Pharmaceutical Detailing and Medication Choice: A Micro-Level Analysis of Statins
Posted: 22 Jun 2007
Date Written: June 18, 2007
Introduction: Pharmaceutical marketing expenditures have grown dramatically over the last decade, yet the effects of marketing on drug choice at the physician level is not well understood. Previous research has tried to answer this question using either aggregated national data and limited physician-level data. In contrast, I construct a unique dataset by combining patient-level prescribing data with pharmacy claims data to better measure the effects of detailing and other health plan characteristics, like copayments and formulary position, on drug choice. This research has important implications for understanding the effects of pharmaceutical detailing, which are becoming increasingly important with the introduction of Medicare Part D and for policymakers examining pharmaceutical regulation.
Objective: The objective of this paper is to estimate the marginal impact of pharmaceutical detailing on drug choice after controlling for other channels, such as copayment and formulary position, and patient and physician characteristics.
Method: Physician detailing data and statin prescriptions (135,947) were collected from a nationally representative sample of 2,442 physicians, and combined with pharmacy prescription data containing copayments and formulary position. The effects of promotion, copayments, physician characteristics, patient characteristics, and drug characteristics on statin choice was estimated using a conditional logit model. To account for detailing being potentially endogenous, a two-stage model was used where the first stage estimated the probability of detailing as a function of total statin prescriptions written by the physician and their location.
Results: The marginal effects of promotion show that increasing detailing by 3 visits (standard deviation) increases the probability of drug choice by 4.6%. The effects ranged from 1.4% to 11.1% based on the specific drug. Marginal detailing effects are increased for newly diagnoses patients (7.45%) and for patients switching to new therapy (6.69%).
Conclusion: These results show that pharmaceutical detailing is an important factor in drug selection and that the effect varies by type of patient and health plan. Promotion increased the probability of newly diagnosed patients receiving older statins, while patients switching from a previous statin were more likely to be prescribed newer statins. Copayments appear to be important in drug selection, but only among non-generic statins. The level of pharmaceutical detailing, in combination with copayment and formulary position, has a strong impact on the choice of drug prescribed to a patient.
Keywords: Pharmaceuticals, Marketing, Drug Choice
Suggested Citation: Suggested Citation