Assessing Beliefs About Depression and Its Treatment

This dissertation is motivated by the clinical problem of determining which of two comparably effective modes of treatment (pharrnacotherapy and psychotherapy) a depressed individual would be most likely to benefit from. A cognitive variable, illness beliefs about depression, is proposed as a promising guide for making such patient-treatment mode matches. A broad area of literature is reviewed in order to learn how beliefs about depression are and have been conceptualized, and to learn what methods, if any, have been previously used to measure them. The review begins with an examination of the prevailing cognitive paradigm and it\u27s failure to consider beliefs about depression as universalized, personal constructs that play an integral role in organizing depression-related behavior. The review includes sections on: beliefs and health and illness behavior, beliefs within the behavioral ecology of treatment for mental illness, the consequences of mental health beliefs and attributions, and beliefs as behavioral expectancies. A section on the social nature of cognition about depression highlights the conceptual grounding of the dissertation in an ethnomedical framework articulated by Kleinman (e.g., 1980). This anthropologically-influenced section also focuses heavily on the work of a British researcher, Rippere, who produced the sole significant body of work on depression- relevant beliefs which the author was able to uncover. The first part of the dissertation culminates in the statement of an original theoretical model in which illness beliefs, as the bearers of perceived illness controllability, are postulated to mediate response to depression. This mediation process is hypothesized to occur through the modulated utilization of highly individual ensembles of different kinds of personal resources for coping with and combatting depression. Paradigms of conscious and unconscious use of these antidepressive personal resources are delineated, and an extended illustration of four distinct sequences of response to the appraised controllability of depression is provided in order to amplify some of the clinical implications of the model. The second part of the dissertation involves the description of a psychometric project to construct a seven-point Likert-type agreement scale for measuring the ideological orientation of illness beliefs about the etiology and treatment of depression--the BADIT (Adult Survey of Beliefs About Depression and Its Treatment). Initial efforts at scale development included: a) a content validity study based on the ratings of six experts, and b) a multifaceted evaluation of the BADIT\u27s reliability based on the responses of a development sample of 198 college students. Following two major revisions, the final version of the BADIT consists The second part of the dissertation involves the description of a psychometric project to construct a seven-point Likert-type agreement scale for measuring the ideological orientation of illness beliefs about the etiology and treatment of depression--the BADIT (Adult Survey of Beliefs About Depression and Its Treatment). Initial efforts at scale development included: a) a content validity study based on the ratings of six experts, and b) a multifaceted evaluation of the BADIT\u27s reliability based on the responses of a development sample of 198 college students. Following two major revisions, the final version of the BADIT consists of four theoretical scales, having five items each. The scales (with their Cronbach internal consistency coefficients) are: Biomedical/Etiology (.72), Biomedical/Treatment (.71), Psychosocial/Etiology (.48), and Psychosocial/Treatment (.67). In order to test some hypotheses concerning the BADIT\u27s structural component validity (Loevinger, 1957), confirmatory maximum-likelihood factor analyses were performed using LISREL (Joreskog & Sorbom, 1986) to evaluate and contrast the fit of eight alternative models for the measurement of illness beliefs about depression. Although in relative terms the author\u27s theoretical four-factor model provides the best fit to the development sample data (X2 [164, N = 198] = 344.30), it is a substandard fit in absolute terms (GFI = .85). Selected results from the empirical study of the BADIT are discussed in relation to needs for further psychometric development, and comments are offered regarding the BADIT\u27s potential as both a clinical and research tool