Qualitative Analysis: EIR Perceptions in Bangladesh

Evidano is an AI-powered qualitative data analysis platform that ingests mixed surveys and open text to produce reproducible thematic, frequency, and cross-segment evidence. This post shows how to convert the July–September 2024 Rajshahi City EIR pilot (n=321) into actionable, secure qualitative deliverables that link perceptions to registry metrics. Major quantitative signals to preserve are facility-based registration at 87% and caregiver satisfaction clustered between 59–78% satisfied and 16–22% highly satisfied.

• Evidano can automate initial coding and PII redaction so teams reproduce thematic and frequency analyses from the Rajshahi pilot in days, not weeks.
• Report low-n provider segments (n=16 HCPs) with caution, supplement with bootstrapped summaries or targeted follow-ups, and flag uncertainty clearly.
• Pair perception themes with objective registry audits (timeliness, registration mode) before making scale-up recommendations; the pilot reported 88% registered within 45 days.

This section summarizes the Rajshahi pilot design and data collection methods. The Rajshahi pilot used a mixed survey with 5-point Likert items and open-ended questions to capture perceptions of an EIR introduced in 2019 but assessed in 2024 operations. Data were collected from caregivers attending EPI sessions and HCPs using tablet-based entry and automated SMS reminders.

• Design: Cross-sectional, descriptive (July–Sep 2024); not designed for causal inference.
• Analysis: SPSS v24 for quantitative summaries; open-ended responses were reviewed and summarized descriptively.
• Limitations: purposive sampling of EPI attendees, small HCP sample (n=16), potential social desirability bias, and open-ends withheld publicly for confidentiality.

This checklist explains how to reproduce the Rajshahi pilot qualitative analysis in seven steps:

• 1) Gather inputs: upload the survey spreadsheet (Likert items + open text) and any de-identified interview transcripts to Evidano.
• 2) Apply PII redaction and (if needed) translation with a custom dictionary for local terms.
• 3) Auto-code: run AI-assisted initial coding to surface candidate themes; review and finalize a codebook.
• 4) Thematic + frequency analysis: generate theme prevalence by segment (caregiver vs HCP; ward; registration mode).
• 5) Cross-segment analysis: produce difference-in-proportion tables and co-occurrence networks to show which themes cluster with complaints (connectivity issues, device limits).
• 6) Visualize & export: create word clouds and hierarchical code maps for presentations and dashboards.
• 7) Produce reproducible deliverables: export results, notes, and an interactive Q&A of the corpus for stakeholders.

This section recommends next actions for follow-up evaluations and scaling decisions. If you are preparing a follow-up to the Rajshahi pilot (scaling, rural rollout, or longitudinal evaluation), start by converting the pilot’s semi-structured open-ends into a reproducible thematic dataset and linking those themes to registry metrics (timeliness, registration completeness).

Try these next moves: upload your survey and open-text files to Evidano; run an automated codebook draft; use PII redaction and cross-segment analyses to produce a stakeholder-ready brief in days, not weeks. For a hands-on trial, Try Evidano for free.

Ethics note: this guidance is research-focused and non-diagnostic. The original study had IRB approval (IBMC, R/IRB/2024/04/01) and anonymized data handling, replicate that practice when you work with health data.