"Back-end Voice Recognition Software"

The National Institutes of Health, Clinical Center is conducting a market survey/sources sought to help determine the availability and technical capability of qualified small businesses, veteran-owned small businesses, service disabled veteran-owned small businesses and/or HUBZONE small businesses capable of serving the needs mentioned below. This market survey/sources sought announcement is not a request for proposals and the Government is not committed to award a contract or purchase order pursuant to this announcement. The information from this market research is for planning purposes only and will assist the Government in planning its acquisition strategy. This is strictly market research and the Government will not entertain questions concerning this market research. The Government will not pay for any costs incurred in the preparation of information for responding to this market survey. Any proprietary information should be clearly identified as "proprietary information". The Clinical Center's Medical Record Department (MRD) is responsible for providing transcription services for a variety of specialized medical reports related to patients' inpatient hospital admissions, day hospital visits, outpatient clinic visits, ambulatory surgery and other complex diagnostic medical studies. Such reports are categorized into four groups including medical reports, diagnostic reports, pathologic reports and letters to outside physicians. The NIH, CC, MRD is seeking a back end speech recognition solution to support its existing dictation and transcription infrastructure. The software must seamlessly integrate within the MRD's existing health information management application infrastructure which is currently supported by 3M's health information transcription platform products including Provider ID TM , ChartScript.net TM , and ESA TM . The desired back end speech recognition solution must be capable of collecting voice files from the NIH dictation system, routing them through a deferred speech recognition processor, and returning a ready-to-edit document in ChartScript TM for a medical editor to review prior to presenting the final document to the dictating clinician for review. The back end speech recognition software must be capable of operating and supported 24 hours a day, 7 days a week, 52 weeks a year. The speech recognition software must be capable of analyzing and utilizing edits made to documents by a medical editor to continually enhance and improve the accuracy of the speech recognition processing software in subsequent dictations completed by each individual clinician. The system must process dictated voice files and present the converted document immediately via a real-time interface processing service. Additionally, the system must be capable of handling provisions for the MRD to apply settings at an individual dictating clinician level to enable reports to be processed via a direct transcription model, therefore, bypassing the speech recognition software if ncessary. The application must be capable of collecting dictated voice files from individual clinicians in order to build the clinicians' voice files. In addition, the system must utilize a continuous learning method and be able to adapt to specific dictating clinicians' dictation habits while also utilizing acoustic learning methods in an effort to maximize the usability rate of recognized text presented in the MRD's ChartScript.net application. The system must be capable of handling approximately 4 million lines of text per year. Additionally, the NIH, CC maintains strict controls over its information technology architecture and as such requires that all connections and data communications occur via a secure connection. Payment for services will be accomplished via a unit cost per line of dictation that passes through the speech recognition service. In responding to this sources sought market survey and to demonstrate capability of providing the back end speech recognition software as described above, vendors must minimally provide information related to the following: 1) specific type of software system utilized/provided/supported; 2) specific transcription platforms and/or speech recognition software supported and utilized by the company; 3) description of related experience in the health information management field particularly implementing and supporting a back end speech recognition system; and 4) overview of information technology security architecture incorporated in the company's current structure. The vendor response shall also include any other specific and relevant information related to the requirements of this project that will enable the Government to determine the capabilities of the company to perform the specialized requirements described in this synopsis. Interested organizations must demonstrate and document in any response submitted, to this market survey extensive experience with and the ability to perform all of the specialized requirements elsewhere described. This notice is a market survey and is for information and planning purposes only and does not commit the Government to any contractual agreement. THIS IS NOT A REQUEST FOR PROPOSALS. The Government will not award a Purchase Order or Contract based upon Vendor responses to this announcement. The Government shall not assume any costs for preparing or submitting any information in response to the market survey or the Government's use of the information. Responses should be submitted by no later than March 30, 2009 to Ann Argaman, Office of Purchasing and Contracts, Clinical Center, 6707 Democracy Boulevard, Suite 106, Bethesda, Maryland . If there are any questions, they may be e-mailed to [email protected].