This ISMP Alert was perfectly timed because we had confusion over this labeling just a couple of weeks ago. I grabbed one of each label type out of the carousel and snapped a couple of pictures. See below. The top image is of the original labeling, the middle image is the new labeling and the bottom image is the two sitting side by side for comparison.

Liquid unit doses should really be packaged in an oral syringe. Oral syringes are clearly marked to indicate volume, which helps avoid confusion like that caused by unit dosed cup. Another thing I would like to see changed is the use of concentrations like 125mg/5mL and 100mg/4mL. Even though these concentrations are clearly the same, you wouldn’t believe how often this confuses people. Labeling should contain the concentrations in its lowest possible volume, i.e. 25mg/mL, and the dose should be clearly marked, i.e. dose = 100mg = 4mL.

The problem is more widespread than most realize and often flies under the radar. I knew about the issues, but wasn’t prepared to deal with them until we went live with our bar-code medication administration (BCMA) system. A side effect of the BCMA system is that it tends to catch things like late and/or early medication administration. That means no more mythical med passes with all medications administered at exactly the same time.

One thing to remember here is that the problem does not reside with the nurses, per se. There are many factors involved. Nurses are frequently asked to do too much with too little time and resources, thus forcing them into undesirable situations. The system is the problem. And as much as it pains me to say, this is one problem where a technology-only solution is not the answer.

The solutions are simple, but not always obvious or practical for many health care facilities. For example, the 30-minute rule could be changed to a 60-minute rule, i.e. medications would need to be administered within 60 minutes before or after their schedule administration time. This was recently done by the American Association for Respiratory Care (AARC) in a position statement that basically said that inhaled medications shouldn’t be held to the same CMS 30-minute rule because “Inhaled medication administration incorporates a unique methodology and has a recognized delivery time between 9-20 minutes, depending on the delivery device used for administration.” The AARC statement is supported by CMS. Or perhaps the facility could stagger standardized frequencies to give nurses additional time to admister medications, i.e. not have all morning medications due at 9:00 am. It may be as simple as moving some medications like aspirin, warfarin or HMG-CoA reductase inhibitors, i.e. atorastatin  and the like to the evening time. How about this one: hire more nurses. I know, easier said than done, but very practical nonetheless. It’s important to remember that this is first and foremost about safe and effective patient care.

Ultimately there isn’t a one size fits all approach to the problem and it is clear that it will be some time before we have a solution, but it is certainly something that needs to be addressed. Unfortunately this isn’t a problem that immediately available technology can fix. ADCs are not designed to be a time saver for nursing. Neither are Bar-code Point of Care (BPOC), a.k.a. BCMA, systems or Computerized Provider Order Entry (CPOE). All these technologies are designed with the idea of improving patient safety through the reduction of hospital related medication errors. We’re going to have to look somewhere else for a solution. Just a thought.

A little bit of information on our labeling process may be in order. To generate a bar-code label with our software we have to enter the particulars of the drug into the application database. One piece of information, perhaps the most important piece, is a unique identifier. We’ve settled on the NDC number, which is often contained in the bar-code on the manufacturers product. With me so far? Good. So to make things easy we typically scan the bar-code on the medication package, which in turn automatically populates the unique identifier field. We do this to mimic the bar-code on the manufactured item as closely as possible; works great most of the time.

Now the problem. Take a look at the image below making sure to pay particular attention to the numbers beneath the bar-codes. Notice the similarities. The first 11 digits, highlighted in yellow, are the same for the drugs on the top and the bottom. The same is true for the first 11 digits on the second and third drugs, highlighted in green. Unfortunately these are completely different drugs. The second item is a 1mL vial of midazolam 5mg/mL injection and the third item is a 2mL ampule of fentanyl 50mcg/mL injection. No problem because the last five digits of the number in the bar-code are different, right? Sort of. Our labeling system truncates the information at 11 digits. So when the pharmacist attached these drugs to our cross-reference file the BCMA system couldn’t tell the difference. Doh! The solution was simple, but only after the mistake was caught. I won’t tell you how we caught the error, just know that it was caught.

The system broke down in several places and no blame is necessary. However there are some important lessons to take away from the experience: pay close attention to what you’re doing, be careful, check the product after you label it and before it goes out, and know the limitations of your technology.

Update (6/28/10):
Some people have asked me how we handled the bar-coding issue above. The issue is actually being handled from two different directions. The solution on my end was really quite simple. For items like those mentioned in the post I remove the first five digits from the unique identifier after scanning the bar-code information into our AutoLabel system. It just takes a second and eliminates the duplicates. See, I told you it was simple.

The second piece is coming directly from Talyst. Someone at Talyst read my blog post and shot me an email regarding the issue. I gave them a little more detail and they put me in contact with a couple of their engineers responsible for the labeling system. They’re currently working on a more technical solution.

I’ve been a pharmacist for more than 10 years now and I’ve make my fair share of mistakes. I would like to think that none of those errors caused harm, but that would be naïve to say the least. And forget about the errors that were never detected because one can only speculate about those.

