By Cynthia E. Keen, staff writer
    November 16, 2012

    Sunday, November 25 | 1:00 p.m.-1:30 p.m. | LL-INS-SU5B | Lakeside Learning Center
    Radiologists who believe that using speech recognition (SR) makes them less efficient and more error-prone in reports should find solace in this poster presentation.

    A meta-analysis of the effects of speech recognition on these factors was undertaken by Dr. Luigi Lepanto, from Centre Hospitalier de l'Université de Montréal, and colleagues. They evaluated three prospective and three retrospective studies with a large amount of heterogeneity out of a group of 24 studies.

    The studies showed that radiologists' time to dictate a report increased by 15% to 190% compared to conventional dictation and preparation by a transcriptionist. Error rates increased with the complexity of the report and were significant.

    Lepanto will discuss additional details of the analysis and discuss the researchers' findings that a gradual, selective implementation of speech recognition will generate the maximum benefit of the technology.