Despite the existence of efficacious interventions for mental disorders, many people who suffer with disorders such as depression fail to seek treatment. This is particularly true for low-income and minority patients. Unfortunately, we have little understanding of why this is so. Studies of barriers to care generally find that structural barriers such as lack of insurance or access to providers only partially explain low rates of care. In this interdisciplinary study, we will focus on women who have not sought treatment for a current episode of major depression, but who have been identified as part of an ongoing study of treatment of depression in multi-ethnic, low-income gynecology patients. Our proposal integrates qualitative and quantitative methodologies using sophisticated discourse software to analyze recordings of evaluation and treatment sessions. The long-range goal of our work is to determine whether cultural beliefs and changes in those beliefs affect acceptance and success of depression treatment in low-income women. The objective of the study is to examine the relationship of explanatory models (EM) of depression and its treatment to clinical outcomes in a randomized clinical trial for major depressive disorder in disadvantaged women. We argue that cultural beliefs inform patient EMs and that initially, these EMs are divergent from clinician EMs. Where there is extreme divergence, we predict that patients will fail to engage in treatment. However, if the EMs of patients and providers begin to converge, we expect that patients will accept and remain in treatment. We will accomplish our objective, through two specific aims: 1) Provide careful and clear operational definitions of EMs in actual use between patients and providers and a clear set of measures for how these models change over time. 2) Examine the relationship of the convergence of EMs to clinical outcome variables of depression remission and adherence to treatment. To accomplish these aims, we will use videotaped data from evaluation and treatment sessions of study subjects enrolled in the treatment study described above. The videotapes will be digitized, edited and transcribed. Finally, using discourse analysis methodology, we will develop a coding system to reliably identify EMs and rate EM convergence. Statistical analyses will identify the relationship of EM convergence to severity ratings of depression and assessments of engagement in and adherence to treatment. The methodology in our study can serve as a prototype for the study of process of care issues in low -income, ethnically diverse populations.
