This post is a continuation; Part 1 is at this link.
In this post, I discuss what's needed to make EHRs more satisfactory and answer questions about my claims.
What’s needed is a paradigm-busting shift in which clinicians and software developers collaborate to create tools that are truly clinically useful and promote (rather than destroy) efficiency. We can start by redefining Meaningful Use – see my blog post at http://curinghealthcare.blogspot.com/2009/05/defining-meaningful-use-of-health-it_02.htmlI then answered a question about my claim that over-reliance on XML and Web services is a big problem with HIT standards:
I'm not saying that XML can't work; just that XML is being over-used. When data sets are rather simple or XML is used for sending messaging operation instructions for web service, I have no problem with it. But when it comes to representing complex data, as in a CDA-based document or a complex form definition file, XML sucks!
Here’s why: When you use XML in such cases: human readability is a joke, verbosity is huge, and parsing requirements and complexities are immense. It also forces conceptual incongruities just to keep data in hierarchies when such parent-child relationships are unnecessary. See http://curinghealthcare.blogspot.com/2009/12/dueling-data-formats.html and http://c2.com/cgi/wiki?XmlSucks.In fact, I’ve seen XML-related complexities bring the Query Health initiative to its knees (see http://curinghealthcare.blogspot.com/2012/08/disruptive-innovation-in-health-it-path_8.html).
I fought against this "let's use XML for everything" mentality since the mid-90's, but to no avail. Now the crap is hitting the fan!
Bottom line is that there are much easier/simpler and more rational ways to represent complex data than XML, even hierarchical data. We've got to start busting maladaptive paradigms.
Regarding the web services issue, I have no problem with this architecture for the most part, but bringing SOAP and RESTful into the DIRECT Project, and relying on them exclusively in other Federal initiatives, adds similarly increased complexity, as well as vulnerability, as I discuss at http://curinghealthcare.blogspot.com/2012/07/oncs-direct-project-in-defense-of.html.
What we need is a hybrid approaches that balances web services and XML with simpler elegant methodologies.
These issues came to a head during the standards-making process at the S&I Framework Structured Data Capture (SDC) initiative where I fought for the inclusion of a much simpler approach to capturing data using electronic forms, but was voted down by the community (see http://wiki.siframework.org/Canddiate+Standards+List+Feedback). To their credit, however, they are allowing our team to pursue a pilot (on our own dime) even though it will demonstrate use of technology/methods that aren’t fully compliant with the implementation guide they are developing.
In any case, the kind of change we need isn't going to happen by EHR industry lobbyists or the Feds ... As I said before, it's got to be led by clinicians and researchers working closely with creative software developers (including nimble EHR/HIT vendors) who focus on designing and deploying clinical tools that are truly useful and useable.In part 4 at this link, I discuss a practical vision of a next-generation EHR system that can be built today, which addresses the many criticisms of what is happening today.