Speech Technologies and Its Impact on Clinical Decision Making Rose - - PDF document

Speech Technologies and Its Impact on Clinical Decision Making Rose Reedy, BSN, RN, C and JAMES ZADINSKY, MS Case Study: Dwight David Eisenhower Army Medical Center This case study provides an overview of one military treatment facility's (MTF) approach to process improvement leading to business process re-engineering, and the process to monitor and track the effectiveness of change. The case study reviews mapping and monitoring strategies success factors as well as potential barriers to a successful process re-engineering implementation effort. Managed care penetration within the Military Health System (MHS) became significant in the 1990's and has increased over the past years. The need to provide excellent health care in a timely and cost efficient manner as well as the resulting reductions in reimbursement for services forced many MTFs to look at how services were being delivered and what those services cost. In response to regional economic dynamics, Dwight David Eisenhower Army Medical Center (DDEAMC) undertook a process re-engineering initiative called RAD-CSR (Radiology - Computer Speech Recognition). The reengineering initiative, which started in December of 1997, culminated with a commitment to reduce three full time support positions (annual cost to the facility of $ 100K) by 31 March 1998 from the department of radiology. A multidisciplinary quality improvement team (MQIT) was used to help navigate through Radiology’s process improvement phase. DDEAMC dedicated a team of six internal staff members to work as the process improvement team throughout the scope of the project. The composition of the team: ! Health Care Administrator (HCA) ! Officer in Charge (OIC), Department of Radiology ! Non-Commissioned Officer in Charge (NCOIC), Department of Radiology ! Informatics Nurse ! Database Administrator (DBA), Composite Health Care System (CHCS) ! Network Manager for the facility's local area network (LAN). Eisenhower Army Medical Center has adopted the FOCUS-PDCA approach for improving

• rganizational performance. The FOCUS-PDCA approach refers to an approach whereby the

activities of a process are captured utilizing graphical models.9 The FOCUS-PDCA approach provides a systematic means for capturing and representing processes while, at the same time, providing a framework for making changes to those processes in order to maximize opportunities for success. FOCUS-PDCA takes a full life-cycle approach that encompasses strategic plan implementation, process reengineering, and technology deployment all of which are coordinated and integrated through the use of a change management infrastructure. A principal goal of the MTF is to promote process improvement efforts specifically related to improvement in efficient healthcare delivery. Process improvement will be accomplished by improving the integration of

existing methods, tools, and techniques and enhancing their use in patient care delivery through the use of emerging information technologies. FOCUS-PDCA refers to an approach to process improvement that takes a comprehensive approach to improving performance by: ! F focusing on an area to improve; focal point: Availability of Radiology Reports

• rganize an interdisciplinary team: MQIT

! C clarify the process or function to improve: Exam interpretation-report generation ! U understand process variation: Exam interpretation-report generation ! S select the process to be improved; specify: Report Generation ! P plan an approach to implement change; blueprint: Process Re-engineering Design ! D do – implement change; execute blueprint: Project Implementation ! C check – evaluate the changed process; audit: Incident Mapping and Analysis ! A act – continue with the effective change, or re-evaluate and continue new process: Analysis The roles of data and other resources help explain the interdependent behavioral relationships between activities that make up a process. An examination of these interdependent relationships helps one to understand the impact of making a change in one area of the organization. The process, Report Generation, like any process, has its own variables. We evaluated four components:

• Input: Data used to produce an output
• Control: Data that converts inputs to outputs
• Output: Data produced by the activity
• System: Resources that provide energy to, or perform, the activity; typically people

Examples of each of the following components are as follows:

• Input: Physician Order, Procedure Scheduled, Exam completed
• Control: Radiologist views and interprets film and dictates results, transcriptionist enters

report into CHCS

• Output: Radiology Report Verified/Available
• System: Administrative, Radiologist, and Medical Staff

Five areas of process variation were initially identified. These five areas are related to waiting for action, sitting in a queue, thus the need for streamlining the report generation process (Figure 2). Since the department of Radiology decided to make a change to the process, Report Generation, the model required that changes be made not only to the other processes but also to how the transactions occur between those processes. Therefore, the department of radiology now requires the radiologist to directly enter their results into CHCS (i.e., a change in the Control: Radiologist interprets films -> enters results in CHCS; on the System: Administrative and Radiologist-> medical transcription position eliminated and Radiologist enters results by dictating

directly into CHCS; and on the process: Report generation -> dictation of results and verify transcription of report in one encounter) (Figure 3), this, in turn, impacted the process: Radiology Report Completion. With radiologists now required to enter results directly into CHCS, the department of radiology was able to redesign the process so that the review, editing and correction of transcriptions were performed by the dictating/interpreting radiologist. Through the identification and analysis of each process that made up a process and its related components, the department of radiology was able to better understand the relationships between processes and the impact of changes to any of those processes or processes. Understanding these interdependencies had a profound effect on helping the department of radiology successfully manage change. Although the radiologists acquired several steps in the process deemed to be clerical, the end result of faster turn-around time for reports validates the worth of the process (Figure 4). This approach to process improvement allowed us to not only analyze the existing

