Provider compliance with treatment guidelines has traditionally been depicted as an issue primarily concerning the clinical behavior of physicians. The models often invoked to study or intervene in the treatment patterns of physicians are individual-based, and focus on various social-psychological mechanisms related to learning or peer- pressure. However, the factors which influence compliance with recommended treatment guidelines in fact constitute a complex web of psychological, interpersonal, organizational, and community-based variables. The guidelines for treatment of early stage (I or II) breast cancer in older women have been clearly and consistently stated since 1985. In brief, recommended primary treatment has been breast conserving surgery (BCS) plus radiation therapy (RT), or modified radical mastectomy. Nonetheless, geographic and hospital variation in the use of BCS, and a high portion of women, particularly older women, receiving BCS without radiation therapy have been reported in the literature. Previous studies of compliance with these treatment guidelines may have been hampered by the difficulty of determining whether radiation therapy was received following BCS, due to the variety of provider locations which provide RT and the different Medicare claims files in which these data could be found. We propose to link a number of extant data files with macro-level contextual data (created indices on linkage to NCI Community Clinical Oncology Programs (CCOP), NCI Comprehensive Cancer Centers (CCC) and radiotherapy research networks; AHA data; Area Resource File and Census data) to a sample of older women with early stage cancer from the SEER- Medicare linked data bases, in order to develop a national sample of such patients with information on their treatments and community settings. Our major interest is to test for the influence of linkage of the primary treatment sites (surgery and/or RT) to research networks, controlling for patient characteristics, other organizational variables, and community context, on treatment compliance. We propose a hierarchical polytomous regression model to estimate the effects of 3 levels of covariates on treatment received. This research will advance our understanding of the context in which compliance with treatment guidelines occurs. In particular, this study will provide important information on the interaction of context (organization and community) and clinical decision-making for aging women with breast cancer- a population that is seldom studied in the compliance literature.
| Laliberte, Linda; Fennell, Mary L; Papandonatos, George (2005) The relationship of membership in research networks to compliance with treatment guidelines for early-stage breast cancer. Med Care 43:471-9 |
