Emotional expressions associated with therapeutic inertia in multiple sclerosis care

Background: Emotions play a critical role in our daily decisions. However, it remains unclear how and what sort of emotional expressions are associated with therapeutic decisions in multiple sclerosis (MS) care. Our goal was to evaluate the relationship between emotions and affective states (as captured by muscle facial activity and emotional expressions) and TI amongst neurologists caring for MS patients when making therapeutic decisions. Methods: 38 neurologists with expertise in MS were invited to participate in a face-to-face study across Canada. Participants answered questions regarding their clinical practice, aversion to ambiguity, and the management of 10 simulated case-scenarios. TI was defined as lack of treatment initiation or escalation when there was clear evidence of clinical and radiological disease activity. We recorded facial muscle activations and their associated emotional expressions during the study, while participants made therapeutic choices. We used a validated machine learning algorithm of the AFFDEX software to code for facial muscle activations and a predefined mapping to emotional expressions (disgust, fear, surprise, etc.). Mixed effects models and mediation analyses were used to evaluate the relationship between ambiguity aversion, facial muscle activity/emotional expressions and TI measured as a binary variable and a continuous score. Results: 34 (89.4%) neurologists completed the study. The mean age [standard deviation (SD)] was 44.6 (11.5) years; 38.3% were female and 58.8% self-identified as MS specialists. Overall, 17 (50%) participants showed TI in at least one case-scenario and the mean (SD) TI score was 0.74 (0.90). Nineteen (55.9%) participants had aversion to ambiguity in the financial domain. The multivariate analysis adjusted for age, sex and MS expertise showed that aversion to ambiguity in the financial domain (OR 1.56, 95%CI 1.32–1.86) was associated with TI. Most common muscle activations included mouth open (23.4%), brow furrow (20.9%), brow raise (17.6%), and eye widening (13.1%). Most common emotional expressions included fear (5.1%), disgust (3.2%), sadness (2.9%), and surprise (2.8%). After adjustment for age, sex, and physicians’ expertise, the multivariate analysis revealed that brow furrow (OR 1.04; 95%CI 1.003–1.09) and lip suck (OR 1.06; 95%CI 1.01–1.11) were associated with an increase in TI prevalence, whereas upper lip raise (OR 0.30; 95%CI 0.15–0.59), and chin raise (OR 0.90; 95%CI 0.83–0.98) were associated with lower likelihood of TI. Disgust and surprise were associated with a lower TI score (disgust: p < 0.001; surprise: p = 0.008) and lower prevalence of TI (OR_disgust: 0.14, 95%CI 0.03–0.65; OR_surprise: 0.66, 94%CI 0.47–0.92) after adjusting for covariates. The mediation analysis showed that brow furrow was a partial mediator explaining 21.2% (95%CI 14.9%-38.9%) of the association between aversion to ambiguity and TI score, followed by nose wrinkle 12.8% (95%CI 8.9%-23.4%). Similarly, disgust was the single emotional expression (partial mediator) that attenuated (-13.2%, 95%CI -9.2% to -24.3%) the effect of aversion to ambiguity on TI. Conclusions: TI was observed in half of participants in at least one case-scenario. Our data suggest that facial metrics (e.g. brow furrow, nose wrinkle) and emotional expressions (e.g. disgust) are associated with physicians’ choices and partially mediate the effect of aversion to ambiguity on TI.