• perations but also to explore alternative ways to redesign processes to improve organizational

performance. The MHS had proven methods for deploying process re-engineering from inception of the program through implementation planning. However, there were no proven cost reductions monitoring and mapping models that DDEAMC could assimilate for its implementation phase. Literature research indicated that initiatives that failed did so during the implementation phase, where cost reductions were a lower percentage of what was originally agreed to during process re-engineering. There had to be a strategy developed to track the progress of DDEAMC’s process re-engineering implementation phase and cost reductions/avoidance that would result from its re-engineering efforts. Projected cost reductions/avoidance had to be compared to actual implementation practices to assure cost reductions occur. A system was needed to determine where slippage was occurring during its process re-engineering efforts and for what reasons slippage was occurring. It became apparent during the first three months of implementation that

• utputs from the measurement and mapping systems were key to DDEAMC leadership’s success

in achieving cost reduction goals. This case study of DDEAMC Radiology starts with process reengineering and cycles through to tracking and monitoring of approved cost savings ideas that preempts business as usual for more effective services and lower costs. There are important factors to the success of any process re-engineering program. We have focused on those essentials that assure cost reductions occur. The following fundamentals are addressed within this case study:

• Essentials needed for a successful process re-engineering program.
• Approach taken by DDEAMC in defining components for mapping and monitoring.
• Critical reports and monitoring tools used during implementation.
• Communication of reports/steps took to analyze and evaluate implementation.
• Potential barriers to implementation.

• Flaws in the integration process.

Essential Factors in a Process Re-Engineering Program There are four common phases for most process re-engineering initiatives, it is very important to discuss critical issues for each phase of process re-engineering. These key issues, if integrated into process reengineering, set the cornerstone for a successful program. Key issues that had a positive affect for DDEAMC’s success are outlined in each of these phases:

• Process Re-engineering Design: Fall/ Winter of 1997
• Implementation Planning: Winter of 1997
• Project Implementation: Winter/Spring of 1998
• Incident Mapping and Analysis: Summer of 1998

Process Re-engineering Design: Tracking of ideas start in this phase. DDEAMC chose to use Visio (Figure 5) to track concepts from inception, through approval, to implementation. The important point made here is the need to have some tracking system in place before the start of process re-engineering. The database will prove to be invaluable for diagramming and monitoring the status and outcomes of each process re-design idea along its path through the different phases of process re-engineering. Although this phase is where most of the discovery and hypothesis development occurs, it is also where accountability for process ownership lies. Usually, the person identified as accountable for the process also is accountable for cost savings associated with the new process re-engineering. Later, during implementation, accountability becomes very important since change is somewhat hard to accomplish without it, especially where staff reductions are to occur. In DDEAMC’s case, the OIC/NCOIC were chosen to accept accountability for process re-engineering implementation and associated cost savings. Any process re-engineering idea, before it can roll forward to implementation, must have a process where acknowledgements must be obtained from stakeholders indicating agreement that the idea can be implemented. An OIC/NCOIC affected by re-engineering becomes part of the stakeholder approval process. During implementation, the OIC/NCOIC is/are accountable for all status reporting on those process re-engineering ideas that falls within his or her span of control. One of the more challenging parts of process re-engineering is to identify those resources that will be saved or utilized by the organization during implementation which are:

• Non-Labor Costs
• Labor and Benefits Costs
• Investment Costs
• Revenue Enhancement Opportunities

Cost savings must be associated with the specific cost center affected. DDEAMC took the approach of documenting cost savings within the affected cost center, with an impact date indicating when savings are to occur. Investment costs were offset against savings based reduction of 3 full time equivalents (FTE).

Another important task is ranking of ideas by complexity. Reasons why this task is important is to help determine needs for enabling services such as Information Management and to determine how many departments will be affected by process change. DDEAMC had representatives from the various services participate on re-engineering team, involving them allowed for accurate costing of process re-engineering and a mutual understanding of the needs a process change will have on the enabling and other departments affected by the change.

• Visualization: Define Scope of Project: Implementation of Computer Speech Recognition with

front-end editing capability, empowering radiologist to dictate, proofread, verify document in

• ne step. At the same time, enable three full time employees to take advantage of the early

retirement plan and eliminate 3 FTE.

• Goal Setting: Prioritization: Each radiologist required a new PC workstation, training in the

use of new dictation/transcription system, training to upload reports into the hospital's information system, and finally elimination of three positions from the Department Of Radiology.

