CPOE Implementation: A Status Report
CPOE is supposed to prevent medication errors, but it still has issues of its own.
Back in 1999, the Institute of Medicine (IOM) published the article "To Err is Human: Building a Safer Health System," which focused on preventing adverse drug events (ADEs).
Computerized Physician Order Entry (CPOE) was touted as a tool to reduce ADEs. Subsequent studies pointed out how it would help prevent medication errors and improve patient safety.
The US government has pushed computerization, as well.
"To improve the quality of our health care while lowering its cost," President Barack Obama said back in January 2009, "we will make the immediate investments necessary to ensure that, within 5 years, all of America's medical records are computerized."
It has now been 6 years, and medical records are still not 100% computerized.
Implementation of CPOE has been slow due to its complexity and huge cost. To further entice hospitals to jump on board with electronic health records (EHR),
the US Centers for Medicare and Medicaid Services
(CMS) sends money to facilities that meet set goals.
EHR systems are not something that can be rushed, but for dollars, workarounds happen. There is also the threat of penalties if systems are not implemented.
As the EHR market has matured, the once-crowded field of vendors has narrowed significantly.
At the end of 2013, just 10 vendors accounted for about 90% of the hospital EHR market: Epic, MEDITECH, CPSI, Cerner, McKesson, Healthland, Siemens, Healthcare Management Systems, Allscripts and NextGen Healthcare, according to Becker's Hospital Review.
No CPOE standardization
CPOE systems are all different, so how are they compared? A hospital may have implemented a CPOE, but does that equate to a sufficient system? Do groups like Leapfrog take into account CPOE errors or just the percentage of usage by prescribers? Do we rate CPOE systems like we rate hospitals?
Data show vendor CPOE market share, but there are no rating systems to evaluate the systems after implementation, or even a list of hospitals that decided to change systems due to issues.
Limited medication profiles
Another issue with CPOE is its lack of a coherent view of a patient’s profile while entering medications. It is also difficult to verify orders without a comprehensive view of the medications that the patient is taking.
This lack of a full picture causes the user, whether prescriber or verifier, to rely on the software alone, rather than a comprehensive approach. Seeing the whole picture while entering and verifying orders would probably decrease errors.
When CPOE systems are used for other tasks aside from entering and verifying orders, there is more alert fatigue.
On the pharmacist verification end, it is common to see alerts of different significance with nothing to differentiate high importance from low. For example, the same type of alert may be used to discuss inventory, prior authorization, and other messages that take away from the verification role, even though many of these alerts previously happened at order entry.
Pharmacists should not have to think pharmacologically and pharmacokinetically about how a medication works along with alerts dealing with inventory, cost, and formulary status that once occurred at the front end. There should be a way to differentiate these alerts and have them fire at appropriate times, rather than during actual medication review.
Tailoring the CPOE to be more user-friendly for the prescriber often comes at the expense of more frustration on the back-end with verification. For example, a CPOE may allow a prescriber to free type directions for medications taken irregularly (3 days a week, different strengths on different days), choose non-formulary medications rather than built-in CPOE formularies, and remove alerts that need to be seen at order entry.
In this way, verification becomes more of an order entry “fix” role that pulls attention from clinical aspects of verification.
CPOE software is also designed under the assumption that prescribers and verifiers are working in a quiet environment, but both sides are working in noisy environments. When a phone is ringing, a patient is yelling, and a nurse is asking a question, quick pop-up alerts may not be enough of a warning. Even the most focused individual will make mistakes.
More duplicate orders
The Journal of the American Medical Informatics Association published a study pre and post-implementation of a CPOE in an ICU and found that duplicate medication ordering errors increased after implementation (pre: 48 errors, 2.6% total; post: 167 errors, 8.1% total; p<0.0001).
Sometimes, there is a lack of integration between laboratory values, both inpatient and home medications, and other data or different modules that do not communicate smoothly.
Last but not least, if the staff is not happy with the CPOE software that is implemented, they are not going to use it as designed.