Vanderbilt University Medical Center (VUMC) has a rich history of encouraging clinical leaders and informatics experts to create innovative solutions for clinical problems. Such encouragement is typified by the homegrown computerized prescriber order entry (CPOE) for both inpatient and outpatient use and the powerful and highly sophisticated electronic medical record systems we developed in-house.
In addition to these systems, we have developed several homegrown electronic dashboards that are routinely used across the clinical enterprise. One of these dashboards is known as the adverse drug event (ADE) dashboard, and this tool has been a cornerstone of clinical pharmacy practice for many years. Although the ADE dashboard platform has been a very useful tool, its development, along with others, hinted that the continual creation and customization of new dashboards for clinical surveillance, intervention, and documentation would begin to tax our resources to the point of rendering the process untenable. Whether homegrown or commercially acquired, clinical surveillance, intervention, and documentation dashboards can be vital to reducing ADEs. In order to standardize the process of creating clinical dashboards to improve safety and mitigate risk, we are currently in the process of transitioning to a commercial dashboard development and management system as part of a overall project to install a new pharmacy practice model.
Our current practice model deploys pharmacists in several centralized areas and satellite pharmacies, which limits the clinical involvement of our front-line pharmacists. Our vision for the new practice model will re-deploy pharmacists out to certain patient care areas where their clinical role will be leveraged to better serve patients, nurses, and providers. A more decentralized role will allow our pharmacists greater opportunities for direct patient care, improve communication with medical teams, troubleshoot problems with the medication use system, and ultimately improve pharmacotherapy and patient outcomes. The electronic dashboard network will help facilitate the new practice model by more efficiently identifying patients that require pharmacist involvement.
Why Dashboard Surveillance?
Although pharmacists have long been regarded as key players in the medication use system, our role in improving therapeutic outcomes and reducing the risk of ADEs has received increased attention from other health care practitioners of late. For some time now, health system leaders have been becoming increasingly aware of the benefits comprehensive pharmacy services have on patient populations and studies have repeatedly demonstrated the value of a wide variety of pharmacist services over the last few decades. Systematic review of the literature provides clear evidence that pharmacist involvement in direct patient care leads to the improvement of numerous therapeutic and safety outcomes including hemoglobin A1C (HbA1c), LDL cholesterol, blood pressure, medication adherence, patient knowledge, quality of life, and ADEs. Decreased readmissions, lengths of stay, and mortality have also been directly attributed to pharmacist interventions.
The Institute of Medicine (IOM) estimates that over 1.5 million preventable ADEs occur each year due to medication errors at a cost of several billion dollars and as a result, the IOM now recommends that pharmacists take an active role in reducing these events. Furthermore, IOM recommends that health care systems, especially inpatient settings, ensure that pharmaceutical decision support be available at all times. Today’s health-system pharmacists understand that ADEs are a costly liability to any patient care setting, and the evidence clearly shows that pharmacist interventions reduce the risk of ADEs. Electronic dashboard surveillance should allow virtually any health-system pharmacy to leverage their pharmacists’ experience and knowledge to reduce ADEs and to document these efforts for regulatory compliance.
Value of ADE Dashboards
For several years, pharmacists at Vanderbilt have used homegrown ADE dashboards to support therapeutic drug monitoring (TDM) services and to display select patients deemed at risk for ADEs related to warfarin, heparin, and aminoglycosides. A variety of rules were built into the operation of each dashboard based on factors that would pose a risk to a patient. If a patient surpassed certain criteria for a rule, an alert would be triggered on the dashboard. For example, if the serum creatinine increased by 0.3 mg/dL for a patient who has been receiving an aminoglycoside, the alert would be triggered. When an alert is triggered, a pharmacist will review the details of the case as presented by the dashboard and subsequently contact the appropriate medical team member to discuss the issue. Medical team members such as medical residents, attendings, or other providers are contacted via pager in most instances and they almost invariably accept the pharmacists’ recommendations. Any interventions are then recorded in the pharmacy’s intervention documentation system.
The first ADE dashboard to be developed at Vanderbilt was for warfarin use. The warfarin dashboard was initially conceived in early 2008 in response to the 2008 Joint Commission National Patient Safety Goal for Anticoagulation. It went live in the summer of 2008 and continued to be developed and improved upon over the next several years. This dashboard was populated with rules that would trigger alerts for INR > 3 and a rise in INR > 0.4 in a 24-hour period. Shortly after our pharmacists began using this dashboard, the amount time patients spent with INR > 4 was reduced by 50%. This was seen as a major achievement, as one of the overarching goals for patient safety at VUMC has long been to reduce the risk of adverse drug events. Additional rules were added to the warfarin ADE dashboard over time including an alert indicating inadequate INR monitoring (no INR for 48 hours).
