There is a joke told by medical students as they begin their surgical rotation. In the joke, a medical student asks the surgeon whether he/she “likes their stitches too short or too long.” When it comes to order sets, a CIO may feel like that student. There is no one-size-fits-all solution to either the number or the content of order sets, only the certainty of not being able to please everyone. In this commentary, I look at some the predictable pitfalls and discuss strategies for getting it right.
The clinical case
Recent publications suggest that we have not yet met the promise of improved care with our EMR. Much of the fault is our incomplete adoption of clinical decision support (CDS). The order set represents one of the most fundamental forms of CDS and is a cornerstone of the build and implementation of CPOE. Order sets can translate evidence-based clinical knowledge into actions at the point of care in support of error reduction, quality improvement, and reduced clinical care variability.
In both paper and electronic practice, “personal order sets” may predominate. Most sets are the proprietary creation of a single practitioner, and exist outside of institutional control or review. In many cases, they contain non-formulary medications and orders inconsistent with hospital protocols, and are all too often in conflict with established consensus guidelines. Although touted as contributing to clinician acceptance of CPOE, they systemize practice variability and are difficult to centrally administer. Wholesale conversion of personal paper order sets to electronic ones is rarely a good idea, so don't do it!
It is helpful to think of the electronic order set as the fundamental disease- or procedure-specific platform for evidence-based orders within CPOE systems. Put simply, by design you must make it easier for clinicians to do the “right thing,” because accepting the set (or sections of it) is so much faster than entering single orders. The rub, however, is in gaining consensus on what constitutes the “right thing.”
For CIOs, the creation of order sets is one of the first battlegrounds to be faced in the march towards CPOE, and often pits the physicians against the IT department and physicians against each other. Successful development strategies start with good governance designed to solicit clinical guidance from physicians, nurses, ancillary services and pharmacists, as well as all individuals responsible for entering and receiving orders. Clinician thought leaders with expertise in clinical domains must assume responsibility for creating, editing or approving order sets that are used commonly in their area of expertise, and they should be responsible for gaining consensus from their cohorts. Physicians sometimes need to be reminded that order sets are not cookbook medicine; an order set does not replace clinical judgment and can contain optional therapies or treatment paths. A physician never abrogates the right to add or delete orders. Working together with IT knowledge engineers, orders should be configured to conform to a hierarchy of rules, standards and principles that makes sense both clinically and design-wise.
When order sets are implemented without rules or governance, chaos can result. I recently consulted for a hospital with over 1,500 personal order sets containing 270 different sliding scales for insulin. There was so much variation in orders for pain medication that 27 percent of patients received different doses of narcotic for the same pain score; and a multiplicity and complexity of potassium replacement orders led to the correct dose being administered only 56 percent of the time. Excessive variability in ordering introduces an unacceptable risk of medication error.
To build or to buy, that is the question
Most hospitals spend vast amounts of time developing order sets. This is a painstaking process that requires the participation and cooperation of disparate stakeholders with widely varying motives and interests. Even under the best of circumstances, organizations can spend six to eight months developing a starter set. Most calculations of the cost to implement order sets fail to consider the expenditure of intellectual and creative capital by clinicians and IT staff to research, design, create, encode and eventually maintain order sets. It is in this environment that vendors like Zynx have strategically placed their pre-configured, frequently updated, evidence-based order sets for sale. While there can be different content, support, cost and integration capability between vendors, they all fill a similar niche. Depending on integration hurdles with your particular EMR software, there can be considerable savings of both time and effort for your clinical and technical teams. ROI calculation is unique to each institution and culture.
The ability to access the intellectual capital of your clinical staff and the ease with which you can reach consensus are the make it or break it clinical decision points when you consider home-grown versus purchased order sets. Order sets authored, maintained and disseminated by specialty groups are already available for many diseases and procedures and can provide the content foundation for a home-grown effort. Unfortunately, because of the present technical lack of standardization, these are not plug-and-play. In our present environment, there can even be difficulty in sharing order sets among institutions using the same software.
Here are a few tips:
Before you start, have in place a formal hospital policy and procedure for the creation, adoption, maintenance and review of order sets. Have defined governance involving administration, clinicians, pharmacy and IT. Don't forget a policy for conflict resolution.
Design your order sets with maintenance in mind. If a drug goes off formulary, you will want to easily touch all order sets in which that drug resides.
Have a formal “style guide” setting forth the principles and standards from which all sets will be designed. Adopt naming conventions, a single lexicon, and archive policies before starting the build.
Design order sets to minimize clicks and scrolling. Have a uniform policy about the use of pre-selected orders. Remember, this is what your physicians will interface with on the first day of CPOE.
Fight off the temptation to allow personal order sets. Just say no. If you already have them, quickly develop a policy for retiring them.
Don't underestimate the task of getting physician cooperation and consensus. It will take longer than you think, so start early in your CPOE implementation. A pre-existing spirit of collaboration or hostility among your constituencies is the best predictor of outcome. Don't burn out your physician thought leaders and champions. You will need them later on.
Don't obsess about the number of order sets. Depending upon the profile of the services you provide and procedures you perform the number varies widely. Start with your highest volume and highest priority conditions and procedures, and build from there. If you have less than 300 or more than 1,000 then it is time to rethink.
The order set process is just the first of the challenges inherent in CPOE. In many ways it sets the tone for the relationship between clinicians and IT. Start early, test the waters, choose your battles, make strategic alliances, and remember, this is the first leg of a marathon, not a sprint.