Biomedical Imaging and the Evolution of Medical Informatics Foreword Submitted to Special Issue of Computerized Medical Imaging and Graphics on Medical Image Databases Smadar Shiffman, M.S. and Edward H. Shortliffe, M.D., Ph.D. * Section on Medical Informatics Stanford University School of Medicine Stanford, California 94305-5479 (415) 723-6979 E-mail addresses of authors: shiffman@camis.stanford.edu and ehs@camis.stanford.edu * To whom requests for reprints should be sent. -1-
The past three decades have witnessed a remarkable growth in the power and capabilities of computing technology, with a simultaneous transformation in many aspects of our society, including medicine and health care. Modern medical centers have become dependent on computers and networked communications, and specialists in biomedical computing are now highly valued as key contributors to the environments in which care is delivered and health data are managed. No aspect of medical practice has been affected more dramatically than has radiology, with computers now playing a central role in both the generation of images and the management of images and reports after an imaging study has been interpreted. Elsewhere in the clinical environment, computers are playing a prominent role in applications as varied as medical-records management, decision support, patient monitoring, and medical education. The recognition that a body of technical expertise is necessary to support the integration of computers into the health-care process has led to the introduction of formal training programs in medical computing, and to the definition of core scientific knowledge within a new discipline known as medical informatics. Medical informatics examines the effective and efficient use of medical data, information, and knowledge, with computers and communications networks providing a natural technical base for much of what the field involves [1]. Rapid growth in medical informatics has been reflected in large national meetings, in close ties among informatics workers from a variety of clinical specialties (for example, medicine, nursing, dentistry, and pharmacy) for which information management has become a central concern, and in the introduction of degree programs and post-doctoral fellowships at several institutions in the United States and elsewhere. Although both medical informatics and biomedical image generation/ management have experienced explosive growth in the last decade, there has been only limited interaction between the two research communities. Research scientists and developers from informatics and biomedical imaging have tended to go to different meetings and to publish in different journals, in part because the technical orientation and the research questions being addressed have often been different. Recently, however, the evolution of clinical systems and data types, coupled with new requirements for efficiency in the health-care system, is demanding that we take a new look at relations between the fields of biomedical imaging and clinical information systems. Although we have long recognized that the two fields share similar underlying methodologies for manipulation of medical information, the nascent state of the emerging technologies had originally mandated that imaging systems exist independent of their clinical information system counterparts. Thus there was often a duplication of system components and capabilities within a single institution, and incompatibilities that prevented data sharing. Recent developments have created both the necessity and the opportunity to share applications; the use of efficient information-management tools that support the -2-
sharing of information is crucial to help organizations respond to increasing demand for high-quality care at low cost. We believe that this need for efficient information-management tools mandates greater cooperation and collaboration between those who build medical imagemanagement environments, typically in radiology departments or engineering schools, and those who construct clinical information systems, typically in medical informatics groups, hospital information systems groups, or computer science departments. In this article we examine the formation of ties between biomedical imaging systems and clinical information systems in light of the cited necessity to enhance information sharing. Radiology can benefit from attracting individuals with a medical informatics background, just as informatics can benefit by attracting individuals with special expertise in imaging and image management. The potential for linkages between systems that manage images and those that deal with other patient information has been recognized since at least the early 1970s. For example, one early radiology system, developed by researchers at Massachusetts General Hospital, included functions for patient-examination scheduling and for film-library management. The system later evolved to include on-line reporting capabilities and links to a surgical-pathology accessioning system for automated access to confirmed pathologic diagnosis [2]. Developers of early radiology systems encountered several tasks that pertained to management of clinical data, independent of data type. Research issues included design of user interfaces, development of methods that supported structured data entry while attempting to avoid overly constrained linguistic expression, and construction of semiautomated tools for indexing and retrieving information. The traditional importance of images to clinical fields as diverse as dermatology, pathology, gastroenterology, surgery, and radiology assured that image management was destined to play a central role in clinical information systems design and construction once the storage and communications technologies to support such data types had matured. Thus, in recent years, the use of computersupported image data has become widespread in a variety of hospital departments. Such data are also becoming an important component of the computer-based patient record. Availability of the most recent image data to any provider who might need them, independent of place and time, is essential for quality care. Facile access to images is also key to the success of multispecialty teamwork [3], which is now the dominant form of health-care delivery in hospitals and increasingly in outpatient settings as well. To allow access to image data, hospitals are beginning to distribute multimedia workstations throughout their institutions. Developments in high-bandwidth communications are facilitating rapid access to distributed image archives from these workstations. At the same time, powerful workstations are increasingly installed in small offices and homes and linked to high-bandwidth networks that support remote consultation and videoconferencing. -3-
