(406) AI Chatbots in the PICU: Parental Enthusiasm Contrasts With Socioeconomic Usage Disparities

Introduction: Parents of children in the Pediatric Intensive Care Unit (PICU) are often overwhelmed by complex medical information and turn to online resources to alleviate stress and anxiety. While internet resources provide generic information, Large Language Models (LLMs) can offer accurate and patient-specific information. We aimed to describe health information-seeking behavior, AI chatbot knowledge, and attitudes toward healthcare-focused AI chatbots among parents in the PICU, hypothesizing that they would have high information needs and interest in AI chatbots.

Methods: We conducted a single-center observational study, surveying English-speaking parents of children hospitalized in the PICU from April-July 2024. We assessed participant demographics, healthcare information seeking, AI knowledge, and views on AI chatbots using a 6-point Likert scale. We used ordinal logistic regression to compare differences in socio-demographic factors and attitudes towards AI chatbots.

Results: Of 240 eligible parents, 134 completed the survey. Respondents were predominantly female (76%), aged 25-44 years (75%), white (66%), had less than a college degree (57%), and earned <$100k annually (72%). Most parents used search engines (86%), while only 9% used AI chatbots to find information about their child’s condition. Parents felt healthcare-focused AI chatbots could provide accurate information (median 4/6, IQR 3-4), address medical complexity (4, IQR 2-4), improve communication (4, IQR 3-5), and if given the opportunity, would likely use them (4, IQR 3-5). Male parents (OR 3.4, 95% CI 1.6-7.4, P< 0.01), those with higher education (at least some college vs high school or less; OR 2.4, 95% CI 1.1-5.2, p=0.03) and those with higher income (>$50,000 vs <$50,000; OR 2.8, 95% CI 1.4-5.7, p< 0.01), were more likely to have used AI chatbots.

Conclusions: Despite low current use, parents expressed positive attitudes toward AI chatbots’ potential to explain medical information in the PICU. Significant disparities in AI chatbot use were reported, but interest in future use of healthcare-focused AI resources was consistent across demographics. Future research should investigate AI chatbots' potential in supporting education in ICU settings and address reasons for usage disparities.