Potential Effectiveness of the InterventionNo significant differences between the means of the first andsecond self-assessments were found for depressive (t8=0.40;P=.70) or anxiety (t8=1.00; P=.34) symptoms. This lack ofsignificant difference was maintained for the thirdself-assessment. Work-related burnout scores were significantlylower after 2 weeks of using the Vickybot (z=−2.07; P=.04),with a moderate effect size (r=0.32; Table 3).To assess the effect of baseline symptoms, we conductedsubanalyses of participants with mild-to-moderate depression(PHQ-9, 5-14) and anxiety (GAD-7, 5-14) symptoms. Nosignificant differences were found in depressive (t4=0.33; P=.75)or anxiety (t4=0.00; P=.99) symptoms between the means ofthe first and second self-assessments.To assess whether the effect of the intervention was influencedby the use of the chatbot (dose-related effects of theintervention), we correlated the number of times that moduleswere performed (active treatment) with the change in the clinicalscales’ scores. No significant associations were found fordepressive symptoms (r=−0.48; P=.16), anxiety symptoms(r=0.12; P=.76), or work-related burnout (r=0.55; P=.13). Wealso correlated the “days using the chatbot” with the change inthe clinical scales’scores, and no significant associations werefound for depressive symptoms (r=−0.23; P=.58), anxietysymptoms (r=−0.41; P=.36), or work-related burnout (r=0.38;P=.39). Despite the nonsignificant results, there was a trendtoward a reduction in anxiety and depressive symptoms withgreater chatbot use. To assess the effect of “clinically significantdoses” of the intervention, we conducted subanalyses of patientswith ≥50% (moderate to high) engagement. No significantdifferences between the scores of the first and secondself-assessments were found for depressive symptoms (t6=−0.82;P=.44), anxiety symptoms (t6=0.33; P=.75), or work-relatedburnout (z=−1.89; P=.06)