new medicine, in particular where access is made conditional. One such condition can be the level of cost-sharing required. Thus, evidence in the field of osteoporosis suggests that patients in some European countries may face challenges in accessing osteoporosis medicines where only 50 per cent of the costs are reimbursed. At the same time, evidence in relation to drugs for dementia, osteoporosis, diabetes, hepatitis C and, to certain degree, cancer, demonstrates that factors other than policies on the inclusion of new medicines in publicly funded systems may be equally or more important in affecting drug uptake. These include: access to (timely) diagnosis; whether or not the disease area is designated a national priority; and the clear identification of responsibilities for managing the disease and the existence of designated care pathways. Each of these factors is discussed in the following sections.

Evidence for All Five Disease Areas Highlights the Key Role of Ensuring Access to Timely Diagnosis to Enable Appropriate Treatment, Including Drug Treatment

For example, regarding osteoporosis, available evidence highlights the role of access to bone density measurement technology (such as dual-energy X-ray absorptiometry) as a potentially greater barrier to treatment than the actual routine availability of drugs under the statutory system. While bone densitometry may in principle be available, related scans may not be (fully) reimbursed, or only reimbursed on certain conditions, which could limit access, with examples highlighted for Germany and France where access to bone densitometry is restricted and osteoporosis prescribing low. While it is difficult to relate these observations directly to variation in osteoporosis treatment uptake, evidence suggests that bone density measurement is associated with anti-osteoporotic drug prescription.

Evidence for cancer also highlights cross-country differences in access to specialists, which likely acts as an important driver for accessing timely treatment. For example, one study found the projected number of new cancer cases per number of medical oncologists substantially higher in the UK than in Austria, Germany, Italy and Sweden. Arguably, the number of specialists can only provide a proxy measure of access to specialist care and it provides little insight into the appropriateness and quality of care delivered, although the observed differences call for further investigation into the accessibility of specialist treatment in the UK.

Evidence for All Disease Areas Considered Suggests That Designating a Given Disease or Condition a National Priority Is Likely to Lead to Increases in Medicines Access and Usage

For example, for dementia, where comprehensive plans for the detection and treatment of dementia have been put in place, these were likely to increase the number of people diagnosed with the disease and, consequently, the number of patients receiving treatment. One such example is the 2009 National Dementia Plan in England, which was associated with an increase in dementia diagnosis rate, and dementia diagnosis rates were highly correlated with prescription rates. The UK was also an example where osteoporosis had been identified as a priority, which was likely associated with a rapid increase in the prescription of osteoporosis medicines.

Making dedicated funds available was also associated with enhanced patient access to medicines as exemplified by the Cancer Drugs Fund in the UK, although questions remain about the resource implications and impact on overall outcomes of assigning specific funds to a selected disease area versus other conditions. Another example of making additional funding available comes from the US where the implementation of Medicare Part D in 2006, which enhanced drug coverage for older people, was associated with greater use of dementia medicines among Medicare beneficiaries.

The Clear Identification of Responsibilities for Managing Disease and the Existence of Designated Care Pathways Are Important Determinants of Medicines Usage

Studies in the fields of dementia and osteoporosis highlighted the importance of identifying clear pathways with assigned responsibilities for managing a given disease or condition. For example, for osteoporosis it was argued that the majority of patients should be managed at the primary health care level by general practitioners, with specialist referral reserved for difficult cases; a core role was also assigned to fracture liaison services, providing a system for the routine assessment and management of postmenopausal women and older men who have sustained a low trauma fracture. Evidence suggests that uncertainty about responsibilities among care providers was linked with patients falling “through the cracks,” hindering access to appropriate and timely treatment, in particular among those at increased risk of fragility fractures. Similar observations were made for dementia, with evidence suggesting that availability of memory services would increase access to timely diagnosis and thus treatment.

Each of the factors described in this study is likely to play a role in explaining international variation in medicines use, but their relative importance will vary depending on the disease area in question and the system context. It is likely that any given level of use of a given medicine in one country is determined by a set of factors the combination and the relative weight of which will be different in another country.

The research described in this article was prepared for the Department of Health within the PRP project "An 'On-call' Facility for International Healthcare Comparisons" and conducted by RAND Europe.

RAND Health Quarterly is produced by the RAND Corporation. ISSN 2162-8254.