I have copied to my blog post #1,469 and #1,470 at the HIMSS Linked in discussion “Top Ten reasons when EMR/EHR Implementations are failing”.
EMR/EHR implementations don’t have to fail!
Knowledge workers don’t relate to structured protocols except when it is crystal clear that such protocols automate tedious, clerical activity, in which case they relate strongly to these. So, no wonder attempts to impose rigid “best practices” do not work in healthcare.
In my area of process automation I see 80/20 scenarios in manufacturing, 50/50 in some areas and 20/80, even 5/95 in some industry areas, healthcare probably being closer to 5/95.
The simple solution for healthcare is to back away from complex IT solutions and consider that what healthcare is trying to do is bring together a patient who has a problem(s) with providers who have skills and availability (both in short supply) and an infrastructure where the two parties can meet.
Some of the work is structured and should be. There are not ten “best ways” to prepare a claim within a healthcare organization – you have an encounter at a particular day/time, there is a code, time duration, and a progress note -, you reasonably want your software to post a billing line without requiring anything more than tagging the encounter with “Completed” versus “No Show”, or “Cancelled”. End of.
In the clinical area things are much less structured. This does not mean we have to throw up our hands and say we cannot have an EMR that provides benefits..
Hybrid software that has its foundation in Adaptive Case Management (ACM)/Business Process Management (BPM) is able to accommodate structured processes comprising linked steps (albeit with embedded branching decision points such that the flow is anything but purely linear) plus unstructured “processes” comprising one step each, thereby accommodating ad hoc interventions. There are no other types of interventions.
The way to accommodate ad hoc interventions is to make an inventory of all standard interventions the organization is capable of rendering and then add one extra step/form that posts a blank memo field from which any ‘order’ can be written.
The clinician either streams the patient onto a best practice workflow or picks items from the inventory or does both. If the patient is on a care pathway, the clinician can skip steps, perform steps out of order, revisit completed steps as well as engage collaborative consultations with peers/domain experts at steps.
“Patients At Portals” adds a new dimension – unlike unthreaded e-mail messages, PAP allows messages to emanate from process steps/inventory steps (points of service), with responses to messages returning to the point of service (in case the performing resource has now gone off shift, in which case you would see a 16-hour delay).
None of this contributes anything unless/until the software presents an easy-to-use interface to the healthcare professional,
How about ONE familiar screen that models what all of us do every day?
If you think about it, the way people work is they attend to their fixed time appointments and, separate from this, they attend to items on a to-do list. No exceptions.
Paper or electronic familiar “agendas” have, for years, accommodated attending to fixed time appointments and slotting in to-do’s in between fixed time appointments. It’s not particularly difficult to replicate this in a software system and if you include a resource allocation module that posts structured process steps and allows posting of unstructured processes, we are done.
Clearly at each step (ad hoc or workflow), healthcare workers need to:
* have any required instructions readily available at each step (not in some Help system that requires that you know the answer to your question in order to find the answer).
* have all forms/certificates needed to record observations/document interventions at each step so that they do not have to hunt for these.
And, when an intervention is completed, all data recorded at the intervention needs to automatically go to the EMR/EHR (date/time stamped, user signature), with anytime recall of data, as it was, at the time it was collected, on the form versions that were in service at that time.
The resource management model provides continuity with the same clinician or takes care of handoff to another (whatever is appropriate).
Using the ‘agenda’ model where the user is “in charge” (follows the best practice, does not follow, does what he/she likes) – there is very little left to object to.
And the icing on the cake is to let each Business Unit design/build and own their processes (with some assistance from IT in respect of rule set building).
This way the process becomes “the application” and there is no more of “.. you need to do things this way because the computer requires it”.