The success of the warfarin ADE dashboard lead to the development of an ADE dashboard for heparin. The purpose of building the heparin dashboard was to alert the pharmacist when certain rate limits were exceeded or there was a lack of appropriate activated partial thromboplastin time (aPTT) monitoring. Like the warfarin dashboard, the heparin dashboard was highly successful, as it helped to reduce the number of patients with aPTT values > 200 seconds.These anticoagulation related dashboards were initially developed in order to meet and exceed The Joint Commission’s National Patient Safety Goal (NPSG.03.05.01) for anticoagulation therapy management. Through the use of these anticoagulation dashboards, our pharmacists carved out a clearly defined and critical role in effective anticoagulation management, and produced a valuable safety net for patients who are on warfarin without an INR or on heparin without aPTT or platelet monitoring. Rules and alerts for heparin-induced thrombocytopenia, enoxaparin dosing errors, and new anticoagulants were added more recently.
Anticoagulation Counseling and Education
One of the unique features of the anticoagulation dashboards is that they provide a vehicle for multidisciplinary patient counseling. Each time an anticoagulant drug is ordered for a patient, a red indicator appears in the electronic medical record signifying the patient must receive counseling on their anticoagulant medication. This bold, red indicator remains until documentation is entered into the medical record that counseling has been completed by the nurse or pharmacist, at which time the indicator turns green. This visual reminder facilitates coordination of what is a multidisciplinary task such that nurses know if the pharmacist has counseled the patient and vice versa without having to review the chart. This measure reduces time wasted determining if counseling has been performed and makes it very easy for supervisors and managers to see that all applicable patients have received anticoagulation education prior to discharge as defined by TJC’s NPSG.03.05.01. The anticoagulation ADE dashboards have enabled us to increase the percentage of patients with documented anticoagulation education from the 30 to 40% range up to 95% on a consistent basis.
Growth of Dashboard Value
Following the development and implementation of the anticoagulation-related ADE dashboards, we began to investigate other medication types that could benefit from active computerized monitoring and intervention functions. One of the first to be identified was our use of aminoglycosides. This dashboard was created to provide preemptive alerts for inadequate monitoring of serum creatinine, abnormal aminoglycoside serum concentrations, and other potentially serious complications, such as acute kidney injury. The aminoglycoside ADE dashboard also came to support a comprehensive TDM consult service by providing a vehicle for communication among pharmacists outside of the medical record. Of particular note, evaluation of the aminoglycoside dashboard showed that appropriate dosing of these agents increased from 40% to 80% with less nephrotoxicity.
Like the other dashboards, the aminoglycoside ADE dashboard became widely recognized throughout the hospital as an essential quality and safety improvement tool and the resulting success of the aminoglycoside dashboard led to the proposal for a vancomycin dashboard concept. It was during this development process that we realized creating homegrown dashboards required immense resources. Development of each of the ADE dashboards required a significant investment from pharmacy, biomedical informatics, and medical staff leadership and as it became clear that the further development of clinically valuable dashboards would continue to grow in breadth and depth, we began to question whether our in-house efforts would be sustainable. Driven by the potential development workload and a desire to standardize the process, we began researching a commercial solution to the creation of a comprehensive clinical monitoring, intervention, and documentation program.
Navigating the Fork in the Road
Given the culture at our hospital where homegrown products are the norm, it was a difficult decision to consider transitioning to a commercial vendor for clinical surveillance. Pharmacy and hospital administration leadership recognized that we had been highly successful in developing our own efficient, automated surveillance tools with a proven track record of reducing ADEs. At the same time, there was serious concern regarding the amount of time and resources spent on developing these tools, and what the time investment would be for continued development. We arrived at a fork in the road when considering the numerous drug therapies beyond warfarin, heparin, and aminoglycosides that are high risk, high cost, or complex enough to require pharmacy surveillance. Though we had dedicated ourselves to the care of patients who were on these therapies, we also wanted to expand the safety net provided by clinical dashboards well beyond anticoagulation and aminoglycosides. It was patently clear that as new drugs emerge and drug knowledge evolves, we would need a much more efficient method of adding, revising, and upgrading our dashboards and alerts. The new system would need to be robust, flexible, scalable, and user friendly for those who would program the rules as well as the end users. In evaluating commercial systems, we found that most vendors offered solutions that would meet our needs.
Several vendors came onsite to provide demonstrations of their software and most of the systems we evaluated were able to integrate ADT (admission, discharge, or transfer) information, drug orders, and laboratory results to create alerts for potential medication complications. Minimal or no technical expertise would be needed on the rule building side, as the rules are not pre-built, but do not require the use of code language or other technical computer skills. Rather, a template allows the user to select the drug, lab, and any ADT data in order to build the rules. While some rules are simple, others are complex to build, even within the template structure.
