Deriving reliable well-documented usage and cost estimates for prescription drugs in a single country is a major challenge requiring a complex study design that takes into account the size and nature of the sources, as well as the quality, specificity, and applicability of the recorded cost estimates. Reliable comparison of costs between countries is even more challenging, and according to a recent report from the U.K. Secretary for Health is "fraught with methodological difficulties." It requires country-specific, comprehensive, well-constructed medical data sources that yield similarly derived information on specific drug usage and costs in people with similar general health status.
As a first step in the construction of a well-defined and interpretable study design that would yield reliable and comparable estimates of physician drug prescribing, we directly reviewed the complete electronic clinical records of a large sample of persons in both the U.S. and U.K. databases to obtain an initial general familiarity with the nature and content of the available information derived from two well-validated, large, comprehensive, electronic medical databases. As anticipated, there were major complex differences among individuals. Many were prescribed drugs on a continuous basis. Some drugs were prescribed for only a short period of time. Some drugs were discontinued earlier than intended, and/or the patient was switched to another drug for the same indication. Usage varied considerably in relation to dose and duration, as well as age, sex, and country.
Given the enormous complexities of medical therapeutics, we initially chose a limited but direct approach to documenting differences in the decision to prescribe drugs between physicians in the U.S. and those in the U.K. First, we restricted the study populations to 1.6 million persons in each country matched on age and sex for the years 2004–2006. We then calculated the percentage of people who were prescribed a drug at least once in a given year according to indication. This evaluation, although strictly limited in scope, demonstrated some notably large prescribing differences among physicians in the two countries. Those in the U.S. consistently prescribed antibiotics annually on average to about 50% more people of all ages over the 3-year period compared with the U.K. However, asthma drugs were prescribed for more people in the U.K. at all ages compared witih the U.S. Overall, physicians prescribed statins for an estimated 40% more people in the U.S. compared with the U.K.
Reliable estimation of differences in cost required consideration of many additional important variables that contribute to drug costs among countries. In the U.S. MarketScan data source, payment for drugs came from private insurance plans with variable drug payment coverage. By contrast, in the U.K., payment came regularly from the government's National Health Service and was generally stable in a given year.
It was evident that it would not be feasible to achieve a reliable and useful estimate of relative drug costs for a large heterogeneous group of prescription drugs. Therefore, to mitigate the effects of the complex interactive elements, we restricted the analysis to a relatively homogeneous group of people who were similar in age (55–64 years) and of the same sex, and derived from similar general medical communities. To avoid the complexities of occasional changes in drug usage over time, we further restricted the study to 1 year, 2005. Finally, in this cost analysis, we restricted the study population to those who were prescribed statins continuously for the entire year. The applied restrictions thus limited the findings to a selected but well-defined segment of the two populations. At the same time, this process allowed us to provide a reliable and consistent estimate of the drug-specific cost/pill that could be applied to also estimate aggregated costs for people who used statins intermittently during 2005.
Statins were prescribed to 32.7% of the U.S. study population and 24.4% of the U.K. population. The estimated annual cost of statins according to country is a reflection of the relative proportion of people who are prescribed the drug multiplied by the cost/pill. In the U.S. population of continuous annual users, the cost ($64.9 million) was estimated to be more than 400% higher than the cost in comparable statin users in the U.K. ($15.7 million). It should be noted that although the cost of statins is substantially higher in the U.S. than in the U.K., an increase in cardiovascular health benefits from higher usage, although uncertain, could be substantially larger as well.
The study design provided strict control for age, sex, calendar time, and duration of use. Differences in these characteristics between countries cannot explain the vast differences in cost. Electronic recording of statin prescriptions is standardized and appears to be of high quality and completeness in both resources.
In addition to differences in overall statin use and per unit costs, another significant factor contributing to the disparity in costs appears to be the availability and utilization of generics. As mentioned above, only generic lovastatin, which was used by just 5% of statin users, was available in the U.S. in 2005. By contrast, two drugs, simvastatin and pravastatin, prescribed to 48% of statin users, were available as generics in the U.K.
It is possible that some of the observed differences in costs for statins could be related to other underlying differences between countries, since the U.S. data related to people who were employed (and their families) and covered by private insurance companies, whereas the U.K. data included both employed and unemployed people whose drug costs were paid for by the government. Nevertheless, within both countries the cost of statins did not materially change throughout 2005, and, therefore, relative pill cost comparisons were steady for the selected populations included.
According to the census bureau, most people in the U.S.—roughly 200 million—were covered by private health insurance for drug costs in 2005. However, for the military, veterans, government employees, the elderly, and other groups, the cost of prescription drugs is at least partly paid by the government. Estimates of drug costs with similar accuracy, transparency, and specificity paid by the government compared with the costs incurred from private insurance companies could provide a useful basis for further insight into the nature of the remarkable difference (at least for statins) in prescription drug costs between the U.S. and the U.K.
Large electronic medical databases are now ubiquitous. They can be and have been successfully applied to observational research in areas such as drug and vaccine safety and medical cost estimation. They can be a uniquely valuable source of knowledge provided that the recorded information is of high quality and the research is crafted with the necessary training and experience.
Pharmacotherapy. 2012;32(1):1-6. © 2012 Pharmacotherapy Publications