Applied Radiology

Editor's Note

A shift is taking place in radiology from hospitals to imaging centers. In certain cases, the hospital is actually setting up their own imaging centers. Imaging centers provide a much simpler venue for the placement of PACS. This is based on the fact that there is significantly less legacy equipment, and it is often possible to take a fresh look at setting up the appropriate infrastructure and networks. By looking at the challenges that have been faced and overcome by hospitals, imaging center management can learn the right things to do.

This is the first in a series of newsletters that will provide input from experts on the opportunities they have created by overcoming diversity in establishing successful PACS entities.­­Ronald B. Schilling, PhD

Productivity and Workflow in a PACS Environment

Janice C. Honeyman-Buck, PhD

Among the potential benefits of installing Picture Archiving and Communication Systems (PACS) and the Electronic Radiology Practice (ERP) were the possibilities of improving productivity and workflow by eliminating film and paper handling, thus allowing technologists, clerks, and radiologists to do a more efficient job. In theory, the improved accessibility to reports and images would also improve the efficiency of the treating physician and would improve patient care with more timely treatment. Although improvements in productivity and throughput have been reported, PACS is not the panacea for all radiology productivity problems. With careful planning and increasing communication among information systems, however, there is hope that automation will improve workflow and result in efficiencies in all phases of the radiologic examination.

In order to make sense of the reports generated by various authors, it is necessary to define the components of a total ERP and how they fit together to make an automated system. For the purposes of this paper, PACS will be the center of the ERP, providing the capability of acquiring, transmitting, storing, and displaying images generated by radiology modalities. The Hospital Information System and/or Radiology Information System (HIS/RIS) provide accurate information about the patient, his or her location in the hospital or clinic, orders for that patient, medical information on the patient, and the completed radiology reports. Radiology orders are typically placed using the HIS/RIS. The Voice Recognition System (VRS) automatically transcribes the report using voice recognition as the radiologist interprets the study. In practices where VRS is not used, some type of dictation-transcription system is used, generally with the radiologist identifying the patient and study, then dictating the results for a transcriptionist to enter into the reporting system or HIS/RIS. Figure 1 illustrates a simplified workflow in an ERP environment.

THE TECHNOLOGIST'S PERSPECTIVE

Before digital radiography and PACS, the general radiological technologist exposed a film-screen cassette that had to be identified as belonging to a specific patient, then processed either in a darkroom or a daylight processor. Computed radiography (CR) equipment replaces the film with a phosphor plate that can be read, erased, and reused, but still needs to be identified and processed. Other than reducing retake rates because of the wider gray-scale latitude, productivity may not improve using just this technology. However, in a study comparing three departments, two with conventional film-screen and one with CR, the mean examination times for the department using CR was significantly lower than the ones using film-screen. ¹ The study reports the average examination time for a two-view chest study as 12.5 minutes with film-screen and 7.4 minutes with CR, while the average time for a three- to five-view spine study as 19.1 minutes with film-screen and 8.8 minutes with CR. When digital radiography (DR) is used, there is greater potential for improving productivity since the technologist is not required to handle cassettes and can concentrate on positioning and exposure. In one study reported earlier this year, film-screen was compared with both CR and DR. ² The authors reported an improvement in throughput for the digital modalities, with DR being the most efficient. Measured in patient throughput per hour for chest examinations, the film-screen room handled 8.2 examinations, the CR room handled 9.2 and the DR room handled 10.7. The more striking statistic in this work is the time measured until images were available for interpretation. With the conventional film-screen system where films were hand delivered to the reading room, the average time from the start of the examination to the time when films were available for interpretation was 29.2 minutes, for CR the average time was 6.7 minutes, and for DR the average time was 5.7 minutes.

Technologists still have the responsibility of assuring the correct patient information is attached to a study. In a digital system without a connection to the HIS/RIS, this information must be entered by hand in what is frequently a time-consuming process. When there is a connection to the HIS/RIS, usually called a DICOM Modality Worklist, the technologist can swipe a bar-code or select a patient from a list and complete and correct patient information will be associated automatically with the images in the PACS. This saves time and vastly improves the accuracy and, ultimately, the usability of the PACS. Two studies comparing throughput for film-screen with DR with no Modality Worklist and DR with Modality Worklist produced strikingly similar results. ^3,4 In each case, chest examination times were measured and the total time in the department was reported. In the first study, the average time in the department was 307 seconds for film-screen, 142 seconds for DR with no HIS/RIS, and 98 seconds for DR with the HIS/RIS interface. ³ In the second study, the average time in the department for film-screen was 338.9 seconds, for DR with no RIS/HIS was 138.8 seconds and for DR with the RIS/HIS interface was 94.9 seconds. ⁴

