OBJECTIVE: Clinician implicit bias can impede patient-centered communication, leading to health care inequities. While the field of implicit bias education is evolving with advances in technology, clinicians' perspectives remain underexplored. This study investigated clinicians' perceptions of educational strategies to complement communication feedback technology in the implementation of an implicit bias education intervention. MATERIALS AND METHODS: We recruited primary care practitioners in remote interviews to brainstorm future technologies for improving clinician awareness of implicit bias in patient-provider communication. Participants completed an online survey in which they rated the priority of educational strategies that could complement the technology. We performed inductive-deductive thematic analysis of the interview data with Implicit Bias Recognition and Management (IBRM) domains as a priori codes and used descriptive statistics to summarize the survey data. RESULTS: Participants (n = 16) proposed how future technology could improve clinician awareness, such as recording visits to help clinicians be more self-aware of their communication; however, some providers expressed concerns regarding feedback fatigue and the potential impact of technology on reducing time spent with patients. Participants recommended incorporating feedback regularly into training, identifying organizational incentives, and debriefing with trusted colleagues and communication experts. DISCUSSION: Participants brainstormed technologies and identified educational strategies, such as discussion with a facilitator, that could promote clinician receptivity to feedback and inform IBRM approaches for clinical ambient intelligence. Yet, challenges remain to incentivizing participation for practicing clinicians, and Continuing Medical Education may be one effective approach. CONCLUSION: The proposed technologies and prioritized educational strategies have the potential to promote health equity by helping clinicians develop skills to manage implicit bias. In the future, these findings could inform IBRM interventions that leverage clinical ambient intelligence.

OBJECTIVE: Artificial Intelligence (AI)-based approaches for extracting Social Drivers of Health (SDoH) from clinical notes offer healthcare systems an efficient way to identify patients' social needs, yet we know little about the acceptability of this approach to patients and clinicians. We investigated patient and clinician acceptability through interviews. MATERIALS AND METHODS: We interviewed primary care patients experiencing social needs (n = 19) and clinicians (n = 14) about their acceptability of "SDoH autosuggest," an AI-based approach for extracting SDoH from clinical notes. We presented storyboards depicting the approach and asked participants to rate their acceptability and discuss their rationale. RESULTS: Participants rated SDoH autosuggest moderately acceptable (mean = 3.9/5 patients; mean = 3.6/5 clinicians). Patients' ratings varied across domains, with substance use rated most and employment rated least acceptable. Both groups raised concern about information integrity, actionability, impact on clinical interactions and relationships, and privacy. In addition, patients raised concern about transparency, autonomy, and potential harm, whereas clinicians raised concern about usability. DISCUSSION: Despite reporting moderate acceptability of the envisioned approach, patients and clinicians expressed multiple concerns about AI systems that extract SDoH. Participants emphasized the need for high-quality data, non-intrusive presentation methods, and clear communication strategies regarding sensitive social needs. Findings underscore the importance of engaging patients and clinicians to mitigate unintended consequences when integrating AI approaches into care. CONCLUSION: Although AI approaches like SDoH autosuggest hold promise for efficiently identifying SDoH from clinical notes, they must also account for concerns of patients and clinicians to ensure these systems are acceptable and do not undermine trust.

Healthcare providers' implicit bias, based on patients' physical characteristics and perceived identities, negatively impacts healthcare access, care quality, and outcomes. Feedback tools are needed to help providers identify and learn from their biases. To incorporate providers' perspectives on the most effective ways to present such feedback, we conducted semi-structured design critique sessions with 24 primary care providers. We found that providers seek feedback designed with transparent metrics indicating the quality of their communication with a patient and trends in communication patterns across visits. Based on these metrics and trends, providers want this feedback presented in a dashboard paired with actionable, personalized tips about how to improve their communication behaviors. Our study provides new insights for interactive systems to help mitigate the impact of implicit biases in patient-provider communication. New systems that build upon these insights could support providers in making healthcare more equitable, particularly for patients from marginalized communities.

Patient-provider communication influences patient health outcomes, and analyzing such communication could help providers identify opportunities for improvement, leading to better care. Interpersonal communication can be assessed through "social-signals" expressed in non-verbal, vocal behaviors like interruptions, turn-taking, and pitch. To automate this assessment, we introduce a machine-learning pipeline that ingests audio-streams of conversations and tracks the magnitude of four social-signals: dominance, interactivity, engagement, and warmth. This pipeline is embedded into ConverSense, a web-application for providers to visualize their communication patterns, both within and across visits. Our user study with 5 clinicians and 10 patient visits demonstrates ConverSense's potential to provide feedback on communication challenges, as well as the need for this feedback to be contextualized within the specific underlying visit and patient interaction. Through this novel approach that uses data-driven self-reflection, ConverSense can help providers improve their communication with patients to deliver improved quality of care.

Identifying patients' social needs is a first critical step to address social determinants of health (SDoH)-the conditions in which people live, learn, work, and play that affect health. Addressing SDoH can improve health outcomes, population health, and health equity. Emerging SDoH reporting requirements call for health systems to implement efficient ways to identify and act on patients' social needs. Automatic extraction of SDoH from clinical notes within the electronic health record through natural language processing offers a promising approach. However, such automated SDoH systems could have unintended consequences for patients, related to stigma, privacy, confidentiality, and mistrust. Using Floridi et al's "AI4People" framework, we describe ethical considerations for system design and implementation that call attention to patient autonomy, beneficence, nonmaleficence, justice, and explicability. Based on our engagement of clinical and community champions in health equity work at University of Washington Medicine, we offer recommendations for integrating patient voices and needs into automated SDoH systems.