An Institutional View of Electronic Records Management: Hospitals & Teleradiology.

Electronic records and recordkeeping systems are ubiquitous in most organizations. Only recently, though, have records professionals begun to realize their full power, potential, and uses. Electronic records have the power to capture ideas, notes, and communications that were previously invisible or ephemeral events such as fleeting face-to-face or telephone conversations.

Implementations of electronic recordkeeping systems can: (1) provide more people with greater access to information and (2) give them the possibility of putting disparate information together in new, innovative ways -- in essence to create new knowledge (Davenport and Prusak 1998). These two factors affect the appraisal of previously established records series, the way these series function in the workplace, and the expectations for the use of these records by different individuals. The legal and economic implications of such changes are only now emerging.

As more and more paper records take electronic form, records managers need to reassess and reanalyze the records with a careful examination of the new organizational communication patterns and functions that electronic recordkeeping systems support. This article examines the trends mentioned and cites as examples emerging applications in the medical arena, particularly in the area of radiology. It is founded on a research study at a large tertiary care medical center (Yakel 1997).

Traditional Radiological Records: The Generation Processes

Radiologists create and interpret images derived from a variety of modalities: radiography, ultrasound, magnetic resonance imaging (MRI), positron emission tomography (PET), and computed tomography (CT). These radiological processes of image creation and interpretation are bounded by the generation of records. A clinician writes a radiological requisition to begin the process; a signed radiologist's report signals its completion.

The most common records associated with radiology are the anatomical images themselves and the radiologist's report of findings -- the interpretation of those images. Radiological processes, though, encompass a number of records-creating events resulting in the creation of a variety of types of documentation.

These documents include a requisition for radiological services, scheduling calendars, the actual images generated, various forms of the report that may range from indexes to images to be viewed, notes in preliminary books, an audiotape of the report by the radiologist, the preliminary -- or unapproved -- version of the transcribed report, and the final, approved report.

Additionally, other types of records may also be used to inform the radiological interpretation, including the medical record, particularly any previous radiological images and reports. The radiological images and final reports are the focus of this article.

Although numerous doctors' offices have access to a basic radiography (X-ray) machine and perhaps ultrasound technology, these images still must be professionally interpreted by radiologists, who are often associated with area hospitals. Patients requiring imaging techniques, such as a MRI or CT scan, must go to a hospital or special center for these tests. Image capture, then, takes place at diverse locations, and in the latter scenario, the clinician does not have ready or easy access to the images and must rely primarily on the report when making a diagnosis.

In many hospitals, the radiological report has traditionally been created orally using a telephone dictation system. Clinicians would either telephone into the system to listen to a verbal, preliminary, and unauthorized report, or they could wait until the dictation was transcribed and then access an approved, written description and analysis of radiological images.

If clinicians wanted to tie the written report to the images, clinicians often had to request the images to study them in conjunction with the written report. The integration of visual and written data has traditionally been difficult for clinicians. Until recently, radiology has been primarily film-based. Film was generated for all types of images, even the MRI, PET, and CT scans that were originally created digitally.

The Push for Digital Images in Radiology

Both radiologists and clinicians have been interested in exploring the potential of digital radiographic image and reporting systems. Their reasons, however, are very different. Radiologists see digital media as a means for reducing the number of lost or unread studies, more accurate studies through the use of computer techniques for enhancing images, easier and more distributed access to images, and reduced radiation doses (Siegel, Diaconis, Pomerantz, Allman, and Briscoe 1995). Radiologists see the transition to digital imaging and electronic communication technologies as a means of better image management. The problem of lost images is cited in numerous studies as a major impetus for the adoption of digital radiography.

Clinicians have very different reasons to push for digitized radiological information. The primary complaint of clinicians is that radiological information is often delayed (Crabbe, Frank, and Nye 1994; Clinger, Hunter & Hillman 1988). Faster access to radiological information, clinicians claim, would assist them in making more timely diagnoses and assessments of patients.

Although not incompatible, the goals of image management and patient care are diverse and signal the strains put onto any system to be used for capturing, viewing, storing, and communicating radiological data. In fact, some clinicians have noted that the starting point for selecting any radiological system should address whether that application will be used primarily by radiologists for the interpretation of images and the creation of reports or for the communication of radiological data and review of the reports and images by clinicians (Barnes, Morin, and Staab 1993).

Electronic Applications in Radiology

Electronic systems for displaying and viewing radiological information are diverse. In some cases, digital images and reports are becoming part of the electronic medical record (EMR). In other applications, radiological data (images and reports or simply reports) are in a separate radiological information system (RIS). In a basic version of this system, the focus is on radiological interpretation. More elaborate RIS systems are referred to as picture archiving and communication systems (PACS). These systems are used for interpreting images and communicating results.

More recently, the ability to network these systems, particularly over secure Internet connections, has created a new class of teleradiology systems that can serve geographically dispersed clinicians and radiologists and act either in synchronous or asynchronous modes (Welz, Ligier, and Ratib 1995). Whatever the configuration of the RIS, several inherent implications for work practices are embedded in these applications. In turn, these implications affect records and records managers.

