Difference between revisions of "What is VistA Really"

Revision as of 21:46, 26 August 2008

VistA is a Electronic Health Record programmed by Federal (US) employees working for the Department of Veterans Affairs (previously Veterans Administration) for several decades, since the late 1970s. Because it was developed by Federal Employees, it is available under a Freedom of Information Act (FOIA) request. The version of VistA available directly from the US Government is usually referred to as FOIA VistA.

But what is exciting about that? What is VistA really? Why do people have such strong feelings about it? This is the WorldVistA wiki, and we tend to be fans of VistA, so here is the page that will give some insight to what is great about VistA.

Disputed History

The history of VistA available from hardhats.org indicates that VistA development began in the late 1970's. VA documents, such as the VistA Monograph date the start of VistA as in the mid 1980's. Why the discrepancy? In short the version at hardhats is the real version. It documents how VistA moved from an underground movement to a celebrated system.

VistA and the vision of VistA was a programmer led counter-culture within the VA. Programmers and administrators often risked their jobs (and some lost them) in order to develop a system that would satisfy clinical needs at the local hospital. Each programmer was primarily interested in the needs of clinical users who worked at the same local VA hospital that the programmer did. However, they were committed to building software that could be used across the VA system.

What is unique about this is that it represented one combined system that was developed by a distributed team. In short, the creation of VistA was a precursor to the distributed "open source" development style way before the concept came into vogue in the 1990's. The best description of this style of distributed development is in a paper called "The Cathedral and the Bazaar" by Eric Raymond.

The VA dates "VistA" from the time when they approved the software. The VA was using VistA long before they approved it, but only in some of their hospitals, instead of across the VA system. In fact, the name "VistA" was not created until much later, the original name was DHCP (for Decentralized Hospital Computer Program). The VA bureaucracy is primarily concerned (as it should be) with the care of veterans. As one can imagine, this is a highly political issue with tremendous pressure from Washington. Enlightenment with regard to the development of VistA is largely independent of bureaucratic ability. When the VA is lead by VistA-enlightened leadership, VistA tends to thrive and move forward. When the VA is led by those ignorant of VistA and software generally, VistA tends to stagnate.

The tension between those who actually make VistA good (the VistA programmers, ADPACs, and CACs) and the bureaucracy in charge of the VA resources still exists. In many ways the purpose of WorldVistA is to ease this tension. WorldVistA is intended to be neutral meeting place for VA programmers, VA bureaucrats and those external to the VA. The only common theme in the WorldVistA community is the interest in VistA and making VistA better.

Licensing of FOIA VistA

Despite the way in which VistA was developed, it is a common misconception that VistA is "Open Source" software. Technically, VistA is in the public domain. As a result VistA can legally be the basis of both proprietary software and Free and Open Source Software. FOIA VistA does meet the strict definition of "Free Software".

The fact that VistA can be proprietary software AND free software creates an ongoing tension between different portions of the VistA community. For instance WorldVistA has a policy of releasing software under the [General Public License (GPL)]. The GPL is a "keep-it-free" license that prevents someone from controlling the software and making it unavailable to others. Some proprietary VistA vendors favor other licenses that allow for proprietarization.

What do you mean by...

The first barrier to really understanding VistA is one of overlapping vocabularies. Often people who are technically savvy have a more difficult time with this than those with less technological experience. This problem is directly related to the age of VistA. VistA developed in parallel with many technologies that we take for granted. Often names for technologies in VistA overlap with the names for more common technologies and give tremendous headaches to new VistA developers.

"RPC" is a good example. When most people outside the VistA Community say RPC they are referring to a common *nix messaging system. This standard meaning of RPC is based on the same basic idea as the VistA RPC mechanism, but otherwise, and in detail, bears no resemblance. CPRS connects to VistA using the VistA RPC methods, so it is an important part of the VistA technology model, but its name often causes confusion with those familiar with the more common meaning of RPC.

There are two things that you can do to avoid being confused: first, do not assume that you know what a "VistA person" means when they use these technological terms, and second, read the Confusing Terms section of this website.

Your programming language is named after a disease?

VistA is written in MUMPS. There are several facts about MUMPS that often confound first time programmers.

