AI workflows for qualitative analysis of IYCF practices

Evidano is an AI-powered qualitative data analysis platform that maps directly onto the protocol tasks: transcription, coding, visualization, and secure storage for sensitive health research.

Using the PLOS One protocol assets (460 mother-infant dyads, 44 AWCs, baseline/intervention/endline), teams can prototype reproducible codebooks and stakeholder visuals in weeks rather than months.

• The PLOS One protocol documents a 24-month, quasi-experimental, mixed-methods intervention in Palghar across 44 AWCs with n=460 mother-infant (6–12 months) dyads (published 15 Jul 2026).
• Collected assets include verbatim IDIs, diet cards, weekly process registers, audiovisual IEC, and WhatsApp logs, enabling multimodal ingestion and linked coding.
• Evidano supports auto-transcription with custom glossaries, AI-assisted coding seeded by IYCF domains, cross-segment frequency analyses, and end-to-end encryption.
• Follow the eight-step checklist in this post to reproduce the protocol’s qualitative analysis and export quote sets, co-occurrence networks, and policy-ready briefs.

The PLOS One protocol offers a 3-phase (baseline, 12-month intervention, post-intervention) mixed-methods project aimed at improving complementary feeding via SBCC, pictorial IEC, hot cooked complementary feeds, and supplementation.

This PLOS One protocol documents the study design and provides pre-specified outcomes and timelines for qualitative and quantitative data collection and analysis.

• Why qualitative analysis matters: in-depth interviews with AWWs, ASHAs and supervisors (IDIs) will surface contextual barriers that determine whether IEC materials and cooking demos are adopted.
• Payoff for your team: transform qualitative field data into prioritized, segment-specific recommendations for program design and policy.
• Full protocol: PLOS One protocol.

The protocol collects verbatim in-depth interviews with frontline workers and supervisors, structured questionnaires and diet cards from 460 mothers at baseline and endline, weekly process registers, audiovisual IEC and cooking-demo videos, and WhatsApp group exchanges between AWWs and project staff.

• Qualitative scope: IDIs with CDPO, supervisors, AWWs, ASHAs, plus open feedback during cooking demos.
• Quantitative complements: eight WHO IYCF core indicators, anthropometrics, process indicators (weekly).
• Mixed data types: audio, video, scanned forms, spreadsheets and short messages, which requires multimodal ingestion and linked coding.

The checklist reproduces the protocol’s qualitative analysis in eight actionable steps that map to ingestion, coding, quantification, visualization, and iteration.

Step 1: Collect raw assets (audio IDIs, diet cards, WhatsApp logs) and scan/upload.

Step 2: Auto-transcribe and translate with a custom glossary for local food names and tribal terms.

Step 3: Import IYCF domains from the protocol as a seed codebook.

Step 4: Run AI-assisted initial coding; review and refine code hierarchy with human coders.

Step 5: Generate frequency tables and cross-segment comparisons (AWC, tribe, pre/post).

Step 6: Visualize co-occurrence networks and export verbatim quote sets for each theme.

Step 7: Use AI chat to draft policy briefs or recommendation bullets mapped to outcome indicators.

Step 8: Plan targeted follow-ups using AI avatar interviews for unresolved themes.

Ethics and safeguards require IRB approval, informed consent, anonymization, and controlled public release for tribal and pediatric research, and the PLoS protocol obtained IRB approval (ICMR-NIRRCH, 13 Nov 2023).

This is research-focused guidance; it does not replace institutional ethics review.

For tribal and pediatric research, maintain informed consent, anonymize identifiers, and store deidentified datasets for public release only as approved.

The PLoS protocol for Palghar offers a rich mixed-methods design that benefits from AI-accelerated qualitative workflows, including faster transcription, reproducible codebooks, cross-segment evidence, and stakeholder-ready visuals.

• If you’re preparing for implementation research or policy briefs from IYCF fieldwork, start by uploading a representative batch (3–5 IDIs + one AWC register) to prototype the codebook and visuals in days.
• Ready to try this on your data? Explore a focused pilot and accelerate synthesis with Evidano by Try Evidano for free, or contact us at Evidano to map your study assets to the 8-step checklist above.