A similar study was performed comparing the length of time required to complete a CT study in a "filmless" PACS department with the time in a film-based environment. When technologists no longer printed images at multiple window and level settings, the time required to complete a CT examination was reduced by 45%. ⁵

THE RADIOLOGIST'S PERSPECTIVE

Although cost justification for PACS has a productivity component, radiologists' time is usually not considered because they are usually not considered a cost center by the hospital. However, several authors have investigated the effect of the ERP, particularly PACS and voice recognition, on radiologists' productivity. One study compared the time required to read CTs from printed films with the time required to read from four-monitor PACS workstations. ⁶ Films were printed using a 12-on-1 format and comparison studies were printed and placed in the film jacket for use. Radiologists were allowed to choose their formats on the PACS workstations and window and level presets were used during the interpretation. A selection of chest, abdomen, and brain studies were chosen for the study. There was an overall reduction of 16.2% in the total time required for CT interpretation with soft-copy compared with conventional film. When comparison studies were used, a significantly greater productivity gain was realized. The greatest time savings observed was with chest interpretations, for which the time savings with PACS and no comparison study was 1.79 minutes and PACS with a comparison study was 4.44 minutes.

Hayt and Alexander ⁷ evaluated the effects of PACS and voice recognition on radiologist's productivity. They found that the combined systems resulted in a decrease in radiologists' productivity. The radiologists were asked to describe how the system effected their reading times, and were given the options of less time, the same amount of time, or more time spent in increments of 25%, 50%, 100%, and 200%. Ten radiologists responded, 2 reported an increase of 25% and 8 reported a >100% increase in their time. However, there were a number of striking positive benefits of the system. The percentage of unreported cases at the end of each month dropped from approximately 25% to 0.3% with the ERP. Clinicians who responded to a survey reported that PACS had saved them approximately 30 minutes per day. With the installation of the voice recognition system, the percentage of reports present in the HIS within 12 hours after dictation increased from 3% to 42%. After PACS was installed, 50% of all examinations had reports available within the HIS within 60 minutes, 86% available in 12 hours, and 96% available in 24 hours. One of the weaknesses in the system studied was the lack of integration between the PACS and voice recognition. Radiologists bar-coded an accession number to bring up the voice recognition software, and then chose a study from the PACS on a separate workstation. A close integration of the two systems would improve productivity by eliminating the extra steps.

THE BOTTOM LINE

An ERP can improve productivity in the right environment for technologists and for medical personnel who need to access images and reports. As reported by Hayt and Alexander, ⁷ a marked improvement in the availability of reports was seen, primarily at the expense of the radiologist. Another benefit of PACS was described by Becker and Arenson, ⁸ who studied the correlation of the availability of images and the mean time to drug therapy. When clinicians had access to digital images, the mean time to drug therapy was 3.3 hours, with film the mean time was 4.7 hours.

Several authors have commented that ineffiencies arose when the PACS, HIS/RIS, and voice recognition systems were not integrated closely. When patient and study information is not available on the modality through DICOM modality worklist, technologists must enter the data manually, which is time consuming and introduces errors. When radiologists must access two separate computers for voice recognition and PACS, they double their searching time and can introduce reading errors by selecting inconsistent image and report data. Paper requisitions and reports are still the norm in many radiology departments, which leads to more inefficiency. When PACS is implemented, many of the routing and prefetching operations must be automated to achieve a high degree of productivity. With full automation, Siegel and Reiner ⁹ reported increased efficiencies for technologists of 20% to 60%, clerical staff of >50%, and radiologists of >40%.

A tight integration of all components of the ERP is necessary to achieve overall improvements in productivity. The Radiological Society of North America (RSNA) and the Healthcare Information management Systems Society (HIMSS) have collaborated to sponsor a "phased series of public demonstrations of increasing connectivity and systems integration," called Integrating the Healthcare Environment (IHE). ^10,11 The IHE demonstrations have encouraged HIS/RIS and PACS manufacturers to agree on ways to use the existing standards to facilitate interaction. Only when we have an integrated ERP with all components communicating and a workflow management layer on top will we truly realize the productivity improvement potential of these new systems. n

REFERENCES

1. Reiner BI, Siegel EL. Technologists' productivity when using PACS: Comparison of film-based versus filmless radiography. AJR Am J Roentgenol . 2001;179:33-37.

