There is a growing recognition that emotional problems are important to physical health outcomes. In response, primary care clinics have introduced self-report checklists to identify patients with emotional disorders such as depression or anxiety. Yet psychodynamic theory posits that certain emotional problems may be unconscious and unspoken, and thus not discernible on self-report checklists, and studies show that checklists do not identify every patient who needs treatment. New clinical tools are needed to identify subtle and complex presentations. We aimed to develop an innovative mixed-methods approach characterizing different types of verbal expression of feelings, drawing on psychodynamic theory and empirical research. We outline the development of the mixed-methods approach, including our theoretical framework and use of semi-structured interview data from Partners in Care (PIC), a randomized controlled trial of quality improvement for depression. We then illustrate the approach with one case: an older female PIC participant who screened positive for depression on all study self-reports. The approach delineates three qualitatively different categories of words-specific feeling words, vague feeling words, and physical words-that were quantified to define a measurable pattern for our participant. Clinicians could be trained to identify these categories of words in the context of a discussion of feelings to better detect and understand subtle emotional problems in patients who have difficulty talking openly about their feelings. Next steps include furthering face and construct validity and test-retest reliability, examining the prevalence of these patterns in a larger sample, and assessing correlates of patterns.
