ACCRUAL TO THE BREAST-CANCER PREVENTION TRIAL BY PARTICIPATING COMMUNITY CLINICAL ONCOLOGY PROGRAMS - A PANEL-DATA ANALYSIS

In 1992 patient accrual to the National Cancer Institute-sponsored Bre ast Cancer Prevention Trial was initiated in the United States and Can ada. The Trial will involve 16,000 women who are evaluated to be at hi gh risk of developing breast cancer, Nearly 250 health care organizati ons are participating in the Trial, including over 40 Community Clinic al Oncology Program (CCOP) organizations which are a I component of th e NCI's national clinical trials program. A previous NCI-funded evalua tion conducted by the University of North Carolina showed the CCOP pro gram to be an effective means of transferring the latest cancer techno logy, particularly cancer treatments, to the community setting. This p aper describes a study designed to evaluate the performance of CCOP or ganizations in the Breast Cancer Prevention Trial. Using data from the first fifteen months of the Trial, the ability of CCOPs to accrue wom en is assessed using panel data estimation techniques. An attempt is m ade to predict accrual by structural, process, and environmental chara cteristics of participating CCOPs. Factors predictive of accrual inclu de month in which accrual occurred and the extent of competition for t rial participants in the CCOP service area. The hypothesized model exp lains slightly over 19 percent of the variation in accrual performance . The analysis demonstrates the utility of a panel data approach to mo deling the dynamics of CCOP participation in a chemoprevention clinica l trial.