Forecasting Visit Completion in Low Literacy Patients with Telehealth Interface Usability

Abstract

Telehealth has emerged as a cornerstone of modern healthcare delivery, yet its efficacy is often compromised by design barriers that disproportionately affect low-literacy patient populations. This study investigates the relationship between telehealth interface usability and virtual visit completion rates among low-literacy patients using a mixed-methods evaluation framework. We conducted a mixed usability assessment involving one hundred and eighty-six participants categorized as having low health and functional literacy. Usability was measured using quantitative metrics, including task completion time, error rates, and System Usability Scale scores, alongside qualitative thematic analysis of retrospective think-aloud sessions. Predictive modeling, specifically logistic regression and random forest classification, was employed to determine which usability factors most strongly predict successful clinical visit completion. The results demonstrate that cognitive friction, navigation complexity, and icon ambiguity are major inhibitors of visit completion. The predictive model identified task completion time during onboarding and subjective cognitive load as the strongest predictors of overall visit completion. These findings suggest that tailoring telehealth user interfaces to the cognitive needs of low-literacy patients is critical to reducing health disparities and ensuring equitable access to digital medicine.