Faster Diagnosis: Qualitative Analysis of Implementation Research

This post shows how qualitative analysis of implementation research converts the PLOS Medicine Ethiopia training study (Jan–Sep 2024, published 29 June 2026) into prioritized themes and operational investments that explain a 54.2% reduction in diagnostic interval.

• The multi-tier onsite training plus 6-month mentorship covered 18 Level‑I clinicians, 29 Level‑II providers, and 1, 020 health extension workers (HEWs) across Jan–Sep 2024, and is reported in PLOS Medicine.
• Quantitative gains were large: median knowledge scores rose to approximately 91–100 and the median diagnostic interval fell 54.2% (56.5 → 25.8 days).
• Qualitative themes (leadership, supply shortages, cultural beliefs, referral coordination) explain faster recognition and referral but reveal infrastructure and bed/medication constraints that delayed treatment initiation.
• The training plus mentorship model cost USD 52, 762, suggesting an affordable quality improvement approach when linked to targeted investments.
• Evidano can reproduce this workflow: centralize KAP spreadsheets, transcripts, and chart notes, run thematic extraction, triangulate with timelines, and export reproducible codebooks and visualizations.

Fast take: The PLOS Medicine study (published 29 June 2026) evaluated multilevel onsite training and a 6-month mentorship package across primary, secondary, and tertiary tiers in Northwest Ethiopia (Jan–Sep 2024), producing large knowledge gains and a 54.2% reduction in median diagnostic interval.

• Study window and trainees: Jan–Sep 2024, 18 Level‑I clinicians, 29 Level‑II providers, and 1, 020 HEWs, reported in PLOS Medicine.
• Quantitative results: median knowledge rose to ~91–100 and diagnostic interval fell 54.2% (56.5 → 25.8 days), while patient delay changed less (27.0 → 24.5 days, −9.3%).
• Qualitative results: interviews highlighted leadership, supply shortages, cultural beliefs, and referral coordination as key themes that explain which gains translated into faster diagnosis and which did not translate into timely treatment.

What happened: A University of Gondar team delivered immersive onsite trainings (10 days for tertiary, 7 days for primary, 5 pictorial days for HEWs) followed by 6 months of monthly onsite mentorship and tele-support, and they adapted when unrest and rains blocked travel using cascade mini‑trainings and biweekly telecalls.

• Data sources and analysis: the team used KAP surveys with 100% response from participants, chart reviews of 100 pediatric patient charts, 18 in-depth interviews (IDIs) and HEW focus group discussions (FGDs); analysis combined Wilcoxon and Kruskal–Wallis tests with CFIR-guided thematic coding.
• Qualitative explanation of outcomes: qualitative findings explained why the diagnostic interval fell (improved recognition, standardized referral forms, regular case conferences) and why treatment initiation lagged (diagnostic infrastructure limitations, bed shortages, and medication constraints).

This checklist gives seven reproducible steps to reproduce the Ethiopia implementation analysis using centralized data and AI-assisted qualitative coding.

• 1) Collect: centralize KAP spreadsheets, audio from IDIs and FGDs, and chart review tables.
• 2) Transcribe & translate in Evidano with a custom dictionary for local terms (for example, local words like "nififit").
• 3) Build an initial codebook from CFIR plus inductive themes using Evidano's auto-suggested codes.
• 4) Run thematic frequency and co‑occurrence analyses and compare segments (for example, facilities with versus without biweekly tele-support).
• 5) Triangulate with timelines: map training dates to patient-journey interval changes (Jan–Sep 2024).
• 6) Generate a stakeholder report with clickable quotes, a visual network, and prioritized barriers for investment.
• 7) Export a reproducible package (codebook and visualizations) for policy and scale-up teams.

Ethics note: This analysis is research-focused and non-diagnostic and requires preserving participant consent and de-identifying patient data before upload.

The source study obtained IRB approval and written informed consent; always de-identify and confirm consent before sharing transcripts or chart data with analysis platforms.

Wrap up: Use AI qualitative analysis to convert interviews and field notes from an implementation study into prioritized, reproducible actions, mirroring the approach used by the Gondar team.

• If you are evaluating an implementation study, use AI qualitative analysis to convert interviews and field notes into prioritized, reproducible actions, as the Gondar team did to pinpoint leadership, supply, and referral constraints.
• Try a guided pilot: upload one workshop transcript and a KAP spreadsheet into Evidano and generate a stakeholder brief in under an hour via Evidano.
• For teams preparing scale-up grants, include an AI-assisted qualitative plan to show how you will link training dose to retention and document system adaptations.

Start a free pilot: Try Evidano for free.