Medication reconciliation in hospitals: an interview with Dr. Leonard Feldman

Interview conducted by April Cashin-Garbutt, BA Hons (Cantab) on 3^rd August 2012

Please could tell us what adverse drug events (ADEs) are and why they occur?

Adverse drug effects are harms caused by the use of a medication. There are many different types of ADEs, including medication errors. For example, you could have an ADE from a medication causing a complication due to idiosyncratic reactions; or you could have an ADE from mistakenly giving medications that together are toxic.

There are all sorts of complications due to medications. Some complications are due to the known side effects of a particular medication.

It is one of the big concerns that physicians should have: that the medications given can sometimes cause more harm than good. It is really important for physicians to think about the medications they are going to give their patients to think about the ADEs that could occur with certain medications.

For instance, it can be potentially lethal if a patient whose kidneys don’t work very well is put on a medication called spironolactone. Even though this medication can help patients with heart failure, it can also cause very elevated potassium levels that can cause fatal cardiac arrhythmias.

A physician needs to think about a patient’s physiology and comorbidities before prescribing a medication. Not every medication is the right medication for every patient with the same condition.

Others complications can be due to the medication being given incorrectly, omitted, prescribed in incorrect dosages or in combination with incompatible medications. These are all examples of medication errors. These types of errors should be preventable.

Your recent study involved a nurse-pharmacist led medication reconciliation. Please could you explain exactly how this worked?

We had two registered nurses (RNs) who were not in any way experts in medication reconciliation. They were charged with trying to figure out what the actual home medication list was for the patients admitted to a Johns Hopkins hospital during this trial.

The nurses would try to see the patient if they were going to stay in the hospital for over 24 hours. They attempted to come up with the gold-standard list for the patient’s home medication through a variety of means: talking with the patient, talking with the patient’s family, calling the pharmacies, calling the primary care doctor, and looking over computer records.

Coming up with that gold-standard list isn’t easy. There isn’t really a gold-standard list when it comes to home medication lists. To get this list, you would have to sit at home with the patient for a week or so and watch exactly what they’re taking. Only then would you know what their real list is.

The nurses did their best to come up with this list. At the same time, the physicians admitted the patient and developed a treatment plan for the patient. The nurse would compare the medications the patient had been placed on when they came into the hospital with the home medication list to see if there were discrepancies between the home medication list that the nurse developed and the current medications.

There were going to be some discrepancies between the home medication list and what the patient was taking in the hospital that were intentional. I gave the example of the patient who has a high potassium level from spironolactone when they’re being treated for heart failure. Suppose the patient came in with a high potassium level, you wouldn’t expect the physician to continue to prescribe them spironolactone. That would be an intentional discrepancy between the home medication list and the list of medications the patient was now prescribed in the hospital.

There are also unintentional discrepancies. If the nurse was not sure whether a discrepancy was intentional or not, typically they asked the pharmacist for advice. If the pharmacist thought it didn’t look intentional, or if the nurse themselves noted it was not an intentional discrepancy, then the nurse would talk with the physician to ask them whether the discrepancy was intentional.

If the physician meant to make the change then it would be counted as an intentional discrepancy; but if the physician didn’t mean to create the discrepancy, then it would be counted as un-intentional. A group of pharmacists and physicians then rated the unintentional discrepancies to assess the likelihood that it would cause harm.

We looked again at discrepancies when the patients were going to be discharged. Here we compared the discharge list of medications to their home medication list. We also ranked these discrepancies at the time of discharge.

Your research showed that such a medication reconciliation process prevented some ADEs. Please could you explain how it did this?

We believe that we potentially prevented some ADEs. We did not have a control group, which would have received usual care. Since it looked like this process worked, we plan to randomise patients to usual care or a medication reconciliation collaboration that will be similar to this research.

We didn’t have a control group to see if they were actually having the potential ADEs that we thought. If there was a potential ADE, we stepped in so that it didn’t happen. This is why we ranked the ADEs. This allowed us to assess whether our intervention prevented harm and how much harm we may have prevented.

