Exploring tuberculosis patients' preferences for AI-assisted remote health management services in China: a protocol for a discrete choice experiment.

INTRODUCTION

Effective health management is critical for patients with tuberculosis (TB), especially given the need for long-term treatment adherence and continuous monitoring. Artificial intelligence (AI)-assisted remote health management services offer a promising solution to increase patient engagement, optimise follow-up and improve treatment outcomes. However, little research has explored TB patients' preferences for these services, and no discrete choice experiment (DCE) has systematically investigated how they make trade-offs between different service attributes. This study aims to (1) identify key attributes of AI-assisted remote health management services that influence TB patients' choices, (2) assess how patients with TB evaluate trade-offs between different service options using a DCE and (3) examine whether preferences vary by sociodemographic characteristics and health system factors.

METHODS AND ANALYSIS

Six attributes were identified through a literature review, focus group discussions and expert consultations. A fractional factorial design was used to generate choice sets while maintaining statistical efficiency and minimising respondent burden. The DCE will be analysed using a multinomial logit model to estimate average preferences. A mixed logit model will be applied to explore preference heterogeneity among participants, incorporating interaction terms with sociodemographic and attitudinal variables. Stratified and latent class analyses will also be considered to further investigate sources of heterogeneity.

ETHICS AND DISSEMINATION

This study complies with the Declaration of Helsinki and has been approved by the Ethics Committee of Wuhan Pulmonary Hospital. All participant data will remain anonymous, and individuals may withdraw from the study at any time. The findings will inform the development of patient-centred AI-assisted TB management strategies and contribute to broader policy discussions on AI integration in TB care. The results will be disseminated through peer-reviewed journal publications, policy briefs, conferences and online platforms.