Keeping three websites going is a lot of work, but entirely manageable with proper planning. Clinical Swift and Clinical Workflow Center were added to the EHR Science family to address specific HIT issues and challenges. All three sites had detailed plans on launch, and plans for each were revised in Autumn of 2015. However, things changed radically (for the better) over the ensuing eight months, so that by the end of July, I needed to stop everything to determine what should happen next.
Like EHR Science, Clinical Swift and Clinical Workflow Center were conceived as primarily educational/informational sites — places where I could opine to my heart’s content about informatics and technology. And, while this aspect still applies, the amount of effort and time spent on understanding new technology and software engineering has managed to change how I view what it is that I should be doing every day.
From theory to practice
In 2011, when I launched EHR Science the idea of writing commercial software was very far in the back of my mind. At that time, mobile platforms, the cloud, and NoSQL were relatively new. The BPMN 2.0 spec was only a few months old, and business process management systems (BPMS) that fully (or near to it) supported executable BPMN 2.0 systems were in the future. Five years ago, my interests leaned decidedly toward the theoretical and conceptual basis of clinical systems, so I buried my head in discrete math and software engineering books for a few years. I wanted to understand software from the ground up and to develop better skills at deconstructing real-world systems. Most of my explorations made their way into blog posts, and frequent readers will note that those explorations radically changed my views about clinical software design (see Is the Electronic Health Record Defunct?).
Concurrent with my studies, I tried tinkering with various technologies — PHP frameworks, NoSQL databases, non-Microsoft development tools and, in July of 2015, Swift and Xcode. Likewise, my explorations in clinical workflow analysis and interactions with Chuck Webster steered me toward BPMS (by way of Petri nets and graph theory). Since December 2015, I have investigated four open source BPM suites that are sufficiently mature to be used to deliver workflow-based applications to small businesses.
Demand for improved EHR workflow support is growing, but it will take a while for deep workflow support to appear in current EHR systems. And when it does, those EHR workflows will be hardcoded, time-consuming to change, and one-size-fits-all. BPM suites (i.e., workflow technology) address these issues head on. Small practices can get the benefit of workflow technology without having an EHR—even if they still use paper records. Five years after setting out to learn as much as possible about software design, the tools have appeared that allow me to turn my studies into applications.
CPC+ is more revolutionary than it seems at first glance. Current EHR systems do not have the process-support capabilities that CPC+ requires, and adding them to current EHR systems will prove to be much more challenging than MU—much more. CPC+ broaches the subject of process support (I’m not sure that program designers recognize this); no matter, the cat is out of the bag.
Apple introduced CareKit in April—a complete surprise to me. Having looked at the framework, I can say without any reservations that CareKit is going to be transformational. Intentionally or not, Apple has created a little clinical software framework. The time required to build certain types of clinical apps is greatly reduced with CareKit. Once one moves past the notion that a clinical care system is necessarily a monolithic software application (with a patient data repository at the gooey center) to thinking in terms of a set of limited applications that have access to a patient data repository, the landscape looks quite different.
The infrastructure that Apple is putting into place to support CareKit, ResearchKit, and HealthKIT will help to create a market for problem-focused apps. Practices may adopt them as desired, and, once again, paper charts will not be an impediment. The cost in time and money to try clinical software in practices will be much lower with problem-focused apps. Practices will be able to experiment without worrying about developing a bad case of buyer’s remorse. CareKit and provides tools for developing apps that will allow practices to experiment with patient engagement and care coordination in ways that are simply not possible now.
The change in view and circumstances brought about by the actions of Apple, CMS, and BPM suite vendors over the last year has forced me to reconsider what I was doing. Accordingly, I stopped everything and made far more detailed plans for all three sites, and for Informatics Squared, Inc. I am polishing up my first iOS app (something for my wife), and I’ve caught the bug. It’s a new day!
Publish or what???
For the past two years, I have been wrestling with the problem of how to address the theory-application divide from the standpoint of publications. More than a few times I have scrapped paper outlines because I could not get the theory-practice content problem resolved. No one hoping to solve a problem wants to wade through a dry theory discussion. Conversely, no one interested in HIT design from a more abstract/mathematical perspective wants to wade through a how-to guide. Not being able to find the right balance has put the kibosh on more efforts than I care to remember. Finally, I have come to terms with this matter. In no small part, this is due to greater interest in HIT design papers from journal editors, book publishers, and conference committees. Thanks to MU and frustrations with usability, there is growing realization that fixing HIT will require more than a few tweaks and some user feedback. In response, I used the time away to outline a few topics for white papers, at least one academic paper, and a how-to clinical workflow handbook.
I spend most of my time on a MacBook Pro. Until a few years ago, I was a Windows guy. The switch to open source and Apple development required a more robust setup. The Mac laptop got an upgrade to 16 GB RAM, and now I’m running Ubuntu Linux and Windows in virtual machines—quite smoothly, I might add.
The cloud has also made an appearance. Cloud providers are making it dirt simple to experiment with public-facing applications, so I am setting up a cloud-based BPM suite test site. What would have cost thousands of dollars to set up 10 years ago, thanks to competition among Amazon, Google, and Microsoft, is now very inexpensive (or even free, for a limited time) for those creating/testing applications.
Behind the curtain…
Life has not been all rainbows and butterflies. All websites have been upgraded to servers that can handle traffic surges better. Each now resides in its own virtual machine, which makes management less problematic. SSL has been added to each site, and security improved (I hope). There haven’t been any breaches, but looking at firewall information has been jarring. Of course, there have been the requisite plug-in update failures that led to hours of emailing with hosting company tech support, reviewing .htaccess files, and learning much more about file permission settings than I care to know. At one point I was completely locked out of all sites by a tech support engineer with too many opinions and too little knowledge. Happily, all problems have been resolved.
Overall, I’ve had a very productive summer and look forward to the possibilities that lay ahead. Until next time…