Each letter in response to our article raises interesting and thoughtful points. Drs. Dolin and Wiesenthal raise concerns that we separated functionality from interoperability costs, whereas Dr. Kretz raises concerns that we applied the SBCDE as a national model of interoperability. We agree that functionality and interoperability go hand in hand, and we did not mean to imply that one could be implemented without the other. However, to model costs, it was conceptually and methodologically easier to approach each set of costs separately. The nation as a whole has not reached consensus about how to approach interoperability. The SBCDE system is one of the most successful early examples of broad clinical data exchange and we were able to access costs for it, making it a reasonable choice. This is, however, only one of the possible architectures that can be used. We agree with Dr. Kretz that it may not be the best approach because the data exchanged were not coded, thereby limiting benefits (1).

Dr. Kretz further states that VISTA should be given significant consideration when developing a model for a national network. Although VISTA has provided tremendous benefits to its population (2), we do not believe that its interoperability approach will be the most useful outside this closed system. Dr. Kretz questioned our assertion that an important financial barrier to wider adoption of health information technology is that benefits accrue to parties other than those who bear the costs. From the policy perspective, who receives the benefits versus who must pay is critical; this issue was addressed by several recent studies (1, 3). It is not simply a question of whether the party bearing the costs accrues benefits greater than the costs. We must also consider the period over which costs and benefits are dispersed and the matter of equitable distribution of costs to those who benefit financially. This is particularly essential for small physician practices, as highlighted by Mr. Doherty. We agree that financial savings from health information technology will be substantial (1, 4).

Finally, we completely agree with Drs. Rothkopf and Jackson about the importance of appropriate user interfaces. For successful implementation of health information technology, we must address cultural issues, speed of the system, appropriate training, technical support, and seamless connectivity (5). All these factors must be addressed to realize the dream of widespread, interconnected information technology and the attendant benefits in safety, quality, and efficiency. The recent experience of tremendous health data problems for Hurricane Katrina victims highlights the urgency of implementing an effective network in the United States.