Evidence of drug efficacy comes mostly from studies of middle-aged populations with few comorbidities, but many changes in drug benefits policies heavily impact elderly patients who frequently receive multiple medications and have numerous comorbidities. The randomized controlled trial (RCT) is considered the gold standard among methods to assess the effectiveness of biomedical interventions. However, randomized health policy trials are rarely done because they are perceived to be methodologically difficult, politically problematic, and costly to implement. We are unaware of any study that studies the impact of drug benefits policy changes by comparing randomized versus nonrandomized controls. In March 1999, Pharrnacare, British Columbia's publicly funded drug benefits program for all residents over age 65, introduced a change in its drug coverage policy but exempted a randomly chosen control group of 10% of physicians. The specific policy change related to respiratory medications, but its importance lies in the fact that this drug class can be seen as representative of many other medication reimbursement policy changes. As changes in drug reimbursement become increasingly common in the elderly, we propose this pilot study to answer important methodological questions related to the quantification of outcomes resulting from such changes. We will compare the results of this randomized policy trial with the findings that would have been obtained from a more conventional pre-post evaluation without randomized controls. Outcomes to be studied will include changes in drug utilization, clinical outcomes such as physician visits, hospitalizations, overall health care expenditures, long-term care admissions, and mortality. To do this, we will make use of data available to us for the entire population of elderly people in British Columbia as well as validated quality-of-life questionnaires for a subsample of patients. Particular attention will be paid to the effects of age, nursing home residents and specific co- morbidities on these outcomes. The findings of this study will be useful in designing and interpreting future studies of drug reimbursement policy changes in older patients.
