The IBM/UAB Computerized Patient Record Project
by Roger Rogers

I want to let you in on some exciting developments with TPF. I have been working on a project that is one of the most exciting efforts that I have had the opportunity to be a part of in my entire career with IBM. The successful completion of this project could have spectacular consequences for TPF and for the medical industry.

We first heard about a potential usage of TPF in the medical industry in October of 1993. We were contacted by an independent consultant who told us that he had approached the University of Alabama, Birmingham (UAB) with an idea for using TPF. The folks at UAB had never heard of TPF (really, imagine that ...). After he had begun some initial work, a project that was kicked off there, he asked us to come down and present to the project director. As John Pilla (TPF Services Manager) and I made the trip I began to think of all the jokes that I could expect (OK, sir, would you like a smoking or non-smoking room, and will that be window or aisle bed ... yes, now you realize to qualify for this rate you will have to book your hospital stay 14 days in advance and stay over Saturday night, ... did I ask about special meals ...). However, we talked about TPF's presence in credit cards, banking, a stock exchange, and ... airline reservations. If there was one thing a computerized patient record needed that we had, it was near 100% availability. The other thing required, not yet realized by many in the industry, was scalability. The doctors we were talking to believed that a "real" computerized patient record would eventually demand the most sophisticated, business-critical, computer solutions available. The match was made, nearly.

In order to bring TPF's strengths to this problem set, we needed to many its strong system infrastructure with the modern capabilities of the workstation and the server. To make a very long story short, we developed an architecture that makes TPF the hub of a three-tiered architecture of workstations, servers, and the TPF object-hub. Objects?! ! ! Yes, with the assistance of the TPF development lab and a team of programmers working full-time on the UAB project, we are developing software that will enable TPF to manage "container" objects in a large complex of servers and workstations. TPF will facilitate the storage and distribution of these objects in a natural partnership with the servers. Obviously, this new work will be valuable not only to the UAB project, but also to all TPF customers and, who knows, maybe a few more customers who "have never heard. There was more than a little skepticism about our ability to pull it off, but most of the skeptics are silent now. Many who were brought in to review the architecture started out highly critical of the choice of TPF (TPF... you've got to be kidding ...), but as we laid out the architecture and the value of capitalizing on TPF's abilities, honed in the intensely challenging environments that so many of you are acutely aware of, the critics went from skeptical, to intrigued, to downright excited.

In October of 1994, UAB officially approved of a project to develop a proof-of-concepts system that will be complete by year-end 1995. By the time you read this article, we should have the hub running and have a radiologist's report flowing from workstation through server, through hub, and back again.

As we incorporate this technology in formal offerings, we will be providing more detail. However, the basic idea is to use TPF as the central repository and distribution agent for objects. We have developed a natural API for use by server programs that will allow the servers to manage the workstations and leave the highly available storage and distribution of information to the hub. This allows application programmers to use all of the capabilities of the servers and workstations to manage and present information and leaves the I/O intensive, and transaction intensive work of storing and distributing the information to TPF.

These developments strengthen TPF's role in the modern IT shop by opening and accommodating its capabilities to new developments. The IBM Worldwide Transaction Processing Solutions organization and the TPF Development Lab are making significant investments in these capabilities to ensure that they bring continued value to our current customer set and to potential new customers looking for solutions to business-critical problems. We're looking forward to a successful TPF entree into an exciting new world, the medical records industry.