Patient Interaction Phenotypes With an Automated SMS Text Message-Based Program and Use of Acute Health Care...

Area of Science:

• Health Informatics
• Patient Engagement
• Digital Health Communication

Background:

• Automated bidirectional SMS text messaging is a key strategy for post-hospital discharge patient-system communication.
• Understanding patient interaction patterns is crucial for optimizing messaging program design.
• Tailoring communication to individual patient needs can improve engagement and outcomes.

Purpose of the Study:

• To identify and characterize distinct patient interaction phenotypes within a postdischarge automated SMS text messaging program.
• To analyze the relationship between patient interaction phenotypes and demographic/clinical characteristics.
• To explore the association between interaction phenotypes and hospital revisit outcomes.

Main Methods:

• Secondary analysis of a randomized controlled trial involving a 30-day postdischarge SMS text messaging intervention.
• K-means clustering was used to identify patient interaction phenotypes based on engagement and conformity.
• Features were engineered to quantify patient engagement and adherence to the messaging program.

Main Results:

• Four distinct patient interaction phenotypes were identified: enthusiasts (high engagement, high conformity), minimalists (low engagement, high conformity), nonadapters (low engagement, low conformity), and high needs responders (high engagement, intense need).
• Significant differences in demographic characteristics (gender, race, insurance) and clinical outcomes were observed across these phenotypes.
• A total of 1731 patients were analyzed, with variations in engagement and conformity patterns.

Conclusions:

• Patient interactions with automated SMS text messaging post-discharge are diverse, necessitating tailored approaches or alternative communication methods for some individuals.
• Patient interaction phenotypes, beyond message content, may offer predictive value for adverse health outcomes.
• Health systems should consider individual patient interaction styles when implementing SMS messaging strategies.