Braces meet bots: straight teeth and straight answers?

Abstract

A Commentary on

Santonocito S, Cicciù M, Ronsivalle V.

Evaluation of the impact of AI-based chatbot on orthodontic patient education: a preliminary randomized controlled trial. Clin Oral Investig 2025; 29:278.

Design

Randomized controlled trial (RCT).

Case selection

A sample of 100 patients was enrolled in the study between June 2023 and August 2024. Participants were randomly allocated into two intervention groups of 50 each, using a computer-generated sequence to maintain a 1:1 ratio. The Chatbot group received oral hygiene (OH) and orthodontic treatment guidance via a chatbot accessed through a QR code, while the control group was given standard educational material by means of information leaflets compiled by the leading scientific societies in orthodontics. Each main group was further divided into two subgroups: Leaflet-fixed therapy (LF, n = 27), Leaflet-aligner (LA, n = 23), Chatbot-fixed therapy (CF, n = 24), and Chatbot-aligner (CA, n = 26). The chatbot was built on the open-source Botpress platform and was refined using expert-reviewed training materials. When the participant opens the chatbot application, the interface displays a menu containing 4 questions, with the 3rd question focusing on home oral hygiene procedures.

Data analysis

The evaluation included a clinical examination at baseline (T0) and after 5 weeks of starting orthodontic treatment (T1), along with a questionnaire at T1 assessing participants’ oral hygiene knowledge and compliance. Modified gingival index (MGI) and plaque index (PI) were used for clinical examination of oral hygiene. The questionnaire comprised four domains: Knowledge Evaluation (KE_S), Understanding Scale (US_S), Compliance and Adherence Scale (CAAS_S), and Satisfaction Scale (SS_S). It employed a 5-point Likert scale, with scores ranging from 0 (strongly negative) to 4 (strongly positive), with intermediate values representing negative, neutral, and positive responses.

Results

MGI and PI showed a statistically significant increase from T0 to T1 in both groups. While comparing the change in MGI from the start of treatment between the groups, a statistically significant increase was found in the control group than the chatbot group. Further analysis reported statistically lower increase in MGI in the CF and CA groups as compared to the LF and LA subgroups. There were no differences between the control and chatbot groups with respect to questionnaire section scores.

Conclusions

The study concluded that the use of an AI-based chatbot positively influences patient compliance with oral hygiene maneuvers. Furthermore, it could not be established whether the AI-based chatbot enhanced the patient’s knowledge, understanding, and satisfaction with the information received compared to the information leaflets.