No electronic file is the same as its paper counterpart. Overt and planned -- as well as unplanned and surprising -- differences will emerge even if the same information is available in these two media. At a minimum, data are juxtaposed differently on the screen and in analog format, enabling or precluding certain relationships and comparisons. Furthermore, search and retrieval are different and, in many ways, are enhanced in a digital environment, though browsing behaviors still may be more efficient in analog format.

Furthermore, information exchange and communication is different. If radiologists and clinicians are using teleradiology technology in synchronous mode, the exchange of information is quicker and the ability to answer questions more timely. Yet, these technologies can also be used asynchronously -- unlike the telephone or face-to-face conversations, which were previously the principal means for clinicians to get follow-up radiological information.

A problem with new radiological recordkeeping systems is that they must interface or at least coexist with paper and electronic legacy systems. Determining a standardized methodology for incorporating older data into the new system is a key concern for hospitals. Because many systems are proprietary, importing data is often difficult. Although rekeying text or scanning images en masse is not cost efficient, not incorporating key or sufficient information on active patients into the current recordkeeping system can impede medical care and create confusion. Understanding the relationships between sequential and occasionally parallel recordkeeping systems is critical for medical personnel and records managers.

The creation of large hospital systems through mergers also means the integration of multiple recordkeeping systems, each containing diverse information such as radiological examinations, blood tests, and drug therapies given to patients. Acronyms for some tests in radiology, pathology, and pharmacology are not standard across the medical profession. Therefore, when hospitals merge and begin to implement joint systems, one set of acronyms must be selected. How will this selection be done? What stakeholders should be represented? How will staff manage moving back and forth between legacy systems with one set of acronyms and new electronic patient record systems with another set? What is the role of the records manager in mediating between these diverse recordkeeping systems and documenting the inconsistencies?

The ability of electronic systems to capture previously invisible information and activities has two aspects. First, teleradiology systems can send images captured at a remote site to an expert radiologist located in a medical center. The radiologist then interprets the image and sends back a report detailing the findings. This ability to bring expertise to geographically dispersed regions is a major selling point for this technology.

Through e-mail, these systems could also potentially capture consultations, discussions concerning those images, and reports. Previously, these consultations were usually done face to face or over the telephone. Making these conversations visible and part of the organizational memory would have implications for work practices as well as legal implications. Furthermore, radiologists do not currently bill for consultations. If these activities can be captured, does that make them potentially billable?

Second, PACS and electronic medical record systems (EMRs) capture metadata concerning who viewed, annotated, printed, and sent specific data and when those activities occurred. This type of trail would be impossible with paper EMRs or radiological reports. The records management system simply does not and could not capture this information efficiently. This ability to capture and replay aspects of invisible work creates additional information richness as well as new accountabilities.

The Future of Radiological Data: More Complex Electronic Records

In the previous discussion, the assumption is that images are static. Yet in the electronic environment, particularly as bandwidth increases, this assumption is false. In the past, images in a radiological specialty, such as ultrasound, had to be selected and fixed for interpretation as well as for a record. In the increasingly digital world, this type of selection may become unnecessary. The ultrasound record could be transformed from a still image to a streaming video detailing a baby's heartbeat or the flow of blood into a newly transplanted organ. In many cases, this video would be a richer record and perhaps a better record.

Developing teleradiology systems feature videoconferencing as well as applications that transmit annotated images (Beard and Nemminger 1998). Some proposed systems are modeling digital images with a voice-over representing the dictated report. Future systems will also include synchronized dictation and pointer movements over the images as features are highlighted with arrows and other annotations. The integration of these annotations into the formal interpretation of the images contains many elements that the old paper-based systems had, but the new electronic record is substantially different.

Radiological records are also moving towards multimedia presentations of information. These displays will include not only the images and the textual interpretation, but perhaps also audio or e-mail communications between the radiologist and the clinician concerning the radiological findings. Multiple versions of the images, possibly containing sequential annotations added by the radiologist or the clinician as a part of those communications, may be possible as well. If the transmission of the radiological report consistently features annotated images, reports may begin to be written quite differently and could integrate references to a variety of annotations and reference points that were previously omitted.

Implications for Records and Information Managers

Records managers cannot assume that electronic records will simply replace existing paper records. Electronic records are often different and add unique information or previously unavailable abilities to retrieve, compare, and analyze data. For records managers, this metamorphosis necessitates first a need to re-analyze records series in light of their transformation from analog to digital. Second, records managers need to be aware of changing work practices surrounding recordkeeping systems. Alterations in work practices are indications of other changes in the nature and function of the records contained in those systems. Process-centered approaches may help records managers understand the extent and impact of these changes.

The functionalities and uses of electronic records also change as information moves from paper to digital media. Software applications change the relationships between data elements and how they are presented. Records managers need to be able to identify the relationships between data elements to ensure their preservation over time.