During my early years of staffing I entered anywhere from 300-400 orders per shift; more with overtime, extremely busy days, etc. Assuming that I was 99.9% accurate, which is a ridiculous assumption, let’s do a little math.

- 300 orders per shift x 4 shifts/wk x 50 weeks/year (hey, everyone gets a vacation) = 60,000 orders entered annually

- With 99.9% accuracy, don’t laugh, that means I committed approximately 60 errors per year. Now I know that accuracy rate is a crazy estimate. I’d believe 60 errors per week, but 60 errors per year, c’mon.

- Anyway, assume 60 errors annually for 10 years = 600 medication errors that reached a patient

- Based on numbers from various sources I’m going to assume that a majority of these errors were harmless. Using similar sources I’m also going to assume that approximately 3% of my errors inflicted damage in one form or another. That means I managed to harm approximately 18 patients during a ten year span in my career. How many of those 18 patients were seriously harmed or even killed secondary to one of those mistakes? We’ll never know.

For those pharmacists that think they don’t make mistakes, think again. There is zero chance that you won’t harm a patient during your career. If you’re fortunate you’ll never know. If you’re unfortunate you have the potential to be fired, lose your license, or end up in prison like Eric Cropp. So much for just culture.

We don’t mean to make mistakes. No one rolls out of bed and says “hey, today it’s my turn to make an error”. By definition an error is unintentional (from Merriam-Webster an error is “an act involving an unintentional deviation from … accuracy“) and everyone makes them. So how do we eliminate them? We can’t because they’re unintentional or weren’t you listening. As long as humans are involved in the process there will be mistakes. And I don’t believe that you can remove humans from the practice of pharmacy because in certain circumstances you need insight and experience that no amount of automation can emulate, yet.

We can, however find ways to reduce errors. I firmly believe that technologies like CPOE, BCMA, CDS, robotics, or some yet to be developed system will prove beneficial in reducing medication related errors, but they will never completely eliminate them. I also believe that diligence and exploration into ways to incorporate new dispensing techniques, better practice models, evidence based practice and workflow modeling can help reduce pharmacy related errors as well. It’s silly to think that technology is the only solution.

There you have it; something I neither wished for nor desire to keep, but remains nonetheless. Just a thought.

Most consumers don’t do a very good job of keeping track of their medications; much less the medication names, dosages, what they are used for and when they were last taken. It’s not uncommon on admission to the hospital for a patient to say things like “I take a blood pressure pill” or “a pain pill” or “a water pill”. As a pharmacist I can make gross generalizations about these medications, and can narrow the options down with aggressive questioning, but can rarely be sure without seeing the medication for myself.

The Electronic Health Record Incentive Program, a.k.a. Meaningful Use guidelines, calls for medication reconciliation to be used for at least 80 percent of “relevant encounters and transitions of care” (page 95). In addition “the capability to perform medication reconciliation is included in the certification standards for certified EHR technology.” This is easier said than done.

Most medication reconciliation begins in the Emergency Department. It is typically a manual system of information collected by nurses who in turn pass it off to the physician for approval. Unfortunately many physicians don’t take the time to scrutinize the medication list which is often inaccurate or incomplete.

The ideal list of medications currently being taken by a patient wouldn’t be generated by the patient at all. Instead the list would be downloaded from a nationally standardized e-pharmacy. Of course no such thing exists, but that doesn’t mean it shouldn’t.

In theory all medications taken by patients are filled in a pharmacy, whether that is a chain pharmacy, community pharmacy or mail order pharmacy. Modern pharmacies are computerized and connected to the internet so that insurance adjudication can take place. The same data should be transmitted to a centralized e-pharmacy where it would be stored and accessed by hospitals during patient admissions. The list would follow the patient throughout their admission and be finalized on discharge. After all, the medication use profile is never more accurate than at the time of discharge.

In the absence of a centralized e-pharmacy, several vendors offer software applications designed to help hospitals maintain a digital medication reconciliation record. Most of these applications can be integrated into the pharmacy information system, making the process a little easier. The solution is not ideal, but it is better than a manual system with pen and paper.

Some vendors that offer medication reconciliation software are listed below.

RxReconcile
MedsTracker
RcopiaAC
MediRec
Mediware’s ClosedLoop Clinical Systems
HCS Medication Reconciliation

According to Warner “On one hospital visit as Mom was being given her morning round of medications the nurse wheeled in the computer terminal with the meds. She scanned my mother’s wrist band, then preceded to scan my mother’s meds. She noticed for one prescription that she only had one pill and not the two that were ordered. No problem, she exclaimed. She scanned the one pill twice and said she would be back with the second. Mom downed the meds. 40 minutes later, I went looking for the nurse and reminded her about the second pill. At first she seemed confused. Then she said, “good eye” I would have forgotten. Intriguingly, the computer wouldn’t have. It thought all was well and good.” So much for BCMA. I wish there was a way to remove the “human factor” from many of our technological advances.

