This project addresses health disparities by investigating a novel computational approach that makes implicit, thus hidden, bias visible. Healthcare bias, based on patients? race, gender, socioeconomic status, sexual orientation, and other characteristics lead to health disparities. Such biases are often unintentional and hidden in communication among clinicians and patients. Although there is broad agreement that healthcare biases need to be better understood, assessed and mitigated, traditional clinical communication training and assessment is removed from actual patient-provider interactions in which bias hides. To mitigate health disparities, we propose social signal processing (SSP) technology that automatically assesses hidden bias during patient encounters. SSP involves machine analysis and feedback on subtle cues (e.g., talk time, interruptions, body movement) that reflect the quality of communication. This technology will automatically capture nonverbal, linguistic, and affective, cues in patient-provider interactions and then provide feedback for improvement, designed in collaboration with patients and providers. Guided by human-centered design, we will engage low income, racially diverse patients and their providers to inform the design of visual feedback from SSP assessment, and then evaluate the efficacy of this novel technology in both simulated and real world encounters. Leveraging our preliminary work and multidisciplinary expertise from our two investigative sites, University of Washington and University of California San Diego, we will partner with academic and community health clinics at both sites to engage underserved patients and providers in three specific aims to: build an SSP model that characterizes communication quality among clinicians and health disparity patients (Aim 1), design SSP feedback that conveys hidden bias to patients and providers (Aim 2), and evaluate the efficacy of SSP technology in controlled and real world clinical settings (Aim 3). Findings will bring insight into social signals associated with hidden bias that we can automatically detect during patient visits, design recommendations for effective SSP feedback for both providers and patients, and evidence on the technical validity and efficacy of SSP technology for improving patient and provider experience of patient-centered care. To mitigate health disparities, patients and providers need unbiased interactions. This project will contribute a novel computational paradigm using SSP that brings human-centered visibility to implicit biases that manifest in healthcare communication and lead to health disparities. Findings will advance biomedical informatics and health disparities research with a novel SSP approach for the next generation of healthcare providers and educators, empower health disparity patients, and promote healthcare quality and equity. Bringing visibility to hidden bias though human-centered SSP has significant promise, particularly in healthcare where patient-centered communication is critical to building rapport, establishing trusted patient-provider relationships, promoting equity, and ultimately mitigating health disparities.

Public Health Relevance

This project investigates a novel social signal processing (SSP) approach to help reduce health disparities by improving patient-clinician communication for low income, racially diverse patients in primary care. In partnership with academic and community-based health systems in Seattle and San Diego, our multidisciplinary team will expand their preliminary work to build an SSP model that characterizes communication quality among clinicians and health disparity patients (Aim 1), design SSP feedback interventions that convey hidden communication signals to health disparity patients and providers (Aim 2), and evaluate the efficacy of SSP interventions through controlled simulations and in underserved clinics (Aim 3). Findings will lead to a novel SSP approach for the next generation of healthcare providers and educators, empower health disparity patients, and promote healthcare access, quality, and equity.

Agency
National Institute of Health (NIH)
Institute
National Library of Medicine (NLM)
Type
Research Project (R01)
Project #
5R01LM013301-02
Application #
10021722
Study Section
Special Emphasis Panel (ZRG1)
Program Officer
Vanbiervliet, Alan
Project Start
2019-09-20
Project End
2024-07-31
Budget Start
2020-08-01
Budget End
2021-07-31
Support Year
2
Fiscal Year
2020
Total Cost
Indirect Cost
Name
University of Washington
Department
Other Health Professions
Type
Schools of Medicine
DUNS #
605799469
City
Seattle
State
WA
Country
United States
Zip Code
98195