Supporting data for "Development, evaluation and implementation of an HPV vaccine chatbot for parents to improve HPV vaccination uptake among middle school girls in China"

<b>Scope of data:</b> This dataset supports research on the development, evaluation, and implementation of an HPV vaccine chatbot intervention for parents of middle school girls in China. The research employed a multi-phase approach: first developing and evaluating an HPV vaccine chatbot based on large language model (LLM) integrated with a China-specific HPV vaccine knowledge base. The second phase implemented the chatbot and assessed its effectiveness in improving HPV vaccination uptake among middle school girls and parental HPV-related literacy through a cluster randomized trial (cRCT) involving 2,671 parents from 180 middle school classes across four sites: Shanghai megacity and three regions in Anhui province (one urban and two rural areas), conducted from January to May 2024.<b>Data files description:</b> The dataset comprises seven key files organized by research phases. Development phase (Chapter 3) data includes: (1) an HPV vaccine knowledge benchmark dataset with 212 validated questions used to evaluate LLM performance; (2) an HPV vaccine question dataset containing 107 parent-relevant questions with responses from different chatbot personas; and (3) a Chinese HPV vaccine knowledge base with 40 core knowledge points specific to the Chinese context. Implementation phase (Chapter 4) data includes: (4) a comprehensive dataset of 2,671 trial participants with demographic information and outcome measures; (5) baseline data for all 3,304 initially consented participants; (6) detailed chatbot engagement metrics for 1,051 intervention group parents; and (7) a collection of 16,950 user questions submitted to the chatbot during implementation.

<b>数据范围：</b>本数据集用于支持针对中国初中女生家长的人乳头瘤病毒（HPV, Human Papillomavirus）疫苗聊天机器人干预方案的开发、评估与落地研究。本研究采用多阶段研究方法：第一阶段开发并评估一款融合中国本土化HPV疫苗知识库的大语言模型（LLM, Large Language Model）驱动的HPV疫苗聊天机器人；第二阶段部署该聊天机器人，并通过整群随机对照试验（cRCT, cluster randomized trial）评估其在提升初中女生HPV疫苗接种率及家长HPV相关健康素养方面的有效性。该试验于2024年1月至5月开展，覆盖上海超大城市与安徽省的三个区域（1个城区、2个农村地区）的180个中学班级，共纳入来自上述班级的2671名家长。<b>数据文件说明：</b>本数据集包含7个核心文件，按研究阶段进行组织。开发阶段（第3章）数据包括：(1) 包含212道经验证题目的HPV疫苗知识基准数据集，用于评估大语言模型性能；(2) 包含107道与家长相关的HPV疫苗问答数据集，涵盖不同聊天机器人角色生成的回复内容；(3) 包含40个适配中国场景核心知识点的中文HPV疫苗知识库。实施阶段（第4章）数据包括：(4) 涵盖2671名试验参与者的综合数据集，包含人口统计学信息与结局测量指标；(5) 全部3304名初始签署知情同意的参与者的基线数据；(6) 针对1051名干预组家长的详细聊天机器人交互使用指标；(7) 试验期间提交至聊天机器人的16950条用户问题合集。