Something to Talk About

A Philadelphia healthcare enterprise lowers costs, eliminates paper and raises efficiency by integrating its Web-based voice technology, PACS and clinical system.

When speech recognition technology first appeared in the toolbox of tech-savvy radiologists, Main Line Health (MLH) Radiology was at the forefront. When its existing digital dictation solution needed upgrading and integration with the PACS infrastructure, MLH was prepared for the challenge of taking speech recognition to the next level.

A Philadelphia healthcare enterprise lowers costs, eliminates paper and raises efficiency by integrating its Web-based voice technology, PACS and clinical system.

When speech recognition technology first appeared in the toolbox of tech-savvy radiologists, Main Line Health (MLH) Radiology was at the forefront. When its existing digital dictation solution needed upgrading and integration with the PACS infrastructure, MLH was prepared for the challenge of taking speech recognition to the next level.

 Located in Philadelphia’ s western suburbs, MLH, which consists of Lankenau, Bryn Mawr, Bryn Mawr Rehab, and Paoli hospitals, performs between 400,000 and 500,000 diagnostic and treatment procedures annually, including breast imaging, interventional radiology, neuroradiology, musculoskeletal imaging, abdominal and pelvic imaging, ultrasound, pediatric imaging, nuclear medicine, CT, PET and MRI.

 To deliver on our goal of providing high quality patient care, MLH needed to pair its radiologists’ clinical experience with state-of-the-art technology. By finding, deploying and using the right technology, MLH would be well positioned to carry out its groupwide hallmarks of quality imaging, friendly care, and fast and accurate interpretation of radiologic studies.

Revamping Radiology

 As MLH sought to create a completely digital, filmless and paperless environment, speech recognition was a major component of the strategy. We began by identifying specific vendor and product selection criteria. This included ensuring compatibility with an extensive list of PACS integration providers, demonstrating a track record of cost savings and fast implementation, and verifying improvement capabilities surrounding workflow efficiencies by eliminating radiology paperwork and film. “We needed an application that could integrate seamlessly with our PACS system, so it was critical to find a vendor with an extensive list of PACS integration partners that could provide a system with proven ROI,” says Enterprise Imaging System Administrator Ken Olbrish.

 A project selection committee consisting of radiologists, a lead transcriptionist, system administrators and radiology and information services management identified the scope and business requirements for a speech recognition solution. RFPs were developed and distributed to vendors, which then were evaluated using a predefined metric to quantify and qualify responses to identify two final vendors.

 The project selection committee participated in scripted vendor demonstrations and went on multiple site visits. The site visits proved very valuable in finalizing and validating new workflow changes. The final vendor selection was made after about six months of very intensive evaluation of available systems that provided integration options with our existing systems and workflow redesign. After a thorough review of available speech recognition systems, MLH ultimately implemented the Dictaphone PowerScribe solution from Nuance.

 PowerScribe was installed and integrated with MLH’ s PACS and clinical systems, enabling clinicians to view radiology reports and images digitally through a Web interface across all four MLH hospitals and seven outpatient centers.

 MLH’ s IT and radiology resources worked with Nuance and McKesson to configure PowerScribe to run directly on the McKesson PACS workstation and to allow the radiologists to have full PowerScribe functionality on the PACS workstations. As both vendors already had a strong relationship, this integration was very easy to set up and was done in less than a week’ s time. MLH’ s IT staff installed the PowerScribe client and microphone on each diagnostic PACS workstation in a single evening and the integrated system was ready to go.

Training

 To secure immediate buy-in from our radiology community that would be the actual users of this technology, MLH instituted a 7-week training period and developed a one-on-one vendor interaction program to teach radiologists best practices from day one.

 This included the use of MLH-specific training guides and daily conference calls to review issues and training progress (i.e., site radiology chairmen called-in to report on their site’ s progress and issues). Also, two-on-one training took place consisting of a radiologist coached by a vendor trainer and PowerScribe super-user. The training period was broken down into multiple sessions instead of one long session, consisting of an initial 2-hour session for voice adaptation and basic system functionality with a follow-up 90-minute session for advanced functions and training. Radiologists were also encouraged to self-edit and sign off during the initial training sessions, and editor reviews for accuracy and content of completed reports were followed up by one-on-one evaluations with radiologists.

