United Kingdom Data Source
The U.K. prescription data were derived from the General Practice Research Database (GPRD) constructed in 1990. The medical information contained in the database has been repeatedly validated and found to be of high quality and completeness for research purposes.[2,3,4] The GPRD contains continuous medical information derived from the electronic records of more than 3 million people enrolled in over 300 general practice outpatient offices throughout the U.K. Participating practices were initially selected to reflect the age, sex, and geographic distribution of the U.K. population. The general practitioners use computers with identical software in their routine practice and have been trained to record demographic data, medical diagnoses, and deaths in a standard format. All prescriptions (coded using Multilex codes [Multilex Product Dictionary used within GPRD data (GPRD, London, United Kingdom)]) are generated electronically, and the details of each prescription are automatically recorded into the patient record. This unique data resource has been used extensively for drug and vaccine utilization and safety research.[4,5,6,7,8]
United States Data Source
The U.S. prescription data were derived from a large, claims-based medical database (MarketScan Commercial Claims and Encounters Database; Thomson Reuters, Ann Arbor, MI). The database contains comprehensive medical information contributed by large self-insured employers distributed throughout the U.S., and a smaller number of health insurance plans. It includes longitudinal data originating from 1995 for about 100 million people younger than 65 years, including family members of employees. Demographic data are recorded, along with information about paid claims for drugs, medical services (with diagnoses), and procedures. Each drug claim includes information on the specific entity dispensed (coded using U.S. Food and Drug Administration National Drug Codes), the date of dispensing, the quantity dispensed, and the intended duration. Most important, it provides the payment for each dispensed supply of drug separately. Insurance company costs are regularly negotiated individually with drug companies. This data resource has previously been used in large targeted drug safety research,[9,10] as well as drug and health care utilization studies.[11,12]
In this matched-cohort cost analysis, we initially identified 1.6 million people younger than 65 years of age in the U.K. who were enrolled in the GPRD. Each person was then matched by year of birth and sex to a person chosen at random from the extensive U.S. files. From this matched pool, we estimated that 280,000 people aged 55–64 years from each country in 2005 were prescribed at least one drug. From this population, we determined the number who had been prescribed statins.
We derived two separate random samples of 100 people who used statins in this age group from each country and reviewed them directly to determine statin usage. Estimated costs in the U.S. were based on a direct individual review of the two separate random samples of 100 patient records. The U.S. drug costs were derived directly from the individual patient's prescription record entered by MarketScan, which contained data on the total amount paid for the prescription, including the ingredient cost, patient contribution (i.e., copayment or coinsurance charges), dispensing fees, and applicable discounts. The average annual cost/patient was calculated as the average cost of 30-day and 90-day prescriptions in the random samples for each individual drug-dose combination multiplied over a 12-month period. We then applied these estimates to the total number of annual users for each drug in the sample to obtain a total annual cost.
For the U.K., we obtained a cost/pill (exclusive of applicable discounts) from the 2005 Prescription Cost Analysis report published by the National Health Service. This series of publicly available annual reports provides data on all prescriptions written by general practitioners in England for each drug preparation over a given year, including aggregate ingredient costs and the total number of pills dispensed. The cost/pill of the most commonly prescribed preparation of each drug-dose combination was selected for inclusion in the analysis. Dispensing and container fees, estimated at a per-pill cost (~U.S. $0.03), obtained from the May 2011 National Health Service Drug Tariff, were not included. Patient copayments, although instituted by the National Health Service for prescriptions dispensed in the pharmacy setting, were excluded from the cost calculations since nearly 90% of patients were exempted from paying them in 2005. Since statins are virtually always prescribed as one pill/day, we estimated the average annual cost/patient by multiplying the average cost/pill for each drug by 365.
Costs for the U.K. were converted to U.S. dollars by using the 2005 Organization for Economic Co-operation and Development gross domestic product purchasing power parity benchmark (U.S. = 1.0, U.K. = 0.636). All monetary units are presented in 2005 U.S. dollars.
Pharmacotherapy. 2012;32(1):1-6. © 2012 Pharmacotherapy Publications