Well, we went live on CPOE on May 15. This was not a big bang but rather a trickle, starting with the hospitalists who volunteered to act as provider champions and were willing to work through the bugs and growing pains with us. It is now three weeks later and we are still with the initial group plus two, and have had more than our share of pains. The following are some of our trials, tribulations, and lessons learned.
To start with, the go-live went fairly well, all things considered. Meditech had a good skilled support person on site and we were able to work through several little dictionary issues and process, process, process challenges. That having been stated, we had a very low “new” patient volume the first and second day, so most of the real problems and issues were not noticed until the third day when the Meditech representative was scheduled to leave. Since we had daily end-of-day huddles recapping the challenges and successes of the day, we quickly got consensus that we needed to ask the resource to stay an additional day. Meditech responded favorably to our request, and we were able to work through many of the growing pains in an immediate fashion, as opposed to having to place tickets, which can take days to even be looked at initially.
Another thing that became quickly apparent is that the way we had built orders in the nursing (PCS) and order management (ORM) dictionaries had impacts in the provider CPOE orders and dictionaries. Since Meditech does not train user sites on downstream dictionary impacts or interoperability, we had no idea of that. One example is that nursing orders include a need to take vitals. Our standing nursing orders include nursing vitals and pulse oximetry. If the provider enters an order for pulse oximetry again, not realizing it is part of the standing nursing order set, it then becomes, in effect, a duplicate order.
Another key challenge related to how lab orders in CPOE handled comments. This was due to how the LAB dictionaries are built and, more importantly, the process around how lab handles comments. I will provide two examples that resulted in different solutions. The first was with add-on orders. The provider would, for instance, order a CBC, then, as is often the case, decide he or she might want to edit that order once placed and have the lab also request a Basic Metabolic Test (BMT or BASMAT as it is known in our facility) or other blood test. To keep the phlebotomist from taking the patient’s blood twice, the provider used to write that this was an add-on test on the paper order. In the CPOE world, a comment field was built and used with the order — only the lab did not get to see the comments field until after the blood work was drawn and processed. In other words, the patient risked having to draw blood twice when the orders were basically placed at the same time. To deal with that issue, the workaround we created was to add an option to the provider order sets that is shown as an Add-on Test and is displayed with the first order option on each order listing.
Another example, again utilizing the need for a comments field in the lab testing, is when a provider orders something like a INR (International Normalized Ratio) and gets the results but does not trust them, so he/she orders the same test two days in a row. Lab may see this and either cancel it as a duplicate or try to contact the provider to ensure that he or she really wanted a second test. We now re-implemented the comments field for the provider to note that they do want the test rerun. The challenge is that the text length is very small and limited since it gets printed on a test tube-sized label. Obviously brainstorming solutions included some very creative ideas such as the provider typing NIKE (“just do it”) into the comment field.
Lab was not the only department impacted by cross-module order communication. Radiology, cardiology, and others also saw challenges requiring either dictionary or process changes. Ultimately, while many of the providers feel there are hidden efficiencies or benefits to CPOE over the paper orders, the one lesson we all wish Meditech would take away from our deployment is that cross-module dictionary education is critical, and so building in a siloed approach is counterproductive. Also, from my personal experience with other vendors like Siemens, Epic, and McKesson, onsite support from vendors during go-live really does require more than three days before transitioning to a call-in model.
But I regress, so back to CPOE lessons learned. One critical problem that we are still dealing with is the printing of discharge medications. This is a critical part of Meaningful Use and medication reconciliation, and is a patient safety issue. In theory, the medication list nursing compiles from home medications should be seen at the time of admission — the providers look at the list and decide what to continue or place on hold. That list then feeds the current meds listing. At time of discharge, the current meds should come over and merge with the original home meds list and the provider then select which ones the patient should continue. Basically, this usually looks like a clean version of the admissions list plus a couple of antibiotics.
The problem we are experiencing is when the discharge list is compiled, it may include other medications that the patient was never on, or so-called ghost meds. We found these are sometimes picked up via the pharmacy dictionary if a dosage does not exist. For instance, the provider may wish to select to continue cefpodoxime proxetil, an antibiotic, but instead on the discharge summary it may read Bactrim Septra, which the provider never selected. While both are antibiotics, they have different uses and strengths, and may not be interchanged in all cases.
Although this is a dangerous scenario and had a Meditech ticket still open on it, perhaps the more problematic issue is that once the discharge medication list is correct on the CPOE page and the provider has reviewed it 2-3 times, it does not print with the same medications that are on the screen. A provider has the potential of thinking the medications they noted on the discharge instruction page printed correctly and may sign that document to hand to the patient without ever realizing the danger or potential for problems. While Meditech is working on that issue for us, the only solution in the interim is provider education and having the providers read the list once printed, doing a one-to-one comparison against what they show on the discharge meds screen. Obviously, this detracts from the provider experience and trust in the system. It also adds to the time it takes a provider to discharge a patient.
In terms of time, please note that most providers will require more time with patients and performing documentation than they did in a paper world. That may affect your RVU or other productivity measures used on provider compensation contracts. This should be noted and addressed as part of the initial deployment strategy. Also, in terms of cost, you may have to pay your physician lead or champion an additional stipend, especially if they are not an employee of your hospital. Additionally, you may need to up-staff some shifts or days when you start in order to keep patient flow where it needs to be. Going big bang will also require much greater levels of support staff. Don’t forget all those pesky huddles and meetings requiring your CPOE team, providers, and ancillary units. Also, don’t be surprised if those teams grow in size and meetings grow in frequency. At the same time, you may also be having simultaneous meetings on additional or new order sets.
In terms of the product itself, provider feedback has been for the most part very positive. While some comment it is not Epic, they do admit to seeing benefits or efficiencies with use that were not readily apparent or noted during go-live. One provider maybe put it best, “Paper and MAGIC was like having a 1970 Ford F-150. It got you were you were going but not with any bells or whistles. The new 6.x product is like having bought a fully loaded 2012 Ford F-150 but is missing a wheel or constantly having something break on it.”
I will not bore you with additional postings on how the CPOE rollout went with other specialty providers like OB or the surgeons. Instead, my last Meditech 6.x posting will be an overall wrap-up article of the experience from start to finish, including where we are today (with over 120 open Meditech trouble tickets), 6.x and our ability to meet Meaningful Use, and where I think the future lies.