• MUMPS is not a C-based syntax, as most modern programming languages are. This means that MUMPS code looks strange to outsiders.
• MUMPS is not white space invariant. This means that you must pay attention to where you put spaces. Many modern program languages allow you to freely put in spaces at any point (following FORTRAN rules originally) With MUMPS you must be careful where you indent using spaces. Putting multiple spaces in some places is not allowed at all, which will really bother you if you like python (which forces indentation as part of the syntax instead of using periods as MUMPS does)
• MUMPS is its own database, but its database is not SQL. It's a Hierarchical Database that is ideally suited to healthcare
• Globals (variables) means something in the permanent database, ie: global across all processes in the system, not the typical programming meaning of global across all the subparts of a single program (variables that are shared within all parts of a single program are called locals in MUMPS).
• MUMPS often behaves like an operating system. Due to space limitations, some original implementations of MUMPS actually were the operating systems on the machine, and this legacy still influences the behavior of MUMPS.

Business people might be a little jarred by some of the business implications of MUMPS.

• There are very very few qualified MUMPS programmers, compared to the much larger number of programmers in mainstream languages. There are more MUMPS people than SNOBOL or many of other programming languages of the same generation, but for the market segment the number of programmers is definitely disproportional. Just because MUMPS is commonly used in healthcare does not mean that it is easy to find non-healthcare programmers who use MUMPS.
• It takes a competent computer scientist much longer to learn MUMPS than other languages because MUMPS maintains backward compatibility to ideas that modern computer scientists don't see except in assembly language programming, such as the GOTO command or line tags.
• Because they are rare, MUMPS programmers are often expensive

MUMPS is too often written off as an obscure language having little or no value. Here are some of the really good things about MUMPS

• Excellent string handling
• The database is really really good, especially for healthcare
• It is unbelievably fast, as it was designed to run on ancient hardware, so it flies on modern hardware.

VistA is a text based (command line) system. CPRS is only the GUI.

Many many people who have used VistA think that CPRS is VistA. Most of VistA is what is known as "roll and scroll" or a terminal application. VistA was designed long before monitors where capable of complex graphics displays. Old monitors that only displayed text were called terminals. Modern computers use Terminal Emulators which allow modern computers to interface with applications as though they were terminals.

Much of VistA is only available on terminals. CPRS is a GUI client written in Borland's Delphi (Pascal) that uses VistA-RPC calls to connect to VistA. Most of the underlying architecture of VistA will not be discussed here, but it is important to clarify that CPRS is only part of what VistA has to offer.

File:Server.png

Many people unjustly dismiss terminal applications. GUIs make learning programs faster, but terminals make using programs faster. Once the initial learning curve is overcome, terminals offer faster data entry and better response times than most modern GUI applications. When given the choice, experienced users often prefer the terminal, and lose productivity when forced to use GUIs. Some of this is because Graphical interfaces use both a mouse and a keyboard, and some of this is because GUIs can change the location of information on the screen, whereas text applications are highly predictable about where they display information.

People who only rarely use a program have to figure out how to use it from scratch every time. For them a GUI is the best program system, because they can use their general knowledge about computer systems to figure out how to use the program anew.

Often clinicians need graphical displays that are not possible in terminal applications, such as for graphing results or seeing trends. CPRS is intended to perform most of the common clinical functions in a GUI environment.

What is CPRS?

CPRS is the dominant GUI for VistA. It takes a small percentage of the underlying VistA functionality and presents it in a graphical fashion. CPRS focuses on those activities that are most frequently used by clinicians. Often those without direct patient contact, such as pharmacists, do not have a GUI based interface, and must use the traditional roll-and-scroll terminal interface. Since so many clinicians are trained with CPRS, and because they often use only this interface, many clinicians assume that CPRS is VistA.

This is not entirely a bad thing. CPRS provides a very well thought out approach to the underlying VistA capabilities. It is relatively easy to use, given the complexity and depth that it contains. A quick look at the Google Scholar search for "VistA" and "CPRS" reveals that CPRS is one of the most studied EHR interfaces available.

CPRS is also very, very customizable. In the VA, the Clinical Application Coordinator (CAC) role, is essentially responsible for the care and feeding of CPRS. A competent CAC is capable of tuning CPRS in numerous ways to the preferences of departments, floors, or even individual clinicians. The CPRS menu is capable of linking to other programs, so that clinical programs that are actually separate applications can appear to be a part of CPRS.

CPRS is often criticized by recent programmers for being "long in the tooth" (that means too old for you Yankees). It is written in Delphi, which is no longer a popular programming language as it was in the mid 1990s. Like the rest of VistA, CPRS was written when computers were much slower, and as a result CPRS is comparatively light-weight. While the community is seeking a Compelling CPRS Replacement CPRS itself continues to run on computers that are too old to run any commonly deployed proprietary EHR systems.

Contact the Editor

This page was originally written and is currently edited by Fred Trotter. If you have comments or suggestions about the contents of the article please contact Fred Trotter directly.