Automated analysis of radiology reports is as accurate as manual auditing, but less expensive overall.
An automated analysis using natural language processing and machine learning algorithms can accurately define if reports are compliant with the use of standardized report templates and language, according to a study published in the Journal of the American College of Radiology.
Researchers from Texas sought to evaluate whether a software program using natural language processing and machine learning could accurately audit radiologist compliance with the use of standardized reports compared with manual audits.
The researchers reviewed radiology reports from a one-month period through their software program. Twenty-five reports for each of the 42 faculty members were also manually audited.
The results showed the manual audit provided a mean compliance rate of 91.2 percent compared with 92 percent for the automatic auditing program. The researchers concluded that the automatic audit was as accurate as manual audits, but with reduced labor costs.