2. Andriole KP, Luth DM, Gould RG. Workflow assessment of digital versus computed radiography and screen-film in the outpatient environment. J Digital Imaging . 2002;15:124-126.

3. May GA, Deer DD, Dackiewicz D. Impact of digital radiography on clinical workflow. J Digital Imaging . 2000;13:76-78.

4. Dackiewicz D, Bergsneider C, Piraino D. Impact of digital radiography on clinical workflow and patient satisfaction. J Digital Imaging . 2000;13:200-201.

5. Reiner BI, Siegel EL, Hooper FJ, Glasser D. Effect of film-based versus filmless operation on the productivity of CT technologists. Radiology. 1998;207:481-485.

6. Reiner BI, Siegel EL, Hooper FJ, et al. Radiologists' productivity in the interpretation of CT scans. AJR Am J Roentgenol . 2001;176:861-864.

7. Hayt DB, Alexander S. The pros and cons of implementing PACS and speech recognition systems. J Digital Imaging . 2001;14:149-157.

8. Becker SH, Arenson RL. Costs and benefits of picture archiving and communication systems. J Am Med Informat Assoc. 1994;15:361-371.

9. Siegel EL, Reiner BI. Work flow redesign: The key to success when using PACS. AJR Am J Roentgenol . 2002;178:563-566.

10. Dreyer KJ. Why IHE? RadioGraphics. 2000;20:1583-1584.

11. Channin DS, Parison C, Wanchoo V, et al. What Does IHE do for me? RadioGraphics ; 2001;21:1351- 1358.

Practical Issues

Digital Detector Systems Technology Part I: Requirements

J. Anthony Seibert, PhD

Note: Part 2 of this article, Digital Detector Systems Technology: Comparisons, will appear in the second issue of this newsletter.

There are a variety of detectors now available for acquiring large-area digital projection radiographs for diagnostic medical imaging. Large field-of-view image intensifier-television (II-TV) camera systems, computed radiography (CR) photostimulable storage phosphor detectors, charge-coupled-device (CCD) and complementary metal-oxide semiconductor (CMOS) cameras optically coupled to a phosphor scintillator, and thin-film-transistor (TFT) two-dimensional arrays directly coupled to scintillator/photodiodes or semiconductor detectors are examples. Unlike analog screen-film detectors that are contrast limited in operation, digital acquisition devices are inherently signal to noise ratio limited, which means that the image quality is dependent on the quantum statistics of the image formation process as well as proper post-processing. In addition, digital systems aim for high spatial resolution and high detective quantum efficiency to achieve low radiation dose. Some digital systems do this better than others; however, cost, portability, handling and practicality are issues that must be considered prior to purchase and implementation.

The promises of the past are quickly becoming the reality of the present when it comes to implementation of digital detector systems and movement to a filmless, all-electronic image distribution and database system. As 60% to 70% of image volume in a typical radiology department is generated by projection radiography, a key consideration for implementation of a Picture Archiving and Communications System (PACS) is the purchase and deployment of digital radiography devices. Questions one may ask are: What type(s) of digital detector(s), how many systems are needed, what are the basic equipment costs, and what peripheral components are necessary for PACS integration? How do digital systems differ from the conventional screen-film detectors that they replace? What considerations are necessary to verify optimal performance, preventative maintenance and longevity? To answer these questions, this article will present a brief review of screen-film detector limitations and the requirements for digital detector systems. Part 2 of this series will present a description of current digital system attributes and functions that can improve medical image throughput and quality.

LIMITATIONS OF SCREEN-FILM DETECTORS

Screen-film detectors have served the medical imaging community very well for more than 100 years, and technological innovations have yielded continuous improvements in image quality. Intrinsic limitations with screen-film radiography have all but halted any future significant improvements achievable with this technology. These include limited exposure latitude (which impacts the ability to simultaneously provide adequate anatomic information in the thinnest and thickest parts of the imaged object), chemical processing (which is often a weak link in the overall imaging system and is becoming an environmental problem with waste disposal), inefficiencies in handling and storage (which requires a huge infrastructure of space and people), and lack of image postprocessing capabilities (which imposes a penalty on information extraction for diagnosis and requires optimization tailored to the detector, not the examination). The film also serves as the acquisition, display and archiving device, which often necessitates a compromise in one or all of these "coupled" tasks. With decreasing costs and improving functionality, digital detectors designed for projection radiography can potentially overcome these limitations.