However, the relationships between data elements are not the only important shifting relationships of which records managers need to be aware. These changes may also signal changes in the status and power of the records creators and users; in turn, such changes may have their own implications for future recordkeeping systems.

The importance of structural and administrative metadata for records managers in this setting is salient. For example, metadata created to maintain relationships among data elements as well as to track actions concerning the record (accessing, editing, compiling) are key elements in maintaining and ensuring authenticity and reliability. Records managers need to become aware of the various metadata elements underlying applications and their functions in supporting authentic records.

The increasing integration of radiological reports (text) with the actual images also calls into question current records scheduling practices. In many states the textual report is the record and primary evidence that the medical test was completed. Although exceptions may be made for chronic illnesses and some diseases, many states allow the routine destruction of radiological images after a certain period of time has elapsed.

The radiological report, however, is retained permanently and incorporated into the medical record. PACS make integrating the text and image easier. In the case of some teleradiology systems, the image may also actually contain interpretative information, or reports may begin to refer more explicitly to image characteristics that cannot be understood without the context of the image itself. Radiological practices need to be monitored in light of scheduling RIS applications.

Finally, records managers must be ready and able to respond to an expanded purview of what might be considered radiological documentation. The ability of electronic systems to capture what was previously invisible work poses potential social, legal, and ethical problems. Records managers should not leave these issues for others to solve. At the heart of these issues is what constitutes the radiological record in this new environment?

Annotations by clinicians and radiologists after the generation of a report were formerly oral consultations, invisible work not captured previously in any hospital system. Are these annotations now a part of the patient record? Are they nonrecords? Should records managers schedule these parts differently? Can the integrity of this record be maintained if the relationships among its components are altered?

Conclusions

The shift from analog-based to digital-based information systems does not mean that records and information elements formerly in paper formats are being perfectly mapped into the new system. Records analysis must be redone, focusing on work processes and changes in the nature, function, and uses of the recordkeeping system. As a result, record schedules can be revisited and the warrants, legal and otherwise, reinterpreted in light of the new records.

Although the radiological environment does have unique records, laws, and regulations surrounding the recordkeeping and a great deal of standardization, the dynamics at play in the transition from analog to digital radiological recordkeeping systems are not unique. Many of the factors concerning the changing nature, function, and use of records as they are "digitized" occur in other settings and present the same important questions and concerns for records managers.

AT THE CORE

THIS ARTICLE EXAMINES:

* how organizational use of electronic records will force information managers to rethink traditional assumptions about records and records series

* the evolution of the nature of radiological records as they move from analog to digital format

* differences in the functionality and uses of radiological records in their analog and digital formats

REFERENCES

Barnes, Gary T., Richard L. Morin, and Edward V. Staab. "Teleradiology: Fundamental Considerations and Clinical Applications. Radiographics 13, no. 3 (May 1993): 673-681.

Beard, David V. and Brad M. Hemminger. "Remote Ultrasound Using Cooperative Video: A Field Study." Behaviour & Information Technology 17, no. 1 (1998): 10-17.

Clinger, Neal J., Tim B. Hunter, and Bruce J. Hillman. "Radiology Reporting: Attitudes of Referring Physicians." Radiology 169 (December 1988): 825-826.

Crabbe, J. P., C. L. Frank, and W. W. Nye. "Improving Report Turnaround Time: An Integrated Method Using Data from a Radiology Information System." American Journal of Roentgenology 163 (December 1994): 1503-1507.

Davenport, Thomas H. and Lawrence Prusak. Working Knowledge: How Organizations Manage What They Know. Cambridge, MA: Harvard Business School Press, 1998.

Siegel, Eliot L., John N. Diaconis, Stephen Pomerantz, Robert Allman, and Brian Briscoe. "Making Filmless Radiology Work." Journal of Digital Imaging 8, no. 4 (November 1995): 151-155.

Welz, R., Y. Ligier, and O. Ratib. "Design of a Cooperative Teleradiology System." Telemedicine Journal 1, no. 3 (1995): 195-201.

Yakel, Elizabeth. "Recordkeeping in Radiology: The Relationships Between Activities and Records in Radiological Processes" (Ph.D. diss., University of Michigan, 1997).

Editor's Note: Initial data for the research on which this article is founded were collected through NSF Grant ECS-94-22701: A Collaboration Testbed in Medical Image-Based Examination, Diagnosis, and Treatment. Funding for selected transcriptions was made possible through a grant from the Blue Cross, Blue Shield Foundation of Michigan Student Award Program Grant, 1996-97.

Elizabeth Yakel, Ph.D., is an assistant professor at the School of Information, University of Michigan. She has 19 years' experience in information management and specializes in electronic recordkeeping practices and computer-supported cooperative work. She is a member of ARMA International, the Society of American Archivists (SAA), and the Association for Computing Machinery (ACM). She is a fellow of SAA. Her doctorate is from the University of Michigan. The author may be reached at [e-mail].