The development of picture archiving and communications systems (PACS) and high-bandwidth networks has resulted in a unified infrastructure for management of the acquisition, archiving, communication, and display of multi-modality image data and video. PACS research has both drawn from and contributed to developments in related areas such as high-bandwidth communications, database management, and integration of multimedia. It has accordingly had an indirect impact on medical informatics research, for example in the areas of telemedicine, clinical databases, computer-aided instruction, and patient simulation. However, PACS research has not yet provided sufficient flexibility to allow efficient sharing of image-related information. An additional foundation needs to be provided for the PACS framework at the level of data organization and database management, which will provide a common basis for representation of heterogeneous knowledge. This foundation must include specifications of data models, an ontology, a structured vocabulary, and links that associate pictorial information with symbolic representation [4], while complying with standards such as HL7 and DICOM [5]. Such a comprehensive knowledgerepresentation scheme will provide excellent support for composition of heterogeneous data within electronic repositories of medical records [6] and for accurate and efficient manipulation of heterogeneous data [7]. The multimedia nature of medical information warrants collaborative work among members of the biomedical imaging, informatics, and clinical communities. Collaboration should build upon common computing methodologies such as database management, pattern recognition, and userinterface design. We believe that such a collaboration will ensure that contributions that are synergistic are incorporated into the development of integrated information systems, simultaneously benefiting each participating field by facilitating its ability to accommodate future changes in the health-care system. For example, medical informaticians can provide expertise to facilitate construction of decision-support tools within the radiology department, and imaging specialists can provide expertise relevant to the management of multimedia data within informatics projects. This potential for synergy has been noted recently by other observers. For example, Kulikowski stated that recent convergence of technological, scientific, and social factors in the practice of medicine is likely to increases the importance of imaging expertise to medical informatics [4]. Formation of laboratories for image-related informatics research within several radiology and medical informatics units reflects a growing recognition of the importance of informatics to imaging research. However, recent participation by scientists at professional conferences suggests that the potential for such synergy between the fields is not yet widely recognized. For example, only a handful of the papers presented at the 1995 Symposium on Computer Applications in Medical Care (the largest annual medical informatics research meeting in the United States, sponsored by the American Medical Informatics Association) discuss processing of medical image information or the integration of images into generic clinical data-management -4-
environments. Similarly, only a small number of papers at the 1995 conference of the Radiological Society of North America (RSNA) discuss informatics issues. On the other hand, the InfoRAD exhibition at RSNA, which covers commercial radiology informatics applications, is an encouraging demonstration of the contribution of informatics concepts to radiology practice. Efforts are necessary on several fronts to stimulate the proposed collaborations. First, training programs in radiology, biomedical imaging, and medical informatics should ideally include in their curricula topics from the other disciplines. Introduction of such topics in seminars, dedicated courses, or internships will familiarize students in each field with the terminology, the research issues, and the culture of the other fields, and thus will help to prepare future experts for interaction with potential collaborators from the related disciplines. Our medical informatics training program at Stanford University School of Medicine has accordingly recently formed relationships with researchers in the departments of radiology at Stanford and at the University of California at San Francisco (UCSF). We have developed an introductory course in imaging informatics, taught jointly by faculty from all three groups, and are encouraging some of our medical informatics graduate students to undertake master's projects or doctoral dissertations in the laboratories of radiology research faculty from Stanford or UCSF. Second, a common language and perspective must be developed, possibly through participation in Internet newsgroups and in special-interest groups within organizations such as the American Medical Informatics Association and the RSNA. A greater emphasis on cross-publication in the scientific journals that cater to the biomedical imaging and informatics communities will also be helpful. The development of a common language and perspective will in turn enhance communication and productivity within multidisciplinary teams. Third, joint research should be formalized through initiatives such as collaborative grant applications. Because much of the research in medical informatics and in biomedical imaging is inherently interdisciplinary, explicit collaborations would strengthen grant applications while helping to ensure optimal staffing for such projects. In summary, the convergence of technologies for image management and clinical data management is leading to the development of a new set of relationships between workers from the fields of biomedical imaging and medical informatics. Interdisciplinary research that bridges these fields can successfully stimulate innovation and investigative productivity while assuring coordination with other clinical computing efforts and helping to promote improved patient care in an era of restricted resources. -5-
References 1. Shortliffe E. Medical informatics meets medical education. JAMA 1995;273:1061-1065. 2. Greenes R, Brinkley J. Radiology Systems. In: Shortliffe E, Perreault L, Wiederhold G, Fagan L, eds. Medical Informatics: Computer Applications in Health Care. Reading, MA: Addison-Wesley; 1990. 3. Lemke H. Future directions in electronic image handling. Invest Radiol 1993;28:S79-S81. 4. Kulikowski C, Gong L. Medical image processing in an era of highperformance computing. In: vanbemmel J, McCray A, eds. Yearbook of Medical Informatics 94: Advanced Communications in Health Care. Stuttgart, Germany: Schattauer Verlagsgesellschaft; 1994. 5. Board of Directors, American Medical Informatics Association. Standards for medical identifiers, codes, and messages needed to create an efficient computer-stored medical record. J Am Med Informatics Assoc 1994;1:1-7. 6. Ratib O. From multimodality digital imaging to multimedia patient record. Comp Med Imaging and Graph 1994;18:59-65. 7. Huang H, Arenson R, Lou S, et al. Multimedia in the radiology environment: Current concepts. Comp Med Imaging and Graph 1994;18:1-10. -6-