Diffusion /Outcomes of Marketed /Unmarketed Cancer Drugs

University Of Pennsylvania, Philadelphia PA

Investigators

Abstract

[unreadable] DESCRIPTION (provided by the applicant): Understanding the roles that physician financial incentives and manufacturer promotion play in the diffusion and use of new technology is of central interest to policy makers. By using an innovative means to identify off label use in SEER-Medicare panel data with over 500,000 patients from 1991 to 2002, three specific research aims will be pursued in informing that understanding. [unreadable] [unreadable] First, the research will quantify the descriptive statistics of off-label use of anticancer drugs. Off-label drugs are those drugs prescribed for diagnoses other than those for which there is FDA-qualified evidence of safety and efficacy. Despite assumptions that fifty percent of anti-cancer uses are off-label, there has never been a valid descriptive quantification of their actual use by physicians on patients. Second, the candidate will address whether reimbursement for chemotherapy drugs influences their patterns of use, and if so, whether such use contributes to outcomes. Profit per chemo administration varies significantly across drugs, time, Medicare carrier jurisdiction, and product label. By exploiting variation across Medicare carriers' coverage policies and reimbursement levels for off-label chemotherapeutics in the 1990's, the candidate will test how financial incentives affected physician use and patient outcomes. Third, the candidate will address whether the prohibition on promoting medical technology for which there is unsatisfactory evidence of safety and efficacy from a regulatory perspective affects patterns of use and quality of care. The candidate will use a court ruling in 1999 that caused a relaxation of this policy to test how it affected off-label drug administrations and outcomes. [unreadable] [unreadable] Patterns of use in the second and third research aims will be estimated with a conditional logit discrete choice model where the dependent variable consists of all drug choices for a physician at a point in time. Outcomes will be estimated using a Cox proportional hazard model, where the dependent variable is patient mortality at a point in time. [unreadable] [unreadable] The overall objective of this research is to further our understanding of the benefits and costs of FDA and CMS policy that respectively restrict off-label detailing but allow for its reimbursement. This research has the capacity to inform policies dealing with product promotion regulation and physician financial incentives, especially in the war on cancer. [unreadable] [unreadable]