It was also reported that you said medication reconciliation could prevent hospital readmissions: what is the reason for this?

This is part of our next study – to see whether this is true. There is a lot of data that shows that patients get re-admitted due to ADEs. This occurs particularly during the first few weeks after they are discharged from hospital.

Our hope is that if we are able to decrease ADEs by getting rid of some of these discrepancies. By decreasing ADEs, we hope to decrease the need for readmission. More importantly, patients won’t have the medical complications from these ADEs.

We don’t have any direct evidence from this study that we will be able to prevent readmissions, but I think it is a reasonable conclusion that if we did prevent many of these rank 2 and rank 3 ADEs (as we ranked them), then patients would be better off and most likely be readmitted less.

You described the medication reconciliation as cost-effective. How was this the case?

Some of this is theoretical as we were only potentially preventing ADEs. We were using data from Bates et al. to try to estimate how much it would cost the system to have an ADE. Based on their data, it appeared that each ADE would cost the hospital approximately $9,344.

We then were able to estimate how much our intervention would cost. Based on this, we were able to determine that if we prevented 1 potentially harmful ADE for every 290 patients then we would offset the intervention cost.

From our data it wasn’t just 1 discrepancy in every 290 encounters that we thought we would be able to intervene upon. It looked that we would be able to intervene upon 81discrepancies for every 290 patients. And so by our calculations, the intervention would pay for itself.

In the United States, people are starting to pay attention to readmission rates – particularly 30-day readmission rates. We did not factor in how much those readmission rates are costing hospitals. Insurance companies and the Federal Government are going to stop paying for those readmissions.

Given the lack of reimbursement for readmissions, the 1 in 290 figure is probably an overestimate of what is required for the intervention to pay for itself. When Bates et al. were calculating at the costs incurred by ADEs, they were not taking readmission costs into consideration. So, I think it is very likely that it is not 1 in 290, but it’s 1 in a much bigger number. No matter what, I think it is a pretty good bet that this would have paid for itself.

Has cost previously been a reason why medication reconciliation has not taken place in hospitals?

Medication reconciliation does take place in hospitals. The joint Commission here in the United States has been requiring hospitals to complete this process. But most of the time it is left up to the physician or the nurse who is taking care of the patient. These people, however, have other jobs to be doing to take care of the patient.

Our question was: does this need to be the job of the physician or nurse taking care of the patient – who have lots of other things they need to be doing – or could this data gathering be done by someone who does not have other patient care responsibilities.

This makes for better use of the physician’s and bedside nurse’s time. You want your physician to have this list, but you don’t want the physician to spend most of their time figuring out the list. It would be better if someone else created a full list and just handed it to the physician.

This would free up the physician to think about other things, such as why is the patient here now, how can I educate the patient about their illness and their medications, how can I make sure that their primary care physician knows what is going on and that their transition from hospital to home is as smooth as possible.

The physician has more time to attend to these important matters because she does not have to spend time calling the pharmacist, speaking to family members back home while they search through all of the medication bottles in the medicine cabinet, or calling the primary care physician to ask about the patient’s medications.

Are there plans to implement medication reconciliation in hospitals from now on?

This is a good first step for us to show that it looks like it would be cost-effective and that it would improve patient care. We really need to convince the system that it does improve outcomes and comes at a reasonable cost.

Our next step is to investigate this in a randomised fashion in our patient population here in Baltimore.

In a recent medication reconciliation study in Boston and Nashville, Pharmacist Intervention for Low Literacy in Cardiovascular Disease, they did not see the benefits that we thought we would potentially see given our study results. This may be due to a very different patient population. Their population tended to have a higher literacy rate. Our patients here in Baltimore struggle more than in some other places with health literacy.

I think there is a much better bet that a medication reconciliation collaboration done here in Baltimore, with a more vulnerable patient population, is more likely to improve clinical outcomes and it would come at a reasonable cost.

