Chest radiographic data acquisition and quality assurance in multicenter studies

Area of Science:

• Medical imaging
• Pediatric radiology
• Public health

Background:

• Multicenter studies require standardized data collection across institutions.
• Standardized reporting forms facilitate reliable data comparison.

Purpose of the Study:

• To present a standardized method for assessing chest radiographic results.
• Developed for a multicenter study on vertically transmitted HIV in infants and children.

Main Methods:

• A standardized form was used to score 3057 chest radiographs from 805 pediatric patients across five centers.
• Data entry utilized a forced-choice, graded response for 12 distinct findings.
• Quality assurance and inter-rater agreement statistics were meticulously documented.

Main Results:

• The reporting form demonstrated moderate to high inter-rater agreement for most radiographic findings.
• Best agreement (kappa = 0.71) was observed for combined bronchovascular markings and/or reticular densities.
• Moderate agreement was found for nodular densities (kappa = 0.56) and parenchymal consolidation (kappa = 0.57), while lung volume agreement was low.

Conclusions:

• The developed tool is effective for assessing pediatric chest radiographs in large populations.
• This standardized method is applicable to both single-center and multicenter studies.
• The tool enhances data reliability and comparability in pediatric imaging research.