A Train-the-Trainer Point of Care Ultrasound (POCUS) Program for Pediatric Pneumonia in a Low-Resource Setting

TL;DR

A train-the-trainer program for lung ultrasound in diagnosing pediatric pneumonia was successfully implemented in Pakistan, showing strong agreement with expert interpretations.

Contribution

Demonstrates the feasibility and effectiveness of a train-the-trainer POCUS program in low-resource settings with local champions.

Findings

Strong inter-rater reliability (κ = 0.83) between trained CHWs and remote experts.

Moderate agreement (κ = 0.66) when novices and local champions interpreted their own POCUS exams.

Hands-on training and local champion support are crucial for successful implementation.

Abstract

Lung point of care ultrasound (POCUS) has the potential to transform pediatric pneumonia care in low resource settings. Prior studies of novice POCUS users in such settings showed high agreement with remote POCUS experts for diagnosing pediatric pneumonia, but use of remote experts may falsely inflate this agreement.

This study aimed to 1. Deliver a train-the-trainer program in Pakistan on lung POCUS for diagnosing pediatric pneumonia; 2. Determine inter-rater reliability between i) study-trained community health workers (CHWs) and a remote expert, with both interpreting POCUS examinations acquired by the CHWs, and ii) study-trained CHWs and local champions, with both interpreting examinations that they had acquired.

Phase 1: Canadian pediatric POCUS experts developed and delivered a lung POCUS training program for two user groups in Pakistan. These groups included local champions (who had POCUS experience) and CHWs (who were POCUS novices). Phase 2: Children with suspected pneumonia underwent two lung POCUS examinations, one by a CHW and one by a local champion. Examinations were recorded and later reviewed by a remote expert for interpretation and quality assurance. Inter-rater reliability was determined.

Two local champions and three CHWs were successfully trained. An analysis of 231 recruited patients showed strong inter-rater reliability between study-trained CHWs and remote expert interpretations (κ = 0.83). In contrast, inter-rater reliability was moderate (κ = 0.66) between interpretations by novices and local champions when these users interpreted the examinations that they themselves had acquired.

Our study showed that train-the-trainer programs are feasible and can be effective, while highlighting the importance of hands-on training and having local champions provide longitudinal support to novices.