No abstract

No abstract

Radiology departments are beginning to embrace new technologies to decrease operating budgets and improve services. One of these technologies is the picture archiving and communication system (PACS). PACS, through immediate availability of images to the radiologist, promises to decrease turnaround times of reports to the clinician. The purpose of this study was to determine if this technology actually decreases the time for referring clinicians to receive reports generated by the radiologist. The time to provide a preliminary report by a resident and time to finalize this report by a board-certified radiologist was retrospectively obtained for 6,022 abdominal and pelvic computed tomography (CT) scans over two 1-year periods from March 1, 1997 to March 1, 1998 and from March 1, 1998 to March 1, 1999. During the first year, interpretation was conducted using hard-copy film and during the second using PACS. In both 1-year periods, MedSpeak voice recognition software (IBM, White Plains, NY) was employed for dictation. The average time for a preliminary report for a abdominal and pelvic CT, dictated by a resident or fellow, to be available in alphanumeric form on the hospital information system using hard-copy film was 3.73 days. The installation of a PACS system decreased this turnaround time to 0.56 days, representing an 85.0% improvement. The time to availability of final reports, ie, signed by board-certified staff radiologists, was 5.49 days in the hard-copy interpretation subset and 5.97 days in the PACS subset. The addition of PACS into an academic gastrointestinal radiology division improves availability of alphanumeric preliminary reports of abdominal and pelvic CTs on the hospital information system (HIS), dictated by a resident or fellow, by 85.0%. There was no impact with a PACS on the time to final sign reports by a staff board certified radiologist as signing patterns remained relatively constant over the two interpretation formats.

Voice recognition represents a technology that is finally ready for prime time use. As radiology services continue to acquire a larger percentage of the shrinking health-care dollar, decreasing operating costs and improved services wili become a necessity. The benefits of voice recognition implementation are significant, as are the challenges. This report will discuss the technology, experiences of a major health-care institution with implementation, and potential benefits for the radiology practice. In late 1994, the effort to develop large-vocabulary, continuous speech recognition systems in American English had progressed to the point where accuracy had become a less of an issue and vendors could focus on specific markets. In a recent survey, the Healthcare Information Management and Systems Society found in a worldwide membership survey that 31% of respondents were eager to install VR systems in their practicesl; no specialty was more suited than radiology. Until now, radiologists have either handled the process of dictation manually or on tape to medical transcriptionists; the process has been time-consuming and expensive. The choice of transcribing radiology reports with VR systems is an obvious focus due to the consistent and predictable vocabulary within dictation practices. Multiple solutions have been generated and we will discuss our experiences at the Massachusetts General Hospital (MGH) with IBM Medspeak/Radiology (White Plains, NY). We review the VR technology, challenges encountered through deployment, implementation strategies, and experienced benefits. Speech RecognitionSpeech recognition, also referred to as speech-totext, is the technology that makes it possible for the computer to translate spoken word into type. The first type of speech recognition is spoken command recognition, also known as command and control, and the second is pure dictation. Command and control dictation handles the recognition of single words or short phrases spoken with continuous speech, such as "Begin Dictation" or "Accept and Sign." Dictation comes in two divisions: discrete (or isolated) and continuous. The discrete dictation technology has lower processing power requirements, but requires the end-user to place a short pause between each spoken word. Condnuous dictation, as the name implies, does not carry this limitation and is used in the radiology implementation. Speech SynthesisSpeech synthesis, also referred to as text-tospeech, is the technology that makes it possible for the computer to produce the phonemes we would make when we read text aloud. Within radiology, speech synthesis allows the radiologist to review the report, as well as in playback during dictation of reports. Speaker Identification and VerificationSpeaker identification and verification are two related technologies. These technologies, unlike speech recognition and speech synthesis, deal with the identity of the human speaker and not with what was spoken or with synthesizing the particular human voice. With speaker verification, technology ...