JMIR Publications

ABSTRACT

Background:

Infant feeding practices including breastfeeding can benefit a range of maternal and child health outcomes. Therefore, parent access to evidence-based infant feeding advice is of critical importance. In recent years, there has been increased use of digital health technologies to support infant feeding practices. Despite its potential, the use of Artificial Intelligence (AI) to complement existing health care and connect families to timely infant feeding support remains relatively unexplored.

Objective:

To explore women’s perceptions of the use of AI-enabled infant feeding support, as part of mobile health (mHealth) interventions. The study investigates: (A) Openness to receiving AI-enabled infant feeding support; (B) Experiences with existing forms of AI-enabled infant feeding support; (C) Preferences for text messages generated by AI versus with Child and Family Health (CFH) nurses; and (D) Opinions on infant feeding topics that could be suitable for AI-enabled support.

Methods:

Two data collection activities were undertaken with women (primary carers) of infants aged 6 – 14 months old. Participant recruitment occurred from a sample of women within the Hunter New England Local Health District (HNELHD) of New South Wales (NSW), Australia, who had received antenatal care. Quantitative surveys were undertaken to explore women’s openness to receiving AI-enabled support for breastfeeding (Objective A). Descriptive analyses were conducted along with logistic regression analyses to explore associations between participant characteristics and openness to AI. Qualitative data collection involved focus groups to explore women’s perceptions and preferences for AI-enabled support including opinions on infant feeding topics suitable for AI (Objectives B,C,D). Thematic Codebook Analysis was used to analyse the focus group transcripts.

Results:

A total of 164 women completed the quantitative survey. Women reported being moderately (53%) open to AI providing recommendations to see a health professional for infant feeding support. Fewer women were open to receiving breastfeeding advice generated by AI (35-45%). There were 12 women who participated in 3 online focus groups. Thematic analysis resulted in 3 overarching themes: 1) Opportunities to fill gaps in infant feeding support with AI, 2) Variable confidence in engaging with AI for information and advice and 3) Potential convenience of AI and mHealth to offer timely infant feeding support.

Conclusions:

The study highlights the potential of AI and key barriers to women’s acceptability and engagement. While women recognised the potential for AI to fill health care gaps in infant feeding support including after business hours, there was less interest in AI replacing ‘in-person’ support or offering written information that can be located via a simple online search. Women’s concerns regarding the credibility and trustworthiness of AI-enabled support should be addressed to maximise their use of emerging AI-enabled tools, embedded within digital technologies and mHealth. There is potential for AI to complement usual care, rather than act as a replacement.