Computerized Provider – or Physician if you like – Order Entry (CPOE) is an older technology that has been in the spotlight for the better part of the past year thanks to the American Recovery and Reinvestment Act (ARRA) and key components of meaningful use. Because of the “stimulus†offered by ARRA many hospitals across the United States will be gearing up to implement CPOE, ready or not. Currently less than 20% of the hospitals in the United States are using CPOE, and only a small fraction of those are using it for all orders throughout their facility (AJHP. 2008; 65:2244-64).
Like many facilities, my hospital is in the process of gearing up for CPOE. We’re in the initial stages where committees are being formed, money is being spent, groups are gathering to discuss who is going to do what and IT, pharmacy and nursing are busy trying to figure out how much impact CPOE will have on their departments. Make no mistake, regardless of the impact, we’re moving forward.
The reasons for implementing CPOE are clear: it has the potential to eliminate illegible orders, eliminate the use of error prone abbreviations, create fewer phone calls from the pharmacy, decrease the number or incomplete orders written by providers, increase formulary compliance, improve provider compliance with hospital guidelines and with the right clinical decision support system can potentially decrease dosing errors, adverse drug events and potential patient harm. (Arch Intern Med 2003; 163: 1409-16 , JAMA 2001;285:2114-2120, JAMA 1998;280:1311-1316, Pediatrics 2009;123;1184-1190, Pediatrics 2010 May 3 [ePub]). However there is enough literature out there to at least question everything I just said (Pediatrics 2005;116:1506-1512, Pediatrics 2006;118:290-295, JAMA 2005;293:1197-1203), but overall I see the benefits of using such a system. There’s a decent summary of CPOE literature in Ann Intern Med. 2003;139:31-39. The information is a bit dated, but it’s a good place to start. Another good reference area is the CPOE page at AHRQ.
In addition to the benefits listed above CPOE has a tremendous upside for pharmacists. CPOE has been shown to potentially reduce pharmacy order-processing time, which in theory frees up pharmacists to concentrate on patient care (Am J Health-Syst Pharm 2009;66: 1394-1398). And we all know that pharmacists at the bedside can reduce medication errors and save money (Am J Health-Syst Pharm 1997;54:1591-1595, Pharmacotherapy 2008:28:285e–323e, JAMA 1999;282:267-270). This is one thing that BCMA hasn’t been able to do for us. We’ve seen benefits from bar coding in terms of inventory control and safety, but the workload in the pharmacy has increased since implementation. However, there are several things that can be done to alleviate the additional workload, including tech-check-tech and re-engineering our workflow, but at this time we haven’t put a lot of effort into improving the situation.
With all that said, CPOE implementation is a daunting task. It’s time consuming and it’s difficult. There is no question that CPOE implementation will require significantly more financial and labor resources than BCMA implementation did. The financial impact alone is enough to give many health care systems pause. I’ve seen numbers in excess of 10 million dollars for CPOE; in comparison we spent less than 1 million on our BCMA implementation, slightly more if you include the pharmacy automation. Still, it is obvious to me that CPOE is a worthwhile endeavor and cost of implementation shouldn’t be the only factor that deters your facility from moving forward. The potential to improve the quality of patient care, save some lives and get pharmacists out of the pharmacy and up on the nursing units is enough to convince me.
I’ve been fortunate in recent weeks to sit in on some of the CPOE decision making meetings, and some of the things that have been identified as necessary for our success include:
Ownership/Governance
The end users must control the design and implementation of our CPOE system. And the end users are primarily the physicians and secondarily the nurses and unit secretaries. Therefore the process must be driven by physicians. A healthcare system can build the best CPOE system in the world, but if the physicians hate it there is little hope for success. The need to engage the end user isn’t unique to CPOE, and should be applied to all technology projects. For example, the need to engage nursing early and often during BCMA implementation.
Following implementation physicians must take ownership of the CPOE system to continuously improve and optimize its look, feel and functionality. Allowing physicians to lose interest will result in a slow, agonizing CPOE death. Don’t think it can’t happen as there is precedent for uninstalling a CPOE system following implementation; <cough> Cedars-Sinai. To prevent physicians from becoming frustrated with the system or from feeling that their opinions don’t count, identify several physician champions to drive the process. Keep the mind set positive.
Understanding that physicians must take ownership of CPOE, hospital administration must provide a strong leadership infrastructure to deal with issues that are bound to come up.
The order set dilemma
Order sets are both loved and hated by pharmacists. On the positive side they hold the key to complete, legible and logical orders. On the negative side they can increase workload for the pharmacy and be just as incomplete, illegible and illogical as their pen and paper counterpart. Make sure to review all current order sets and physician templates prior to starting your CPOE build.
We started the process of critically evaluating every pre-printed order form that comes through the pharmacy. Let me just say it’s been interesting.
All orders on each order set must be complete from start to finish: drug, dose, route, frequency, indication in addition to criteria for use when multiple drugs are used for the same prn indication, i.e. mild pain, moderate pain, severe pain, etc. It’s important to remember that physicians are not experienced with order entry, so what makes sense to you as a pharmacist might not make sense to them. Oh, and make sure all your order sets have a similar look and feel. This gives the physicians some familiarity in which to work.
Building order sets and having them make sense is a real hassle so consider using something like ZynxOrder to help ease the pain. I recommended using ZynxOrder to manage our order sets, but unfortunately the system is expensive and our leadership decided to go in a different direction. The ZynxOrder system is designed to help build evidence based order sets supported by current literature. In addition it offers version control and an upload tool for moving order sets from the development environment into our pharmacy system, i.e. Siemens. Maybe next year.
Design protocols for use across multiple order sets. By this I mean develop standardized order set components for pain, bowel care, DVT prophylaxis, GI prophylaxis, etc. At the same time make sure to remove all range orders like morphine 1-10mg IV every 2-4 hours prn severe pain. The double range makes no sense. In fact the use of a range in frequency makes no sense.
Make order sets that are applicable across a discipline, not physician specific, i.e. there should be only one order set for routine postpartum care not a different order set for each OB-GYN physician. You’ll have a fight on your hands over this one, but stay strong because the end result is worth it.
Clinical decision support (CDS)
Invest heavily in clinical decision support (CDS). In order to take advantage of all that CPOE has to offer you absolutely have to use a robust clinical decision support system. Without CDS all you have is a glorified word processor. One word of caution, be careful not to expose the physicians to too many alerts. There’s a fine line between helpful and annoying. Most alerts in the pharmacy system are annoying. Physicians won’t put up with it. An article published in the Archives of Internal Medicine (Arch Intern Med 2003;163:2625-2631) showed that physicians overrode 91.2% of drug allergy and 89.4% of high-severity drug interaction alerts with no statistically significant change in ADEs. So make sure pop-up alerts and warnings are appropriate.
Support, training and hardware
Make sure your IT department is ready for a lot of hand holding. We did this with BCMA and it seemed to work well. And for the love of Pete, make sure your hardware is plentiful and up to date. One of the most annoying things in the world is logging into one of the computers in the hospital only to have it go up in flames when you’re in the middle of something. Murphy’s law tells us that this will always happen when you’re in the middle of the biggest order of the day and you don’t have a moment to spare. This goes for wireless connectivity as well. Moving into a patients room with a COW only to find that you have no Wi-Fi can really put a wrinkle in your day.
Conclusion
That’s it. I’m knee deep in CPOE stuff at the moment and don’t think I’ll be coming up for air anytime soon. At least things will be interesting for a while.