Time Series Modeling in Medication Prescribing -- Scientific Vs. Promotional Influences
Principal Investigator: Randall S. Stafford, MD, PhD
Funding Agency: Agency for Healthcare Research and Quality
Physician decisions regarding medication prescribing have a large impact on health care costs and clinical outcomes. Surprisingly little is known, however, about the relative impact of a range of factors that may influence physician behavior in the prescribing of medications. By extension, the ability to explain and project trends in medication prescribing is limited. This study will develop a model of the simultaneous influence of multiple determinants on trends in medication prescribing. Time series regression techniques will be used to estimate the relative influence of temporally varying predictor variables, as well as discrete temporal events, on observed national changes in the prescribing of pharmaceuticals for asthma, depression, and hyptertension. Information on trends in drug prescribing in these conditions from 1978 through 2001 will come from the National Disease and Therapeutic Index survey available through IMS HEALTH (Plymouth Meeting PA). Data on potential predictors of trends will be developed from a variety of sources. By constructing a comprehensive model of the temporal determinants of prescribing behavior, this study will assess the relative influence of clinical guidelines, clinical trial results, pharmaceutical promotion, drug prices and specific competitive features of the pharmaceutical market. Time-series regression modeling will be employed to test specific hypotheses regarding the role of these factors. Results from our analysis will be relevant to efforts to design and implement clinical and health policies to encourage the effective and cost-effective use of medications.
View a list of recent publications and abstracts from the Time Series project.