REQUIREMENTS/CAPABILITIES OF DIGITAL DETECTORS

Design criteria for digital detectors must, at the minimum, emulate the capabilities of screen-film systems in terms of spatial and contrast resolution as well as field-of-view (FOV) requirements. A 400-speed screen-film detector achieves 5 to 7 line-pairs per mm (lp/mm) spatial resolution (corresponding to an equivalent pixel size of 100 to 70 µm), in a wide variety of format sizes tailored for the examination. For mammography, the screen-film detector provides 15 to 20 lp/mm (33 to 25 µm pixel size) in 18 cm ¥ 24 cm and 24 cm ¥ 30 cm field areas. Thus, the requisite spatial resolution and FOV of a digital detector depends on the imaging task. Clearly, for most general diagnostic imaging examinations (except possibly digital fluoroscopy) a minimum pixel size of 200 µm (2.5 lp/mm) is required, and preferably 100 µm (5.0 lp/mm). Digital detectors for mammography require greater resolution; currently, systems with 100 µm to 50 µm sampling pitch are available, but still deliver less than the corresponding screen-film capabilities they are replacing. One fundamental advantage of digital detectors is flexible contrast and spatial frequency enhancement, which the human observer perceives as higher spatial resolution, despite the actual spatial resolution limit. On the other hand, many digital detectors have a fixed FOV (usually large size for digital radiography, e.g., 35 cm ¥ 43 cm, and usually small size for digital mammography e.g., 18 cm ¥ 24 cm), which can limit their effectiveness in a general radiography room because of positioning issues and significant "wasted" image areas.

Another important image quality parameter is contrast resolution of the output image. When the continuous analog signal is converted to a digital signal, in addition to spatial sampling errors, conversion to a finite range of digital numbers (quantization errors) occurs. With insufficient sampling and/or quantization steps, increased "digitization" noise reduces the signal to noise ratio (SNR), which is the ultimate limit to the amount of contrast enhancement that can be achieved. In addition to spatial sampling, the analog to digital conversion accuracy depends on the number of "bits" of the analog to digital converter (ADC). Generally, most digital detectors require 10 to 16 bits of information (1024 to 65536 unique digital numbers-graylevels) to maintain the ability to achieve contrast detection differences that are comparable to screen-film detectors. In some digital systems, non-linear amplification (e.g., logarithmic) of the analog signal prior to digitization effectively distributes the digital numbers to minimize quantization errors over the full range of signal. Most digital systems use 10 to 12 bits (4096 graylevels) in the final output image, such that each pixel requires 2 bytes of storage space.

With a properly designed digital detector, the incident X-ray exposure over a clinically relevant range (e.g., an exposure similar to the screen-film detectors that are being replaced) should be the major contribution of the system noise (other noise sources include digitization [ADC] noise, scintillator structured noise, electronic, and amplifier noise). When X-ray quantum fluctuations dominate the overall noise, variations in incident exposure will proportionately change the SNR by the square root, according to Poisson statistics that describe the process, and the system is said to be "X-ray quantum limited." Digital post-processing of the acquired image data can increase radiographic contrast only to the point that noise becomes objectionable, such that the output is "SNR limited" rather than "contrast limited" as with screen-film detectors.

In a digital system with sufficient bit depth and resolution, increased SNR can be obtained by simply increasing the radiation exposure to the detector up to a saturation level or to a point where other noise sources begin to dominate, but the cost is increased exposure to the patient. Quantitative detector analysis can determine system performance using objective measurements. The presampled modulation transfer function, MTF(f) is a measure of the detector object transfer efficiency as a function of spatial frequency (the signal). The noise power spectrum, NPS(f), is a measure of the noise characteristics of the detector (the noise). When scaled by a conversion constant, a ratio of MTF ² (f) to NPS(f) generates the noise equivalent quanta, NEQ(f), as a function of spatial frequency. This is an estimate of the equivalent number of X-ray photons per unit area (usually mm ² ) that the detector effectively responds to. With ever-increasing incident exposure, the NEQ increases to a point at which the system saturates and/or other noise sources dominate. The NEQ(f) then decreases. At a given spatial frequency, the NEQ represents the output signal to noise ratio, squared: SNR ² out.

