Role of the Next Generation Internet in Healthcare David W. Forslund Deputy Director Advanced Computing Laboratory Los Alamos National Laboratory MS B287 Los Alamos, NM, 87545 voice: (505) 665-1907 fax: (505) 665-4939 email: dwf@lanl.gov Role of the Next Generation Internet in Healthcare The Challenge Healthcare presents a national grand challenge in its requirements to integrate a very diverse and somewhat poorly defined but very high volume information set for easy use by a healthcare provider and potentially even the patient. The trend toward a more distributed healthcare model makes this challenge even greater. Enhancing the backbone bandwidth of the Internet will help make the movement of this information feasible in the future. However, enabling the wide spread delivery of real-time audio, video and image data to the desktop will increase the complexity of the data management proportionately. It will also require further evolution in the remote storage and organization of medical information. This will put even greater pressure, then, on the ability to navigate quickly through and analyze large data sets as might be encountered with complex imaging techniques used to diagnose patients. We believe that the access and movement of large data sets is potentially much more important in telecollaborations than the ability to see one's counterpart on a video screen. Issues and Approaches The current trend in the industry for scalable delivery of information over wide areas is the CORBA (Common Object Request Broker Architecture) technology from the Object Management Group (OMG). It is thought by many to have the best chance of coordinating information between disparate sources and managing the complex interactions (including security) required through its Object Management Architecture (OMA). However, it is known that there are some limitations on this technology in the high performance arena when one attempts to deliver high-volume data rapidly (e.g., Schmidt and Gokhale, SIGCOMM Conference, 1996). The problem of congestion, flow-control and efficient marshalling of objects needs to be addressed even more vigorously if we are to move to the next level of object technology use. Otherwise, the potential gains from powerful multimedia collaborations between physicians over the next generation internet may not be realized. An additional component of this is to extend the CORBA model to include transparent but efficient linkage of highly parallel computational systems. Particularly in healthcare, complex, high volume datasets will have to be judiciously managed to achieve the full power of this new network. The complex nature of the multimedia electronic medical record is an ideal platform to validate the use of the next generation internet. Showing its utility in providing decision support, for example, through remote high-performance computing would be quite beneficial in healthcare. The ability to correlate vast amounts of image and time-based information for timely delivery to the physician is extremely important and valuable to the nation. It could enable a significant shift in the way healthcare is managed by a primary care physician. An example of this is to have high-resolution video of some of the actual patient encounters recorded in the medical record. The judicious balanced application of high-performance server technology with high bandwidth networks with every increasing desktop performance needs to be optimized both for performance as well as cost. The combination of these tools should enable new types of decision support tools to be applied in healthcare. The ability to link these high end resources to low end networking facilties implies some additional challenges to preserve or even enhance quality of service in terms of network availability and response. Our experiences are that network latency is frequently much more important than network bandwidth for collaborative applications. Solutions and examples By building on previous successes in managing complex, distributed, multimedia medical information (e.g., http://www.acl.lanl.gov/TeleMed), we believe it should be possible to utilize the Next Generation Internet to provide secure, real-time analysis and delivery of 3D datasets to the physician using standard (but probably extended) protocols running between high performance computing systems and the high-end desktop. For example, in the effort to track emerging infectious diseases it becomes necessary to correlate data from a variety of sources to determine trends to see if some new disease is appearing. This may require the comparison of 3D CT or MRI data which is stored at a wide variety of institutions. Besides being computationally intensive, this requires high volume data movement in order to do the computations. In addition, the epidemiologist may need to "navigate" through these data sets remotely in order to determine trends that were not previously expected. We see a transition from the current methodology used by the CDC into this new much more data intensive environment in which near real-time assessments of disease trends are being accomplished. Another example, is the real-time MRI work in which brain function is being mapped as the motor functions are being exercised. This requires substantial investment in computational tools as well as high-data rate MRI equipment. The proper use of the next generation Internet could make this diagnostic tool much more widely accessible by enabling the computational tools to be a shared resource used by multiple data acquisition centers. Again the efficient management of the information generated is crucial to its successful use in healthcare. ************************************************** * Stephen Tenbrink email: sct@lanl.gov * * Los Alamos National Lab Phone: 505-667-4935 * * PO Box 1663, MS B255 Fax: 505-665-7793 * * Los Alamos, NM 87545 * **************************************************