Once it was implementation was complete, pharmacy could manage the development of the tool without additional informatics resources. Some of the systems had a particularly user-friendly interface that could filter by patient care area, and this was an important consideration because one of the goals of transitioning to a commercial product was to use the surveillance dashboard to advance our new practice model. Under our previous system, the homegrown ADE dashboards were largely used by select clinical pharmacists and pharmacy residents; this was not in keeping with our desire to expand the clinical responsibilities of all of our pharmacists. Therefore, we sought to leverage the new system to expand the use of the surveillance dashboards to include more pharmacists. Thus, the new surveillance tool would be an opportunity to launch our new practice model out to all pharmacists.
Another factor that influenced our decision to change to a commercial product was that we were already using a commercial system to document our interventions. We felt that if the surveillance and documentation systems were integrated it would facilitate proper intervention and record keeping processes, and result in improved documentation of pharmacist activities. Increasingly, pharmacy departments are being asked to provide data on their activities by state boards and accreditation agencies, so we viewed this as an opportunity to leverage the surveillance system to support documentation.
Do Not Reinvent the Wheel
After the new dashboard system was installed, our first goal was to recreate the same rules that were so successful with the homegrown system. This process highlighted some of the strengths and weaknesses of both the homegrown and commercial systems. While the differences are numerous, they are mostly minor. For example, the homegrown system could extract almost anything from the medical record and incorporate it into the dashboard whereas the new system requires a bit more direction. Further to this, the homegrown system could trigger an alert based on a medication administered whereas the new system triggers an alert based on the drug order from the CPOE system—this difference matters, since not all medications that are ordered are actually administered.
Another notable drawback we identified was that the homegrown system could calculate the creatinine clearance for a patient regardless of whether there was a patient height recorded in the record, whereas the new system requires a height, without which the rule will not work and no alert will be triggered even if the patient has renal failure. These differences have a direct impact on the type of rules that can be built, how those rules are built, and the scope of services that pharmacy can provide. As we adopt to this new way of doing things, we are still creating some workarounds for rules that worked well in the homegrown system, but do not work as well in the commercial system.
Despite some of these obstacles, there are several advantages to the commercial system that will facilitate further design, development, and implementation of additional dashboards. For example, the program allows for full integration with our clinical documentation system, so the pharmacist does not have to leave one system to document an action in another system. We now have autonomy with regard to writing and testing new rules and do not have to wait for or depend on IT support. Furthermore, vendor training and support for the product is included, making it easier to bring users up to speed, and we also have access to a large network of other system users who share experiences, ideas, and solutions.
Today, we are starting our transition to the new dashboard system with great excitement and enthusiasm. The first set of clinical rules have been built and tested, and we are in the process of training our staff on the new software. There are differences in the workflow between the old and new system, but our personnel are beginning to adapt to these changes. In the near future, we will add new rules and expand the number pharmacists who are trained to use the system, as this will be a cornerstone of our new pharmacy practice model.
Bob Lobo, PharmD, FCCP, BCPS, is the clinical pharmacy services and drug policy manager at Vanderbilt University Hospital in Nashville, Tennessee. He is also an adjunct faculty member at the University of Tennessee and Belmont University School of Pharmacy in Nashville. Prior to his current role at Vanderbilt, Bob practiced as a clinical pharmacy specialist in internal medicine for over 15 years. He also is an active member of the American College of Clinical Pharmacy, ASHP, and the Tennessee Society of Health-Systems Pharmacists. Bob received the Distinguished Service Award from the Tennessee Society of Health-System Pharmacists in 2008.
- Asolva, Inc
- Baxter Healthcare Corporation
- bioMerieux, Inc
- Cerner Corporation
- Health Care Systems, Inc
- IBM Watson Health/Truven Health Analytics
- Integrated Informatics
- McCreadie Group
- Mediware Information Systems, Inc
- MedKeeper (A Grifols Company)
- Midas+ Solutions, now part of Conduent Health Analytics Solutions
- OZTech Systems, Inc
- PatientSafe Solutions, Inc
- Pharmaconomics, Inc
- Philips Healthcare
- QSolutions, Inc
- Wolters Kluwer
Part 1 of a 2-Part Series: Elements of a USP <800> Compliant Cleaning Program
Conduct a Drug Diversion Investigation
Special PP&P Buyer's Guide: Temperature Monitoring
Develop a Pneumococcal Vaccination Program
New & Noteworthy
- In The Loop!
- Digital Edition
- Special Announcements