A measure of a detector's efficiency in capturing a given incident radiation flux is a ratio of the effective number of photons detected per unit area to the actual number of photons incident per unit area, the detective quantum efficiency, DQE. The DQE is calculated as a function of spatial frequency as: DQE(f) = NEQ(f)/q. The incident radiation flux, q, is usually determined from computer simulations or estimates of beam energy and penetrability for a given exposure. From the above description, the DQE is also calculated as: SNR ² out / SNR ² in. A perfect detector system has a DQE(f) of 100% at all spatial frequencies. "Real" systems lose efficiency over smaller areas (at high spatial frequency) due to the inability of the detector to efficiently capture X-ray information and/or have additive noise, such as electronic amplification noise, pixel drop-out, or phosphor "structure" noise (among other sources), all of which can mask the "true" signals. Some digital detectors can use the incident radiation more effectively, and thus can reduce the exposure to the patient for a given SNR. In general, DQE(f) measurements for digital detectors range from <10% to as high as 80% for large area objects (low spatial frequency range). The actual DQE depends on the detector design and X-ray converter characteristics. For high spatial frequencies (detail information) the DQE drops to the point at which the system can no longer retain the identity of a small area input signal. To be useful, it is important to determine the incident exposure range over which the DQE is quoted because some systems operate more effectively over a given exposure range (e.g., mammography requires much more radiation exposure than conventional imaging). Knowledge of the DQE, NEQ, NPS and MTF for digital detectors allows objective comparisons that can assist in the determination of appropriate and reasonable performance for a particular imaging application.

THE FINAL WORD: PART I

The minimum technical requirements of a digital detector system are based on the optimal characteristics of the screen-film detector, which represents a huge challenge for system designers. There are several digital detector system technologies currently available that have achieved such status, and have demonstrated the ability to go beyond the capabilities of the analog film image by using to advantage the separation of acquisition, display and storage of the x-ray projection image. In Part II of this article, a comparison of the available digital detectors, including specific implementation details, cost, and use issues, elucidates the benefits and capabilities of these systems in a clinical environment.

Economic Issues

Transition Strategies for New Technology

Ronald B. Schilling, PhD and Edward V. Staab, MD

Often the introduction of new technology into an existing radiology service happens with the recognition of the opportunity to provide better service and increase profits. Occasionally, this is prompted by outside influences from clinicians or administrators. If the new technology is mature or relatively focused, as is the case with most modalities, then few individuals are involved and the introduction is straightforward. But what happens when the new technology is not as mature, involves much of the staff in the department, or perhaps changes the way the business of radiology is done?

There are several examples that have taken place in recent years. Radiologists have experienced such major changes with the introduction of a picture archiving and communications system (PACS) or a radiology information system. Both of these have far-reaching implications for radiology department staff as well as those that the department serves. The merger of hospitals and service entities is another example. To a lesser extent, but still important to a number of constituents, is the introduction of a voice recognition system or the development of a telecommunication service to remote service areas.

This article will outline the strategies often employed, either implicitly or by thoughtful planning of each step, for such transitions. The less carefully planned process usually leads to several false starts. It is likely that the full benefit of a complex new technology will not be realized for years, if ever, without careful planning. A series of tools to help with the strategic plans for implementation of a new complex technology has been published as a series in this journal. ^1-4

The final step of introducing a new technology involves a careful consideration of transition timelines. This is euphemistically described as the "80-mile up" view of the entire process. But before that view, let's look one more time at the process of strategic thinking.

STRATEGIC THINKING

Thinking and planning are everyday activities. The sequence in which we complete our thinking and planning steps is an important factor in achieving maximum effectiveness. Ideally, we should devote an adequate amount of time to thinking about all sides of a particular problem, before progressing too far with planning or implementing a solution. Many of us learn this lesson the hard way. We sometimes forge ahead with elaborate plans, before focusing enough attention on our ultimate goal. We have to go back to the beginning to rethink our options, before resuming the planning and implementation processes.

Strategic thinking tools (or frameworks) can be used to aid the process of strategic thinking. One of the most attractive attributes of such tools is that they can be understood and applied by radiologists without any formal business training. The broad acceptance of these tools comes from the fact that they are easy to learn and highly effective. Essentially, this is because the tools communicate at a fundamental level between people. For example, in a regional healthcare community with people of diverse backgrounds, these tools enable everyone to cross boundaries and relate effectively at a common, fundamental level.

A specific tool known as the "Growth Development Curve" (GDC) will be used to present the "transition strategies for the radiology team." The GDC is a tool or framework for thinking about how the activities of an organization or department change as a function of time. In a well-run organization or department, all members of the team require the same understanding of progress versus time, and where the focus of the entity is at any point of time. In this manner, all parts of the team can work in synchronization and the synergy between elements of the organization or department can be optimized.

