Conversations With AI Chatbots Increase Short-Term Vaccine Intentions But Do Not Outperform Standard Public Health Messaging

NCT ID: NCT07132125

Last Updated: 2025-08-20

Basic Information

• Recruitment Status: COMPLETED
• Clinical Phase: NA
• Total Enrollment: 930 participants
• Study Classification: INTERVENTIONAL
• Study Start Date: 2025-03-17
• Study Completion Date: 2025-05-22

Brief Summary

The goal of this clinical trial is to test whether short conversations with large language model (LLM) chatbots can persuade vaccine-hesitant parents to vaccinate their children against human papillomavirus (HPV). The study compares two chatbot styles to official public health information and to a no-message control. Parents of HPV-eligible children first complete a survey about their attitudes toward the HPV vaccine and their main concerns. They are then randomly assigned to read public health materials, have a three-minute conversation with either a default-style chatbot or a conversational-style chatbot tailored to their concern, or receive no message. The main outcome is change in intent to vaccinate immediately after the intervention, with follow-up surveys at 15 and 45 days.

Conditions

HPV Vaccination Intent

Study Design

• Allocation Method: RANDOMIZED
• Intervention Model: PARALLEL
• Primary Study Purpose: PREVENTION
• Blinding Strategy: NONE

Study Groups

Default-style LLM chatbot

A large language model chatbot (GPT-4o) that delivers tailored persuasive messages to parents about HPV vaccination, using the model's default style with longer, structured responses. Participants engage in a three-minute, multi-turn conversation, with responses personalized to their top-rated vaccine concern identified in a pre-intervention survey.

Group Type: EXPERIMENTAL

Default-style LLM chatbot

Intervention Type: BEHAVIORAL

A large language model chatbot (GPT-4o) that delivers tailored persuasive messages to parents about HPV vaccination, using the model's default style with longer, structured responses. Participants engage in a three-minute, multi-turn conversation, with responses personalized to their top-rated vaccine concern identified in a pre-intervention survey.

Conversational-style LLM chatbot

A large language model chatbot (GPT-4o) that delivers tailored persuasive messages to parents about HPV vaccination in a conversational style, with short responses. Participants engage in a three-minute, multi-turn conversation, with responses personalized to their top-rated vaccine concern identified in a pre-intervention survey.

Group Type: EXPERIMENTAL

Conversational-style LLM chatbot

Intervention Type: BEHAVIORAL

A large language model chatbot (GPT-4o) that delivers tailored persuasive messages to parents about HPV vaccination in a conversational style, with short responses. Participants engage in a three-minute, multi-turn conversation, with responses personalized to their top-rated vaccine concern identified in a pre-intervention survey.

Public health informational materials

Official public health informational materials on HPV vaccination from the Centers for Disease Control and Prevention (U.S.), National Health Service (U.K.), or Public Health Agency of Canada, matched to the participant's country of residence. Materials are 589-680 words in length, cover HPV risks and benefits of vaccination, and are presented for three minutes before participants proceed.

Group Type: ACTIVE_COMPARATOR

Public health informational materials

Intervention Type: BEHAVIORAL

Official public health informational materials on HPV vaccination from the Centers for Disease Control and Prevention (U.S.), National Health Service (U.K.), or Public Health Agency of Canada, matched to the participant's country of residence. Materials are 589-680 words in length, cover HPV risks and benefits of vaccination, and are presented for three minutes before participants proceed.

No Message

No Intervention

Group Type: NO_INTERVENTION

No interventions assigned to this group

Interventions

Default-style LLM chatbot

A large language model chatbot (GPT-4o) that delivers tailored persuasive messages to parents about HPV vaccination, using the model's default style with longer, structured responses. Participants engage in a three-minute, multi-turn conversation, with responses personalized to their top-rated vaccine concern identified in a pre-intervention survey.

Intervention Type: BEHAVIORAL

Conversational-style LLM chatbot

A large language model chatbot (GPT-4o) that delivers tailored persuasive messages to parents about HPV vaccination in a conversational style, with short responses. Participants engage in a three-minute, multi-turn conversation, with responses personalized to their top-rated vaccine concern identified in a pre-intervention survey.

Intervention Type: BEHAVIORAL

Public health informational materials

Official public health informational materials on HPV vaccination from the Centers for Disease Control and Prevention (U.S.), National Health Service (U.K.), or Public Health Agency of Canada, matched to the participant's country of residence. Materials are 589-680 words in length, cover HPV risks and benefits of vaccination, and are presented for three minutes before participants proceed.

Intervention Type: BEHAVIORAL

Eligibility Criteria

Inclusion Criteria

• Adults aged 18 years and up
• Living in Canada, the UK, or the USA
• At least one child eligible for HPV vaccine based on country-specific guidelines who has not received any doses of the vaccine or whose vaccination status is unknown

Exclusion Criteria

• Failure to pass attention checks
• Identified as non-human respondents by reCAPTCHA or other automated bot detection

• Minimum Eligible Age: 18 Years
• Eligible Sex: ALL
• Accepts Healthy Volunteers: Yes

Sponsors

University of Pennsylvania

OTHER

Sponsor Role: lead

Responsible Party

Responsibility Role: SPONSOR

Locations

Participants are recruited online via Prolific

Philadelphia, Pennsylvania, United States

Countries

United States

Other Identifiers

LLM_857299

Identifier Type: -

Identifier Source: org_study_id