Feasibility of AI-Enabled Chatbots for Pre-consultation in HIV Care in Northern Nigeria
Synopsis
This cross-sectional study surveyed 427 adults on antiretroviral treatment (ART) at a large tertiary referral center in Kano, Nigeria, finding that 75.2% were aware of AI chatbots, 72.6% had ever used one, and 66.5% had used chatbots for HIV-related queries (most commonly general HIV/ART information 36.3%, checking ART side effects 23.4%, and preparing questions for clinicians 15.0%), and identified factors independently associated with HIV-related chatbot use, including younger age, post-secondary education, being married, shorter ART duration, presence of comorbidities, smartphone ownership, internet access, and English proficiency.
Interpretation
AI chatbot use for HIV-related information is already common in a routine HIV treatment program in Africa: 72.6% of respondents had ever used a chatbot and 66.5% had used one for HIV-related queries. Evidence on AI chatbot use within routine HIV treatment programs in Africa was previously limited; this study addresses that gap with a systematic sample survey of 427 ART patients. Based on a cross-sectional systematic sample survey of 427 adult ART patients using a validated, culturally adapted interviewer-administered questionnaire; the sample size and sampling approach support estimation of use proportions.
The most common uses of HIV-related chatbots were obtaining general HIV/ART information (36.3%), checking ART side effects (23.4%), and preparing questions for clinicians (15.0%). The study characterizes patterns of chatbot use in HIV care rather than only reporting whether chatbots were used. Use patterns come from self-reported data in the same questionnaire, with proportions based on responses from 427 participants.
Independent predictors of HIV-related chatbot use included younger age (<20 vs ≥50 years: aOR=3.35), post-secondary education (aOR=3.23), being married (aOR=2.26), shorter ART duration (<1 year vs >6 years: aOR=1.91), presence of comorbidities (aOR=2.61), smartphone ownership (aOR=2.13), internet access (aOR=6.67), and English proficiency (aOR=2.11). Guided by the Socio-Ecological Model and the Technology Acceptance Model, the study used multivariable logistic regression to identify independent predictors, providing quantitative evidence on who uses these tools. Multivariable logistic regression provides adjusted odds ratios with 95% confidence intervals for each factor, with internet access showing the strongest association (aOR=6.67).
The findings support clinician-endorsed, multilingual, and low-bandwidth chatbot designs in African HIV treatment programs, alongside safeguards to reduce misinformation and ensure equitable integration. The study translates current use and its predictors into concrete directions for chatbot design and integration. This recommendation is based on descriptive and associative findings from a cross-sectional survey and represents an applied interpretation of the results.
Perspective
The study applies to adult PLHIV on ART at a large tertiary referral center in Kano, Nigeria; its prevalence, use patterns, and predictors can inform chatbot design and integration in similar HIV treatment programs in Africa, and the findings support further work on clinician-endorsed, multilingual, low-bandwidth designs and safeguards to reduce misinformation.
This is a single-center cross-sectional survey using self-reported interviewer-administered questionnaires, and the identified factors are associations rather than causal; the specific chatbot platforms used, content accuracy, and effects on clinical outcomes are not described in the text, so readers may watch for future studies validating these findings across more centers and longer follow-up.