|
HL7 standards are the most widely deployed standards across the healthcare industry. Because of these standards, vendors no longer have the headache of custom designing and programming each system for applications. And the healthcare industry has access to best practices for processes such as the collection of patient attributes or a standard set of interesting events.
However, we must not forget that HL7 standards are a set of loosely defined standards. Unlike some standards that enforce strict guidelines, HL7 standards were designed to find a quick solution to healthcare integration woes. As a result, there are many different versions and aspects of HL7 floating around, and interfacing between disparate systems can be extremely difficult.
 Additionally, various parts of the standard are ambiguous at best and include significant options. This leads to different vendors making their own interpretations. As a result, one of the main challenges of integration is simply being nimble enough to adapt to these interpretations.
Here are five key considerations that will help organizations simplify HL7 integration, improving the efficiency of their healthcare information exchange systems:
1. Do the Simplest Translation Possible
The shortest distance between any two points is a straight line. The same principle is true for integration. When integrating two systems, the risk of lost or mis-translated information is reduced tremendously if you keep the translation in the same language. For example, if you are translating text from Quebecois French to Parisian French, it doesn't make sense to translate it into English first. The same rule applies to HL7 integration.
More than 90 percent of interfaces require transforming or translating one version of HL7 v2.x into a different version of HL7 v2.x; basically translating one dialect into another. HL7 v3 is XML-based and a complete rewrite of the standard. It adds huge levels of risk and time to attempt to integrate systems by translating from HL7 v2.x into XML and then back into a different dialect of HL7 v2.x. Keep the language consistent.
2. Use Tools With Good Visibility into the Transactions
Integration is always a moving target with vendors constantly updating their products. These changes/updates often alter their interfaces in terms of what data they send/receive or how it is formatted. When looking for an integration solution, like an interface engine, make sure there is clear visibility around the transactions. When you go live with your interface, the support team must be able to quickly look up all the transactions in to and out of a system. This also helps reduce time and cost of an implementation.
3. Work from Real Data, not the Standard
Integration is a real-world problem, not a theoretical one. Many of the major healthcare systems were built more than a decade ago. These systems are in existence because they work, not because they strictly adhere to the standards. When it comes to real-world interfacing, much like the first tip, the shortest and most accurate path to interoperability is the best. It is important to have the flexibility to work with the real data coming from a system or required by a system.
Some of the catchy descriptions for HL7 include "HL7 is a non-standard, standard" or "HL7 is a great standard, everyone has one." As discussed earlier, the leniency associated with the standards means they should work with the organization's existing data and interfaces, rather than forcing organizations to build something from scratch that can work with the standard.
4. Plan Your Data Model
The goal of interfacing should always be to keep your work, and your interfaces, as simple as possible. By planning your data model in advance, or building a more efficient data model on top of a sub-optimal legacy one, you can avoid the need for multiple redundant mappings within your interfaces. When you standardize and normalize the data, both within your internal data structure and your interfaces, you can easily to maintain existing code and bring new interfaces online. For example, if you store a patient's first and last name in both your lab result and pharmacy database tables, then you've doubled the amount of mappings required for each interface. This type of redundancy can quickly balloon, leaving your integration engineers with massive amounts of glue code to maintain at each site. A well-planned data model will pay huge dividends as the number of interfaces you maintain grows.
5. Consider Your Monitoring Strategy
Half of the work of integration is keeping your interfaces up and running. Make sure the solutions you adopt have support for monitoring your interfaces and notifying your technical staff when problems occur. Ensuring your staff has all the tools and training required to not only create interfaces, but also monitor and maintain them, is the key to the success of your overall integration goals.
Most healthcare organizations can overcome the challenges of integration by adopting these best practices. Hospitals and healthcare networks can minimize the time and cost of integration, ensure real-time access to the information they need and thereby help increase the efficiency and improve overall patient care.
Eliot Muir is president and founder of iNTERFACEWARE.
|