Through the Patient's Eyes: Application of GoProEmo Technology to Improve Resident Clinical Interviewing Skills in 2030.

A long-standing problem in graduate medical education has been the inability to provide residents with accurate, timely, and actionable feedback based on authentic patient encounters. Opportunities are limited for attending physicians to directly observe and critique residents' clinical encounters; in addition, these encounters rely on the attending's perception of the interaction, which often lacks the patient's input or perspective. Thus, due to limitations in the quantity and quality of direct observations, these encounters may not accurately reflect the resident's overall performance, and often can't be more broadly generalizable. In this setting, we chose to study the impact of the audio-visual emotional assessment tool, GoProEmo, to enhance the quality and quantity of feedback provided to residents.In 2030 a total of 160 internal medicine residents (postgraduate years 1–3) at a university-affiliated tertiary care academic medical center were enrolled. Residents wore the GoProEmo body camera for all clinical encounters on a general medicine ward rotation over 4 weeks. Patients consented to wearing the GoProEmo receptor patch, which synced to the camera, on admission to the hospital.The GoProEmo technology uses a novel assessment tool known as body language coding to evaluate patient emotions. Patient variables include heart rate, blood pressure, cortisol level in perspiration, facial expression, and muscle tension. These factors are analyzed continuously and recorded for later review as a “stoplight” indicator (green, activated/engaged; yellow, confused/emotionally tense; red, emotionally distressed). On playback mode, the stoplight indicator is synced to the audiovisual recording in real time, allowing for pinpoint identification of emotionally activated moments in the interview and resident responses to those moments. This intervention provides opportunities for objective feedback about the learner's ability to adapt to patient emotions or confusion, and for learner reflection regarding opportunities for improvement of their clinical interviewing skills. The percentages of green, yellow, and red interactions per encounter were also tracked for each learner over a 4-week rotation.The primary outcome was the percentage of emotionally concordant learner responses to interactions that were coded either yellow or red in week 4 compared to week 1. We found learners had a significantly increased percentage of concordant red interactions at the conclusion of the intervention compared to week 1 (46% versus 88%, P < .001). Yellow interactions also showed increased concordance at week 4 (28% versus 64%, P < .001).The secondary outcome was the overall percentage of green, yellow, and red interactions for each learner. We noted significantly fewer yellow interactions in week 4 compared to week 1, with a concomitant increase in green interactions (P < .001 for each). Exploratory outcomes that we deemed hypothesis-generating included analyses of heart rate, blood pressure, cortisol levels, and their relationship to the “color” of the interaction.The GoProEmo technology is proprietary to our hospital and was approved for use in this study by the Institutional Review Board committee. Development of the GoProEmo receptor patch was outsourced to a biomedical technology company funded by the study grant. A total of 100 patches were made at a cost of $20 per patch and are reusable.Acceptability of the intervention was assessed at the patient and provider levels using focus groups. Patients found that the implementation of personal visual recording devices improved communication with providers and enhanced the perceived level of compassion. Privacy concerns were addressed during the consent process, and data were permanently erased after review. Learners were initially concerned about being recorded, but focus groups at the end of the study revealed that, overall, residents appreciated the improved feedback quality.