Abstract

Background: Evidence from VA and non-VA settings demonstrates widespread racial disparities in healthcare delivery. Our prior work suggests that providers with higher levels of “cultural competence” (CC) deliver more equitable care. But how CC translates into better care is unclear. We will use data from our current project, “Opening the Black Box of Cultural Competence” (aka Black Box), in which we are analyzing communication content from audio recorded primary care visits. We will complement these content analyses with computerized linguistic analysis methods to generate and test novel measures of patient-provider communication and examine their role in disparities in patient-provider relationship quality. Significance/Impact: Delivering high-quality care to all Veterans is central to VA's mission. We will address a previously unexplored area that represents a potential target for reducing disparities in VA care. Our study also addresses several VA HSR&D priority areas, including promoting health equity, improving primary care practice, and advancing data science. Innovation: Most studies of healthcare communication have focused on communication content (i.e., what is said). By examining linguistic style (i.e., how things are said), we will address a relatively unexplored potential source of racial disparities. In addition, by applying computerized, natural language processing (NLP) methods to evaluate patient-provider communication, this study will develop and test potentially scalable tools and metrics that can be implemented to provide real-time feedback to improve patient-provider interactions as part of a learning health system striving to improve the delivery of high-quality, equitable care. Specific Aims: 1) Apply computerized text analysis tools to transcripts of primary care visits to generate measures of patient and provider linguistic style and style matching (LSM). 2) Test associations of: a) Veteran race and provider CC with LSM and provider linguistic style; and b) LSM and provider linguistic style with the quality of patient-provider relationships. 3) Qualitatively explore examples of visits with high and low LSM and with provider linguistic style patterns associated with high and low relationship quality. Methodology: In the Black Box study, we are analyzing communication content, using the Roter Interaction Analysis System, directly from the audio files of 408 primary care visits at 4 geographically diverse VA medical centers. In the proposed project, we will transcribe the audio files and apply computerized, dictionary-based lexical analysis tools to evaluate functional and semantic speech patterns and LSM between patient and provider. We will test the associations described in Aim 2 using patient and provider survey data collected in the parent study. Finally, we will qualitatively review selected transcripts to evaluate the mechanisms by which LSM, and provider linguistic styles associated with relationship quality, are achieved. Implementation/Next Steps: This pilot study is designed to develop novel methods and measures rather than lead directly to a larger intervention study. The VHA Office of Health Equity (OHE), our current partner on the Black Box study, will be our primary operations partner. We will also engage the Office of Patient- Centered Care and Cultural Transformation. We will review our findings with these stakeholders to plan next steps in translating our findings into improvements in patient-provider communication quality and equity.

Dimensions for VA

Dimensions for VA is a web-based tool available to VA staff that enables detailed searches of published research and research projects.

Learn more about Dimensions for VA.

VA staff not currently on the VA network can access Dimensions by registering for an account using their VA email address.     Search Dimensions for this project

DRA: Health Systems Science

DRE: Research Infrastructure, TRL - Applied/Translational, Technology Development and Assessment

Keywords: Cultural Competence, Disparities, Natural Language Processing, Patient-Provider Interaction, Research Tools, Socioeconomic Factors

MeSH Terms: None at this time.