The utility and cost effectiveness of voice recognition technology in surgical pathology

Abstract

Voice recognition (VR) technology in computer systems converts speech directly into electronic text. In pathology, VR holds promise to improve efficiency and to reduce transcription delays and costs. We investigated the utility and cost effectiveness of targeted VR deployment in surgical pathology. A VR system was deployed for entry of gross descriptions of biopsies and of low to moderate complexity specimens and for entry of final reports for specimens not requiring microscopic analysis. Templates for VR were developed for all reports. Free-text speech entry was used to enter information not covered by templates. Voice converted to text by VR crossed over an interface into the anatomic pathology laboratory information system. Tallies were kept of whether individual specimens were entered by VR or by conventional dictation. A computer program was written to analyze the number of lines of text entered through VR. Cost savings were calculated based on per-line transcription costs from an outside agency. Over 18 months, gross descriptions for an average of 5617 specimens per month were entered via VR, corresponding to 70% of all gross specimens processed by the laboratory. A mean of 106 gross-only final reports per month was entered through VR. VR facilitated same-day processing of specimens received after the previous day processing cutoff time (average 35 specimens per day). VR generated an average of 23,864 lines of text per month, translating to $2625 savings per month. Estimated payback period for VRT as implemented is 1.9 years. The use of VR for gross descriptions of biopsies and low to moderate complexity specimens and for gross-only final reports in surgical pathology facilitates data entry, reduces transcription costs, and contributes to improved turnaround time. Development of templates is important to successful implementation of VR in surgical pathology.