 Although use of the speech solution was not mandated, and physicians did have the choice to self-edit their patient reports or to use transcriptionists, more than 80 percent began self-editing immediately following implementation. This exceptionally high rate is attributed to the method in which MLH introduced the new system to its users. By training the radiologist team to use the PowerScribe system and by demonstrating the value that speech capabilities can bring to radiologic operational efficiency, MLH kicked off its new speech initiative with a high adoption rate. Since then, the bar was raised even higher to an 89 percent self-editing rate across the facility’ s 39 radiologists and 18 residents.

Today’ s Image of MLH

 In addition to a very high self-edit rate, we improved clinical image and information distribution throughout our multiple locations and reduced transcription time and the subsequent enhanced report turnaround time from nearly 23 hours in 2004 to just 3.5 hours in 2007. Additionally, the number of transcriptionists has decreased from 12 full-time and five part-time onsite manual transcriptionists to just five.

 While the total number of MLH onsite transcriptionists decreased, the remaining five have expanded their responsibilities and operation workload beyond transcribing to include a variety of tasks from basic clerical duties to assisting radiologists with any reporting issues they may encounter. They also provide assistance to many of the supervisors in their daily duties and have cross-trained to provide backup for various other positions in radiology, including reading room coordinator, front desk reception and nuclear medicine department secretary. The transcriptionists provide continuous monitoring to ensure that all studies are reported on and any Q & A issues are resolved in a timely fashion. 

 Even though reducing the number of transcriptionists that were required to be onsite was part of this project’ s goals, one of the primary objectives of expanding our use of speech recognition technology was to make report turnaround time as short as possible, which pleases both referring physicians and patients. “Rather than have our team of radiologists dependant on transcriptionists, we wanted the radiologists to have the control,” says PowerScribe system administrator Ann Marie Gravelle.

 By integrating PowerScribe with our McKesson PACS, our radiologists are empowered to do everything they need to do on one computer. Our team can dictate right into the PACS system, review the report, make edits as needed, and avoid any paper build up, thereby making the process paperless.

 With speech integrated into PACS, our radiologists can immediately access images via the PACS system and dictate into the speech recognition system. Having the information and image simultaneously available allows reports to be generated at high speeds. Subsequently, radiologists can view an image, dictate their diagnosis, sign-off on the report and have it automatically faxed to the referring physician within minutes. The benefit is not only in a shortened report process turnaround time, but also in improved patient care due to that shortened turnaround time. Patient care has been improved with speech recognition since in many cases the ordering physicians have the actual reports before the patients reach their offices or homes.

Value Revealed

 Staff physicians love the addition and use of PowerScribe because they get reports back much faster and can give results to their patients sooner than ever before. In addition to speedy delivery and overall improved efficiency, our radiologists are receiving positive feedback from staff physicians. The majority of radiologists like the new systems because reports can be created, edited, and sent in one fell swoop with no waiting or having to be available later to complete a report.

” We needed an application that could integrate seamlessly with our PACS system, so it was critical to find a vendor with an extensive list of PACS integration partners that could provide a system with proven ROI.”

—Ken Olbrish  Main Line Health

 In addition to improving the efficiency of the radiology team, speech recognition is far less expensive than traditional transcription. Outside of the initial upfront expense, voice recognition dramatically reduces the cost of off-site transcription. In total, MLH has saved on average more than $500,000 per year since its implementation. This cost includes money saved on resources needed for onsite manual transcriptionists, decreased outsourced transcription services for report completion and savings associated with a streamlined employee base.

 Because PowerScribe is Web-based, its architecture enables radiologists and referring physicians to have immediate access to images and reports from any location via Internet. For the MLH community, this means radiologists and editors can go home and sleep well at night knowing their job is completed. Referring clinicians make an important impact on radiology business and speech recognition has dramatically improved the time in which radiologists can turnaround high-quality, accurate reports to referring physicians.

 Because of the success in Radiology, we’ re looking at expanding speech recognition technology to other areas of the enterprise. MLH’ s future plans are to establish its Health Information Management (HIM) department as voice-enabled. This progression is a direct result of the operational and quantifiable benefits that MLH has already experienced. Within a few months, MLH will launch a speech recognition initiative by integrating iChart within its HIM department. The ultimate goal in all of this is to continue to deliver on our hallmarks of quality imaging, friendly care and fast, accurate interpretation of radiologic studies.

SEPTEMBER 2007