From Interviews to Action: AI Qualitative Analysis

The November 2026 PLoS ONE study found that introducing balanced energy and protein supplementation through routine antenatal care requires clearer definitions, stronger provider training, reliable supplies, culturally relevant demand-generation strategies, and better monitoring. The PLoS ONE study reached these conclusions through a mixed-methods analysis of stakeholder interviews and national guidelines.

• Between August and December 2023, the study team conducted 13 semi-structured interviews with global and country stakeholders and reviewed antenatal-care guidelines from six countries.
• In November 2026, the study reported that Ethiopia, Malawi, and Mozambique had revised national antenatal-care guidance in ways that created an enabling environment for balanced energy and protein supplementation.
• The study reported that stakeholders faced confusion about the definition of balanced energy and protein supplementation, weak provider capacity, stockouts, household sharing, selling of supplements, and limited routine monitoring.
• A 2021 systematic review cited by the study found that balanced energy and protein supplementation reduced small-for-gestational-age births by 29%, low birthweight by 40%, and stillbirth by 61% across nine low- and middle-income countries.
• The study demonstrates how AI qualitative analysis can help researchers compare themes across countries, stakeholder groups, and implementation domains while preserving a transparent path back to interview evidence.

The November 2026 PLoS ONE study used a mixed-methods landscape analysis to examine whether routine antenatal care could deliver balanced energy and protein supplementation effectively. The researchers reviewed national antenatal-care guidelines from Colombia, Ethiopia, Ghana, Malawi, Mozambique, and Nepal, and interviewed 13 stakeholders from governments, international nongovernmental organizations, multilateral organizations, and donor groups.

The researchers recorded, transcribed, and analyzed interviews using Dedoose qualitative software version 10.0 between August and December 2023. Three research team members developed a deductive codebook from a conceptual framework, refined the codebook with emergent themes, analyzed transcripts independently, and resolved early coding discrepancies through consensus.

The PLoS ONE study identified four major implementation domains: the enabling policy environment, service readiness, product supplies and logistics, and monitoring. The analysis also surfaced cross-cutting issues involving acceptability, counselling, cultural norms, household sharing, resale, local production, and sustainability.

One Malawi NGO stakeholder told the researchers, “It is important to ensure that there are capacity building activities as well as proper training to provide or support quality counseling [for BEP].” The quotation illustrates why qualitative analysis adds operational context that a numeric indicator alone cannot provide.

A multilateral organization stakeholder described the definitional problem directly: “If you look at the [WHO] guideline, it’s not clearly defined in terms of what the exact composition of BEP is.” AI-assisted analysis can identify repeated concerns like definitional ambiguity, but researchers must retain the original quotation and its speaker context.

Health policy teams should treat implementation barriers as interconnected themes rather than isolated complaints. The November 2026 PLoS ONE study linked unclear product definitions with provider awareness, training needs, supply-chain classification, counselling, and monitoring requirements.

Researchers should separate frequency from importance when interpreting interview data. The study reported that stakeholders repeatedly raised stockouts, sharing, selling, and weak monitoring, but the purposive sample was designed to provide contextual insight rather than statistically representative prevalence estimates.

Cross-segment analysis can make policy evidence more actionable. Comparing government, NGO, donor, and multilateral stakeholder interviews can show where perspectives converge, such as the need for training, and where implementation responsibilities remain unclear, such as whether balanced energy and protein products should be classified as food or medicine.

Researchers should preserve quotations alongside coded themes. A government stakeholder in Ghana said, “If we were to introduce BEP supplementation via ANC, we would need to integrate BEP indicators into our existing registers and reporting system.” The quotation connects the theme of weak monitoring to a specific operational requirement.

AI qualitative analysis is most useful when it accelerates organization and comparison without replacing methodological decisions. Researchers should define the unit of analysis, review suggested codes, check negative cases, document translation decisions, and validate synthesized findings against the source transcripts.

The November 2026 PLoS ONE study shows that qualitative evidence can clarify why a promising maternal nutrition intervention may face barriers in policy, training, supply chains, community acceptance, and monitoring. AI qualitative analysis can help health researchers move faster from transcripts and documents to comparable themes, while human researchers remain accountable for interpretation and evidence quality.

A practical workflow is to transcribe and organize source material, apply transparent thematic and cross-segment analyses, review quotations and outliers, and connect every recommendation to supporting evidence. Teams that want to test this workflow can try Evidano for free.