Abstract
Background
Chest radiography is the standard for diagnosing pediatric lower respiratory infections in low-income and middle-income countries. A method for interpreting pediatric chest radiographs for research endpoints was recently updated by the World Health Organization (WHO) Chest Radiography in Epidemiological Studies project. Research in India required training local physicians to interpret chest radiographs following the WHO method.
Objective
To describe the methodology for training Indian physicians and evaluate the training’s effectiveness.
Materials and methods
Twenty-nine physicians (15 radiologists and 14 pediatricians) from India were trained by two WHO Chest Radiography in Epidemiological Studies members over 3 days in May 2019. Training materials were adapted from WHO Chest Radiography in Epidemiological Studies resources. Participants followed WHO methodology to interpret 60 unique chest radiographs before and after the training. Participants needed to correctly classify ≥80% of radiographs for primary endpoint pneumonia on the post-training test to be certified to interpret research images. We analyzed participant performance on both examinations.
Results
Twenty-six of 29 participants (89.7%) completed both examinations. The average score increased by 9.6% (95% confidence interval [CI] 5.0–14.1%) between examinations (P<0.001). Participants correctly classifying ≥80% of images for primary endpoint pneumonia increased from 69.2% (18/26) on the pretraining to 92.3% (24/26) on the post-training examination (P=0.003). The mean scores of radiologists and pediatricians on the post-training examination were not statistically different (P=0.43).
Conclusion
Our results demonstrate this training approach using revised WHO definitions and tools was successful, and that non-radiologists can learn to apply these methods as effectively as radiologists. Such capacity strengthening is important for enabling research to support national policy decision-making in these settings. We recommend future research incorporating WHO chest radiograph methodology to consider modelling trainings after this approach.
Prediction of Pulmonary to Systemic Flow Ratio in Patients With Congenital Heart Disease Using Deep Learning–Based Analysis of Chest Radiographs