• Accountability for Process Ownership: Department Of Radiology, with the final accountability

going to the radiologists who have the power to make the system successful or crash.

• Identification of Stakeholders: patients, ordering providers, transcriptionists, radiologists,

computer analysts, and LAN manager.

• Identification of Resources: from time commitments for the stakeholders, through system

implementation and troubleshooting, provider training, etc. Implementation Planning: This phase is where the re-engineering initiative transforms to a project. This phase allows for project planning and brings those identified as accountable for project implementation to the forefront. With these individuals clearly identified as the ideas’

• wners, coordination of services within a project’s timeline can start occurring. The following list

identifies important elements that DDEAMC used to assure successful implementation planning:

• Mutual agreement to start date for the project: 1 Feb 1998
• Identification of milestones, which if not met would stop progress of the project:
• Installation of new workstations: 14 Feb 1998
• Purchase cost of $ 118K
• T&E of radiology module of software complete: 28 Feb 1998
• Radiologist training complete by 31 March 1998
• Release date for department of radiology 3 FTE: 31 Mar 1998
• Annual cost avoidance of $ 100K
• Agreement of support from MTS until 'bugs' worked out of system
• Agreement of support from MTS until all radiologists utilizing the system
• Removal of previous system: 1 May 1998 - slipped to 1 June 1998
• Elimination of support contract from previous radiology system: 1 June 1998

• Annual cost avoidance of $ 15K
• Identification of responsible parties for each milestone step and completion date.
• Identification of each implementation step and date of completion.
• Determination of dates when capital is needed, if any.
• Identification of cost impact dates, when costs can be reduced from the budget.
• Mutual agreement to end date for the project: 31 March 1998 - slipped to 1 June 1998.

Since all MTFs have limited resources to dedicate time and energy to process re- engineering, ideas must be ranked during this phase by identifying those projects with high cost savings and low implementation complexity to initiate first. Ranking projects using cost and implementation complexity as the determining factors will, in most cases, yield the highest savings within the first half of the implementation phase. This phase fosters consensus building around resource utilization among departments and service areas, such as information systems, education, facilities, communications, and human resources. Project Implementation: For DDEAMC, the radiologists did not realize the impact of what was committed to during process re-engineering and implementation planning until the implementation phase began. Up until this phase DDEAMC’s project team worked directly with the radiologists and assisted in planning and coordinating various needs of the teams. Part of the project team’s focus turned to incident tracking and monitoring while other members of the team started planning for the next wave of process re-engineering. Incident Mapping and Analysis: Most process re-engineering programs fall short of their cost reduction goals. Problems usually occur after the implementation phase begins because lack of follow through, accountability, and responsibility to assure costs are removed for the

• perating budget. Also, the opportunity to not reduce costs arises if there are poor measurement

systems in place to monitor cost reductions. Understanding cost variances that can cloud cost savings from process re-engineering initiatives versus volume changes or shifts in services becomes a challenge for those who have poor monitoring and tracking systems. Incident tracking and monitoring is a crucial component, which can afford an MTF successful cost reductions. Components Needed For Effective Mapping and Analysis The hypothesis tracking database mentioned earlier in this case study is the principle tool used to supply up-to-date information on implementation progress. During initial startup of implementation, radiologists offered input to the development of the reports and determined appropriate reporting cycles to update leadership. It was determined by the radiologists that each month during the fiscal year updating of the database would occur and reports depicting each radiologist’s progress would be generated to management. Reports and Monitoring Tools Used during Implementation There are various reports focus groups selected as part of the error tracking and monitoring process which are:

• Monthly Utilization of Medical Transcription Service (MTS) Report —The report shows the

number of radiology exams transcribed by the MTS. (Figure 6)

• Monthly Exam Turn-around-Time (TaT) Report—The report shows average TaT per

radiologist, or per location. (Figures 7 and 8)

• Monthly Incident Report—The report shows problematic or lost films. Radiologists are

required to comment on corrective actions planned to bring the step back on track if slipped. (Figure 9)

• Support Services Tracking Report—The report shows incident in implementation steps

deemed as milestones that have support services, such as information systems. The report is distributed to those services so appropriate action and collaboration can occur.

