Distributed Systems

Index - Major Sections
Home

_______________

Index - Same Level Subject

Index - Child Subjects

Introduction

In discussions concerning distributive processing, the topic soon turns to “middleware,” Common Object Request Broker Architecture (CORBA), Interface Definition Language (IDL), and many other equally scary acronyms. However, in the new paradigm, there is no need for this elaborate subterfuge.

As an example, HL7 is one of the best known standards in the United States and was developed to facilitate the exchange of information between health care organizations. However, after 13 years, it “still has interoperability limitations that result in proprietary implementations utilizing interface engines to meet the need of connecting smaller systems into the domains of the large mainframe repositories.” (Forslund & Kilman, 2000) In other words, it does not work.

When data is transfer only between one client and one server in a straight forward transfer, the problems of requiring data to be transformed is minimized. Almost all-recent technology is capable of using the lowest denomination of TCP/IP and ANSI text files.

Because many developing countries have only recently begun building their programs, they are luckily enough not to have made a big investment in older technology.

Problems with Distributed Systems in Health Care

Difficult to develop Metadata file or Source Dictionary
Bandwidth intensity
Most patient data that is stored is "off-line"
If data is "on-line" querying a production systems is a no no.
necessary to have security key (login, password) for each and every data access across locations

New Paradigm

Even if distributed systems worked to access data, it was incapable of providing decision support or research opportunities. Distributed systems can access data and read from that record, but so what? You can read a particular patient’s record at a particular location, but it was extremely difficult to analysis that patient’s information across all locations and for all times. In order for that to happen, it was necessary to “gather up” all the data and bring it back to one central location to perform the analysis. And now, if we tried this with more than one patient across all locations and across all times…well, we could be out of luck. Big difference in only being able to access data on a global basis and being able to use that data for decision support and research. If used only for access, only small amounts of data need to be transfer, but if that data needs to be analyzed ALL the data from ALL locations will have to be transfer to one location.

What is needed is the ability to use data to make informed, evidence based, decisions, about health care.

Clients move about and attend different health care organization units (HCO’s). This data must be tracked in order for governments to be able to control their health care budgets. To be able to fight the new emerging and reemerging diseases and to provide care that is effective and efficient “there must be a longitudinal medical record to establish long-term effects of slowly evolving diseases such as Hepatitis C and HIV, not to mention management for chronic illnesses like diabetes…” (Forslund & Kilman, 2000). This should not only applied to the one client, but to all clients across all platforms and across all times. A distributed system can not do this.

The only way to do this is by setting up a Centralized Data Warehouse (CDW).

Today with the evolution of the WWW, data replication, “store and forward” technology, and distributive processing, it has become possible to integrate all the key elements and data into one data warehouse without a large investment. For only the cost of a simple low-end computer and a Web browser, any HCO’s can publish their data to a Web Server and centralized data warehouse anywhere in the world.

Back to Top