The framework for the GDC is simply a series of curves as shown in Figure 1. Each curve represents a different area of focus. It is evident that the areas of focus occur in sequence and may contain a certain amount of overlap. This is a very important aspect of the GDC. The team must determine what the areas of focus are, and the order, start time, (and completion time) for each of these areas.

A well-managed team will review the GDC several times a year. The head of each department should be represented at a team meeting, all seated around a table. When the dialogue becomes focused on the prioritization of resources required to complete each area of focus and the interaction of items between areas of focus, the result is a set of activities that fit together, and are the most important areas of activity for the team. The activity list by department can be managed using the "Sheet of Music." ⁴

TRANSITION TIMELINES

The GDC helps us focus on the issues that need to be addressed to determine the transition timelines.

Figure 1 lists five major areas that must be considered in sequence to introduce a major new technology successfully. Several of the more important subtopics that must be addressed within these major timelines are discussed below.

Understand the business--Under the topic "Understanding the business," it is important to:

1. Identify the leader or champion of the new technology. This may be a team.

2. Evaluate and quantify the current operations.

3. Set functional goals for the system or technology.

4. Recognize the amount of effort needed to communicate the changes that will take place in the environment.

5. Identify all individuals affected by the new system and to what extent they will be affected.

6. Establish cross-disciplinary needs and opportunities.

7. Establish an organizational structure.

Technical process-- The second curve illustrates that soon after you begin the process of really understanding the business and setting the framework for success, you must begin to look at the technical issues. Notice that each of these curves overlap and are continuous.

Under this topic we suggest that you determine what is available in the market place today and to what extent that will meet your needs. Depending on the complexity of the technology, it may be useful to use outside consultants to assist with the evaluation and education of the internal group. It will be necessary to include all-important stakeholders early in this evaluation process so they can understand why you have decided on a particular technology or vendor.

You will also need to plan the transition by asking the usual questions: who will be affected, how will the technology impact their work, and when will this take place. Two other questions that may be appropriate is where will the transition take place and when will it occur. It is possible that the transition will be in multiple phases, as we have learned from the introduction of PACS into a film environment.

Purchase process-- The purchase plan will begin once the acquisition appears to be a serious request. Frequently, this will include the development of a request for proposal. Although not always necessary, the process of developing such a request will raise questions not otherwise considered. The business plan will include the identification of capital and consideration of lease versus purchase options. There are numerous issues that must be considered at this stage, which are beyond the scope of this presentation.

During these stages, site visits and vendor presentations will usually take place. One technique that helps with comparing vendors is to invite them all to present their products and system on the same day. We have convened a stakeholder group at an offsite meeting and listened to each vendor, usually at hourly intervals with time added for questions. This allows the group to formulate many questions and understand the nuances of the differences between vendors more clearly.

Installation-- Assuming a decision is made and that the purchase goes forward, the next phase is the installation. This is when the work really begins. A formal plan for implementation of the technology should be established. It will be necessary to consider the staging of the installation, the delivery and storage of the equipment, the validation and acceptance of the technology, and, finally, the monitoring of the system for reliability. A training program that is simple, available, and repeatable will need to be established. The need for celebration at intervals from "kickoff" to completion of major steps should be planned.

Monitor and revisit-- The various parameters that will be used to monitor the effect of the new system should have been considered. Usually, these are identified in the first phases of this process. If the results are not what had been anticipated, then further evaluation should ascertain what might be done to increase the effectiveness of the system. Methods to solve small problems, such as help desks and trained service technicians, will have to be established.

CONCLUSION

We suggest the following steps to create the GDC and initiate a well thought out plan of action:

1. For the technology introduction under consideration, identify the people or groups who will make up the transition strategy team.

2. Using team brainstorming, determine an initial set of major areas of consideration that are required, in sequence, for the technology introduction.

3. For each of these major areas, assign a leader to investigate the necessary requirements to complete the designated area.

4. Using team brainstorming, review the results of step 3 and determine if modifications to step 2 are required. Establish consistency between steps 2 and 3, resulting in completion of Figure 1--including percent completion and timing for each major area.

5. With the results of step 4 as a roadmap for the technology introduction, the transition strategy team can now prioritize activities leading to a plan of action.

6. The action plan, including all of the base assumptions leading up to it, should be monitored and reviewed on a periodic basis.