Custom IVR Applications

This section of our technical library presents information and documentation relating to IVR Development and custom IVR software and products. Business phone systems and toll free answering systems (generally 800 numbers and their equivalent) are very popular for service and sales organizations, allowing customers and prospects to call your organization anywhere in the country. The PACER and WIZARD IVR System is just one of many DSC call center phone system features..

What Is IVR?. An Interactive Voice Response (IVR) processes inbound phone calls, plays recorded messages including information extracted from databases and the internet, and potentially routes calls to either inhouse service agents or transfers the caller to an outside extension.

Contact DSC today. to learn more about our IVR services and IVR application development software.

Natural speech recognition promises to bring unprecedented speed and convenience to the creation and management of medical data

by David Essex

The use of computers to recognize human speech has a decade-long history in medicine. Since the late 1980s, radiologists and other specialists have used expensive, dedicated systems that employ specialized vocabularies to recognize reports dictated in an awkward, "discrete" speaking style. But that picture began to change sharply in 1996 with a technological breakthrough: the commercial release of low-cost, PC-based software that could recognize normal-sounding speech, spoken at typical speed and without unnatural pauses.

So-called continuous speech recognition is now sparking a revolution in the way voice technology is used in healthcare organizations. MIS departments are for the first time taking it seriously as a data-input method. Broader disciplines with equally broad vocabularies, such as general and emergency medicine, are already using dedicated speech recognition workstations to create regulation-compliant reports and computer-based patient records (CPRs) that tie into existing networked systems. Some hospitals have reduced their medical transcription staff, or are outsourcing the work more cheaply with turnarounds measured in hours instead of days.

But perhaps the most important effect of this new age of speech recognition is the boost it is likely to give to the long-held goal of 100 percent electronic patient records. By making it easier for even technology-phobic doctors to enter data quickly and accurately, continuous speech recognition is helping to stoke the electronic pipeline with the critical mass of information needed to ensure that most medical knowledge is available online.

Speech recognition has an unusual number of factors working in favor of its quick, widespread adoption, unlike most new technologies for healthcare, which often collide with organizational, human and technical inertia. Three forces stand out.

Speaking openly

First, continuous speech recognition requires the least amount of staff learning and adaptation of almost any other input method. It's designed, after all, to conform itself to people's most natural way of communicating. Doctors, notoriously harried and reluctant to change their methods without a fight, can typically learn how to input data into a voice-enabled medical record program with only a few hours training. "Physicians choose the path of least resistance," says Octavio Diaz, MD, chief of emergency services at Quincy (Mass.) Hospital, a 290-bed facility that handles 30,000 annual visits to its emergency department. Diaz plans to install three workstations running Lernout & Hauspie's (L&H) Kurzweil Clinical Reporter and MedQuist's transcription software. Doctors will get only five hours of training, plus a 90-day proficiency period during which they will gradually take on more complicated reports.

The dozen or so vendors, resellers and healthcare organizations interviewed for this article reported no significant problems in getting personnel up to speed quickly. (Some of these, admittedly, make speech input a condition of employment.) Setups that outsource the transcription and correction parts of the process are even easier to sell to caregivers, many of whom are used to dictating notes into handheld microphones.

The ability to slide almost seamlessly into existing office systems and procedures is another advantage that continuous speech recognition has over most other information technologies. Numerous medical record systems vendors, such as McKessonHBOC and Eclipsys, have begun adding voice recognition as an optional input method. Similarly, it's a small step for organizations that are already sending dictation recordings to an outside transcription house to route those same recordings to in-house speech recognition stations controlled by medical transcriptionists.

Speech recognition also has a compelling return on investment (ROI) rationale, though few organizations appear to have procedures in place to track long-term costs and benefits. The case requires little evidence, especially if current transcription costs and turnaround times are already known. On the cost side, the PC, voice board, microphone, recognition software and basic medical vocabulary can be had for less than $3,000; systems that add medical record reminders, regulatory requirements and specialized vocabularies cost upward of $10,000. Network bandwidth may also require upgrading to accommodate the increased traffic, and a dedicated phone line is required for off-site transcription and verification.

The benefits compound, however. Per-line transcription costs may drop as much as one-half, thanks to greater efficiency and reductions in staffing--a not altogether positive effect (see "Medical Transcriptionist's Friend or Foe?" page 62). Medical records can be available almost instantly, or more typically, within 24 hours. Caregivers can use time formerly spent on record keeping to see more patients, then bill more reliably--and at the highest allowed reimbursement level. Additional money may be saved by avoiding fines for fraud and noncompliance that can come from having inadequately documented Medicare claims.

This demand for more and better documentation is fueling the immediate demand for speech recognition. The biggest motivator is new documentation guidelines proposed by the federal Health Care Financing Administration (HCFA) and expected to be adopted, after much delay, as early as this spring. Also, some private insurers have begun to refuse payment to claims that don't meet strict documentation rules, according to several doctors interviewed.

Commercial solutions are all built on recognition software "engines" from just four companies: Dragon Systems, IBM, L&H and Philips. Purveyors of dictation machines and devices, such as Dictaphone, transcription companies such as MedQuist, and various CPR vendors now sell speech recognition modules, which typically work with the companies' existing systems. Each of the four engine vendors has either a dedicated sales effort or a department devoted to vertical applications in medicine.

There are no known comparison reviews of the medical versions of the four speech engines, and even the vendors themselves admit their core products provide similar functionality, with recognition rates in the 97- to 98-percent range. Dragon Systems (the first, by just weeks, to introduce a PC-based continuous speech product) works with value-added resellers to build medical applications, as does Philips. In addition, Dragon's flagship NaturallySpeaking software is sold with optional medical vocabularies and a macro language that allows users to build sophisticated records templates, says one satisfied physician who's been using it since last summer.

IBM and L&H both have longer, broader histories of selling into the medical market with specialized workstation-based software and vocabularies for radiology, pathology and other specialties that lend themselves to highly structured medical reporting. L&H sells Kurzweil Clinical Reporter for Windows NT (originally an MS-DOS application first sold by Kurzweil in the late 1980s). L&H has also begun to offer medical versions of its VoiceExpress software, which lets you use speech to control a document's format. IBM's vertical product is called MedSpeak, but Big Blue also sells medical versions of its more mainstream, lower-cost ViaVoice speech recognition software.