About a year ago, when our experience with our EMR approached the 5-year mark, I thought it was time we approached more advanced projects to automate our workflows. In a practice like ours, which includes surgery, some of the most repetitive, time-consuming, labor-intensive workflows are those that prepare patients for the operating room. Analysis of this workflow reveals there are 2 types of components: very simple and very complex. There is not much in the middle.
Complex steps, performed by an MD or nurse, include:
- Design of surgical procedure
- Location of procedure
- Choosing preoperative labs, imaging, consultations (i.e., cardiology)
- Interpreting results of the above testing
- Communication with the referring physician and preop consultants
Simple steps consist of generating the paperwork to reflect the medical decisions made in the complex steps:
- Chart notes documenting need for surgery and surgery design
- Initiate preop workflow with appropriate staff (i.e., surgery scheduling)
- Preoperative History and Physical
- Informed consent forms
- Preop and postop orders – implementation of above complex decisions
- Preop and postop patient instructions and other supporting information
With these thoughts in mind I began creating a set of body-site specific templates with the sole purpose of initiating and directing the above workflow steps from a single computer screen. We have been using 2 of these templates for about 6 months; I would describe the results as a good first attempt. This project is one of the few done by a physician that directly addresses workflow as well as documentation.
Three challenges make this process slow and laborious.
The first challenge is to automate the simple steps as fully as possible while leaving the complex steps in full control of the physician. This is much more difficult than it sounds. Too much automation will usurp control of clinical decision making from the physician; too little results in failure to fully leverage the EMR technology and may make the workflow more cumbersome than it was to start with.
The second challenge is making the templates user-friendly, which in this case means doctor-friendly. That means getting rid of all the screen clutter and “white noise”, keeping only what is useful and necessary. Many of the EMR’s built-in, “hard-wired” screens fail to do this well.
The third challenge is getting the code written correctly so the template interfaces properly with the EMR itself. This includes trivial but maddening issues such as preventing duplicate actions from duplicate button clicks. This also becomes frustrating when the template you are creating attempts to overcome shortcomings of the EMR itself.
Currently my conclusions are:
- The effort is worthwhile. Although the workflow engine is far from perfect and far from finished, it is quite useful “in the trenches” taking care of patients in a busy clinic. Every time I use the workflow engine I save $6.50, my estimate of what I would pay our staff to do the paperwork by hand.
- Some of the grunt work required to design and write a workflow engine is unavoidable. An EMR that is flexible enough to write useful custom workflow engines must by definition require a high level of effort to customize.
- Software improvements are required in order to make workflow engine creation practical for most practices. My IT experience is about as good as you will find among non-IT-professionals (35 years) but my skills are pushed to the limit making these templates. That is unacceptable. Notwithstanding conclusion #2, the situation can and must be improved.