• Executive Status Report—The report shows results of updated information received from all

team members. DDEAMC’s project team has responsibility for tracking this information and updating the database. Communication of Reports and Steps Taken To Analyze and Evaluate Implementation Progress Communication about progress of implementation is extremely important for all team

• members. Communication about areas that are falling behind allows quick response by

managers and leadership to help keep implementation projects on track and identify those areas that need to be looked at more closely to meet cost reduction goals. The cycle of status reporting serves as the communication tool used to keep all informed about implementation progress, which starts with the accountable manager. It is the manager’s responsibility to review implementation steps in progress, behind schedule, or completed during the month in review. Managers send completed status reports back to the DDEAMC project team for updating of the

• database. Once the database is updated with any activities that change the status of an

implementation step or cost savings impact, the executive status report is run for leadership. Status reported to leadership is based on exception so only those ideas with implementation steps behind schedule are reported. The executive overview report contains analysis of trends that the project team has identified. Certain trends discovered during implementation could have adverse impact to the program cost savings goals. Adverse trends are stratified into two distinct areas, operational and systemic:

• Operational trends: Usually due to problems with coordination between departments or

waiting for a vital component or supply item needed to move the idea forward. Although

• perational trends pose less of a threat to the long-term cost reduction goals of the program,

they can identify areas within the organization that need better collaborative efforts or logistic problems that could be improved. The majority of the operation problems revolved around radiologists becoming familiar with and accepting the new technology, becoming comfortable

with the various dictation input devices, learning the new macro commands available to facilitate the transcription process.

• Systemic trends: These are more serious in nature and pose the greatest threat to the

success of a process re-engineering effort. Systemic issues usually can be tied to poor planning during process re-engineering. Some key elements that are systemic in nature are:

• Stakeholders might have felt pressure to approve ideas they felt were too aggressive.
• Appropriate types and levels of staff were not involved in process re-engineering.
• The idea’s re-engineered process could be very sound, but there might be a lack of
• wnership of the change needed to bring the idea to fruition.
• Continual issues between managers and support services. Support services might have

made resource commitments to assist implementation of the idea but due to other priorities or a shift in the implementation time line, those resources are no longer available. For DDEAMC these revolved around unrealistic timelines for installation, training, acceptance and full utilization of the new system. Not enough time was allowed to accommodate neither individual training requirements, as many of the radiologists were not computer literate, nor the best dictators and had to learn these skills in a timely manner. Potential Barriers to Developing A Successful Incident Mapping and Analysis Program One might have the most appropriate tracking and monitoring systems in place for their process reengineering initiative, but if there is a lack of senior leadership affirmation and support to see the initiative through to fruition, lost cost savings will occur. Having a way to track and report re-engineering initiative is necessary. If there is no means of tracking implementation progress, there will be no way to verify cost reductions attributed to process re-engineering. Lessons Learned During DDEAMC’s process re-engineering phase, the department of radiology had buy-in from all but one radiologist, unfortunately all aspects of the transcriptionists duties were not taken into account prior to eliminating the positions. The other support staff within the department has had to absorb the miscellaneous tasks that were part of the transcriptionists' duties. DDEAMC tried to be proactive, envisioning what would be best for the institution as well as the staff while improving the patient care process. Some delays experienced at the beginning of implementation were due to:

• Conflicts arose between ideas implementation dates and the facilities project dates.
• Delay in getting equipment essential to installation.
• Delayed network interconnection
• Difficulties encountered in connecting to CHCS
• Timing issues with CHCS
• Limitations of CHCS and KEA!, the terminal emulation s/w

• Correction limitations of CHCS/KEA! – inability to work with text within those applications
• Difficulties in connecting an NT network to the LAN
• Slowness in updating files on the LAN due to bottlenecks ie; 10mbps connection vs 100mbps
• Software difficulties
• Speech application issues in customizing the system for physician and CHCS
• Automation difficulties
• Delay in obtaining biometric hardware(fingerprint authenticators)
• Software delays for completing templates
• Too time consuming
• Keystrokes + mouse clicks
• Proof reading, editing - secretarial work
• CHCS Downtime - radiologists unable to dictate to CHCS
• Radiologist verbalize dissatisfaction with need to proofread and edit their dictations
• Training per radiologist averages 5 hours
• CSR accuracy of 95% to 100%
• $118K - 12 complete systems, training for 10 radiologists
• Pentium 300mhz Processors, 192 MB RAM, 512 KB L2 Cache, 40GB Hard Disk

Front-End CSR works for Radiology and promotes and improves the quality, accuracy, and timeliness. Education of responsibilities, actions required to be taken, and by whom during a process re- engineering’s implementation phase is extremely important. Communication through manager and supervisor regularly scheduled meetings help remove barriers that might exist between staff and groups or departments. Using a mechanism to report progress of implementation projects helps keep affected staff in the loop and informed. DDEAMC has continued the process of re-engineering during the fall of 1999. DDEAMC applied the business intelligence solutions gained during the initial process re-engineering; they became valuable tools to improve the chance of success for new initiatives. DDEAMC is also developing a comprehensive approach to mapping and monitoring quality and service indicators that are being identified as important and crucial for the success of patient care and services for the region.