In another example of problems experienced by Warner and his mother, he chronicles a near overdoes of chemotherapy as their oncologist and nurse practitioner failed to adjust her dosage based on renal impairment. “This is where the example comes in. Mom started chemo on New Year’s Eve. As the nurse practitioner was preparing the dose she said that this could cause a reaction in her kidneys. My sister put the brakes on at that moment and reminded the NP that Mom was working with a kidney specialist. This brought a halt to everything. They went back to Mom’s file, found the pertinent info, were surprised to see it, called the oncologist, re-issued the dose of chemo, and finally all moved on.”

I don’t think most people realize how common the above examples are in healthcare. People who believe this type of thing is rare are naïve. As an industry, healthcare not only has a long way to go in catching up with technology, but a long way to go in taking care of patients as well. Perhaps the Obama administration should focus on that.

I always pay close attention to medication use with the members of my family. I question physicians and nurses all the time and have forced the issue on several occasions. This has allowed me to catch numerous errors and blunders over the years. What do families that don’t have a healthcare professional in them do?

I wish Warner and his mother all the best and hope that everything goes smoothly from here on out.

At our facility, allergy information is collected by nursing staff at the time of admission. While competent and intelligent, nurses are not generally in tune with the intricacies of taking a good drug allergy history; they have other pressing matters to deal with. As a result the accuracy of drug allergy information located in the patient’s medical record is often less than complete.

Pharmacists on the other hand are interested in everything related to medications, which includes allergies. Most patients don’t know whether or not they’re allergic to a medication, much less what the details are surrounding the allergy. Believe it or not, not all medication allergies are, in fact, allergies. Some perceived allergic reactions can be classified as an adverse reaction, a drug intolerance or simply an expected side effect.

If you get stomach upset from aspirin that’s not an allergy. How about if I fall asleep after taking a Valium? Definitely not. Well, what if I feel dizzy with my new beta-blocker dose? Nope. I get constipated when I take my morphine. Take a stool softener. What if I turn red and can’t breathe after taking penicillin? Yep, you’ve got an allergy.

Did you know that food allergies can affect how you tolerate medications? No. Well they can. For example, if you’re allergic to eggs you may not be able to receive certain vaccines and immunizations, and if you have a severe peanut allergy you should be careful with certain inhalers used to treat asthma and COPD. While not as glamorous as a good medication allergy, this is important information.

The trick with any good allergy history is to dig deep into the matter. Pharmacists, and other healthcare professionals, need to know what medication was taken, when it happened, what the reaction was, how long after taking the medication, how long the reaction lasted, whether or not the patient has taken any related medications and how they were tolerated, etc. It is imperative that pharmacists have this information if they are to make accurate decisions regarding drug therapy. Technology alone can’t do this.

During my fourth year as a pharmacy student at UCSF I was assigned to a general medicine team as part of my clinical rotations. The team consisted of an attending physician, a chief resident, three junior residents, two medical students and one pharmacy student; me. Every so many days this particular team would be responsible for the patients admitted to the hospital through the emergency department. Part of my job was to obtain a detailed medication history from every patient admitted to our service. The medication history included detailed information on the patient’s allergies. You wouldn’t believe the stuff I found by questioning patients. Of course, after taking the history I would enter all the information into the hospital information system for the rest of the medical team.

This brings me to my point; no matter what advances are made in technology, a change in practice in necessary to bring about a safer way to collect allergy information from patients. The best way to accomplish this is to move the pharmacist away from the physical pharmacy and into the trenches where they can interact with the patients. This is especially true in high-risk areas like the emergency department where patient information is first collected. I realize that computer technology is great, but at this time it cannot replace a good pharmacist. However, it can help a pharmacist do his job better and more efficiently. Just a thought.

There are two medications contained in the hand written orders below. I double-dog dare you to find them. It’s kind of like a “Where’s Waldo” puzzle. Leave your guesses in the comment section of this post. Good luck.

.

Agrawal A. Medication errors: prevention using information technology systems. British Journal of Clinical Pharmacology. 2009 ;67(6):681-686.
The article covers all the usual suspects when it comes to reducing medication errors via technology. Technologies discussed include CPOE, barcoding, BCMA, medication reconciliation, personal health records, automated dispensing cabinets and decision support systems. No great amount of detail was presented, but the article is well referenced. The entire abstract can be found here.

Cheung K, Marcel L. Bouvy, Peter A. G. M. De Smet. Medication errors: the importance of safe dispensing. British Journal of Clinical Pharmacology. 2009 ;67(6):676-680.
The article discusses several strategies for reducing dispensing errors, including barcoding and automated carousels. For each strategy presented, the authors provide some level of support found in the literature. The article is worth adding to your collection. The entire abstract can be found here.

One final item worth mentioning is a brief editorial written by J K Aronson, the President of the British Pharmacological Society. In it he states “Computerized systems can contribute to prevention as well as detection, but they are expensive and can generate their own forms of error. Simpler and cheaper methods are available and should be widely implemented. For example, error reporting is important in both detection and prevention, and pharmacovigilance has a role to play. However, chief among the preventive methods is education.” – I find this statement both insightful and accurate.