# Human cues in eHealth to promote lifestyle change: An experimental field study to examine adherence to self-help interventions

## TL;DR

This study explores how adding human-like features to a digital health app affects user adherence to lifestyle interventions.

## Contribution

The study experimentally examines the effects of visual and relational human cues in text-based conversational agents on user adherence and working alliance.

## Key findings

- Using human cues in a conversational agent did not improve the working alliance with users.
- A strong working alliance was positively linked to higher adherence to the intervention.
- Visual cues decreased adherence, possibly due to confusion about the agent's nature.

## Abstract

eHealth lifestyle interventions without human support (self-help interventions) are generally less effective, as they suffer from lower adherence levels. To solve this, we investigated whether (1) using a text-based conversational agent (TCA) and applying human cues contribute to a working alliance with the TCA, and whether (2) adding human cues and establishing a positive working alliance increase intervention adherence. Participants (N = 121) followed a TCA-supported app-based physical activity intervention. We manipulated two types of human cues: visual (ie, message appearance) and relational (ie, message content). We employed a 2 (visual cues: yes, no) x 2 (relational cues: yes, no) between-subjects design, resulting in four experimental groups: (1) visual and relational cues, (2) visual cues only, (3) relational cues only, or (4) no human cues. We measured the working alliance with the Working Alliance Inventory Short Revised form and intervention adherence as the number of days participants responded to the TCA's messages. Contrary to expectations, the working alliance was unaffected by using human cues. Working alliance was positively related to adherence (t(78) = 3.606, p = .001). Furthermore, groups who received visual cues showed lower adherence levels compared to those who received relational cues only or no cues (U = 1140.5, z = −3.520, p < .001). We replicated the finding that establishing a working alliance contributes to intervention adherence, independently of the use of human cues in a TCA. However, we were unable to show that adding human cues impacted the working alliance and increased adherence. The results indicate that adding visual cues to a TCA may even negatively affect adherence, possibly because it may create confusion concerning the true nature of the coach, which may prompt unrealistic expectations.

•Effects of human cues are rarely studied in text-based conversational agents (CA).•Incorporating human (relational and visual) cues did not affect the working alliance.•Yet, a good established working alliance contributed to higher adherence levels.•Adding visual cues decreased adherence, possibly due to confusion about the CA's true nature.

Effects of human cues are rarely studied in text-based conversational agents (CA).

Incorporating human (relational and visual) cues did not affect the working alliance.

Yet, a good established working alliance contributed to higher adherence levels.

Adding visual cues decreased adherence, possibly due to confusion about the CA's true nature.