A Cross-Sectional Study of Online Survey Respondents' Knowledge and Attitudes Toward Delirium

Delirium affects approximately a quarter of hospitalized adults and is associated with higher mortality, cognitive decline, and prolonged hospitalizations. Patient and caregiver education may improve outcomes. However, limited literature exists on the general public's current understanding of this condition.

To assess knowledge and attitudes toward delirium via an online convenience sample.

Cross-sectional survey using Amazon's Mechanical Turk between October 2023 and January 2024.

Adults 18 years old and older, living in the USA, and fluent in English.

We assessed delirium knowledge using a 15-question knowledge assessment covering three categories (risk factors, symptoms, and management). We compared mean scores between knowledge categories using one-way analysis of variance (ANOVA). We used univariate and multivariate logistic regression models to identify demographics associated with low delirium knowledge scores.

Among 397 respondents, 52% were male, 67% were White, and the median age was 44.6 (IQR 36-52). Participants reported living in 43 states. Participants answered 59% (SD 14.8) of knowledge questions correctly; respondents performed highest within the risk factor category (74%, SD 19.6) and lowest in management (47%, SD 21.1), p = <0.01. Few knew that bedrest is not recommended for delirium and that symptoms do not develop gradually (18% and 25%, respectively). Most (83%) agreed that more patient and caregiver education on delirium is needed. In multivariate analysis, individuals with higher incomes were less likely to score in the lowest quartile compared to those earning under $50K (e.g., OR 0.60 for persons earning $50-100K, 95% CI 0.29-0.91). Finally, low self-confidence in delirium knowledge was not associated with low knowledge scores.

Respondents had misconceptions about delirium that could be addressed in future educational initiatives, including the importance of early mobilization. Further work is needed to characterize what types of knowledge are valued by patients and caregivers and are critical to facilitating engagement.