Taking into account everybody’s different computer systems and their patient population characteristics, it may not be the case that one answer is right for every hospital. This may need to be individualised to the patients’ needs of that particular hospital.

Hopkins at this point owns 6 different hospitals, and I don’t think that all of our hospitals would necessarily benefit from the same intervention. I think it is important to know your patient population and the struggles that your patients deal with, as compared to what may be going on in other places.

It also may be an intervention that you don’t apply to every patient. In our next study we are probably only going to enrol patients who have a low health literacy rate. And so, we might be able to target this within specific patient populations within our hospital system. This may also make the system more cost-effective and more effective in terms of the outcome.

What are the main challenges in medication reconciliation and how do you think these can be overcome?

I think the main challenge is the amount of time that it takes to come up with a home medication list. I think it is overcome by using a team approach to try to determine what that home medication list is.

The whole point of our study is to show that the physicians or the nurses shouldn’t be trying to do this on their own. It takes a team approach. We need members of that team who have time dedicated to the task of determining what the home medication list is.

How do you see the future of patient safety progressing?

There has been an amazing change in the importance of patient safety over the last decade. When I was a medical student, 15 years ago, the focus on patient safety was not nearly the same as what it is now.

I think the world of patient safety is only going to grow. Johns Hopkins has now started the Armstrong Institute for Patient Safety and Quality. When an institution like this creates a whole institute to focus on patient safety, then I think it shows how far that issue has actually come.

The Federal Government, the insurers, the hospitals, the outpatients facilities are all looking at patient safety. When this many groups are immersed in the subject, then change is bound to happen. Hopefully, it will be for the better.

Do you have any plans for further research into this area?

We would like to further evaluate this idea, specifically in a patient population with a low literacy level. We are also experimenting with who should be involved in compiling the home medication list.

In our last study, we used registered nurses. In our next study, we are probably going to use 4^th year pharmacy students. This is a group of individuals who are not paid at all, and could really learn a lot by being involved in the medication reconciliation process. This may mean that the intervention would be almost cost free, yet would probably result in the same improvements for clinical outcomes for patients.

This time our hope is that we can show that we have an intervention that costs next to nothing and verifiably improves clinical outcomes.

Would you like to make any further comments?

I think this is an extremely important issue that for too long has been relegated to a back seat. We have been relying on people to compile these lists who just don’t have the time to do it well.

I think most people involved in healthcare right now are beginning to see that healthcare cannot just be physician-driven. We really do need to work as teams and these teams can result in better outcomes for the patients, better satisfaction for the physicians and so forth.

My physicians would be very happy to give the responsibility of compiling the list of home medications to someone else. We need to continue as medicine evolves to make sure that we are using these team approaches to lead to the best clinical outcomes for our patients.

Where can our readers find more information?

They can find the paper here: http://onlinelibrary.wiley.com/doi/10.1002/jhm.1921/abstract

Readers can also look at our Johns Hopkins hospital group here: http://www.hopkinsmedicine.org/the_johns_hopkins_hospital/

About Dr. Leonard Feldman

Lenny Feldman, M.D., SFHM, FAAP, FACP is an Assistant Professor in Internal Medicine and Pediatrics and a hospitalist in the Division of General Internal Medicine at Johns Hopkins University School of Medicine. He serves as the program director for the Johns Hopkins Med-Peds Urban Health Residency, the track director for the Osler IM Urban Health Primary Care Track, as an associate program director for the Osler Internal Medicine Residency, and as the Director of the General Medicine Comprehensive Consultation Service.

As a hospitalist, program builder, and a specialist in resident and hospitalist education, Dr. Feldman has focused his research on consult medicine, safe transitions, on-line education, and systematic reviews.  

Other interests include patient education, urban medicine, evidence-based medicine, and quality improvement. He is the editor-in-chief for SHM Consultative and Perioperative Medicine Essentials for Hospitalists and an Associate Editor for the Journal of Hospital Medicine.