Evidence-based medicine grows up: by strategically addressing the challenges and user...

Clinical research has demonstrated that specific medical interventions are the most effective practices for treating certain medical conditions. When information about these interventions is institutionalized and shared with physicians through appropriately designed clinical processes and databases,

Embedding leading practices within an information technology framework can reinforce their use by providing tools for physicians where and when they need it. This is the most critical step in sustaining a successful evidence-based medicine (EBM) program.

As hospital executives make progress in implementing EBM systems for their organizations, they must consider how to evolve their existing clinical information systems to integrate the various technologies and tools that support evidence-based care. They must also determine how to make these systems easy for physicians, nurses and other clinicians to use at the point of care.

The Challenge: Delivering Sufficiently Detailed Data

Although an increasing amount of solid data supports the value of evidence-based medicine, many of the commercially available EBM tools provide high-level guidelines rather than the kind of detailed data that physicians need to support their clinical decisions on the front lines of providing care. While general guidelines are useful, they are typically not sufficient.

Physicians are looking for systems that provide specifics, such as the right drug at the right dosage for a patient with specific symptoms. The target is the capability of achieving compliance with 80 percent of appropriate patients within a specific diagnosis category. Systems that fail to provide information at this level are much less likely to be adopted by the medical staff.

Since many commercial clinical information system vendors are relatively early in their development of the technology capable of supporting advanced EBM functions, some leading hospitals have chosen to develop more sophisticated solutions internally. However, the time and financial investment necessary is too great for most hospitals. Instead, they are pacing their implementation, planning to have the process and content foundation ready to go once the vendor has developed a more advanced EBM tool.

Incorporating the Evidence

Hospitals need to incorporate EBM into various functions of their existing clinical information systems, such as the order entry components or the rules and alerts. The quality of the evidence included is critical: too high a bar, and very few studies will be considered and much care will be unmanaged; too low a standard, and current common practice will become the "evidence" with no real quality improvement. Finding the right level of rigor must be driven by content experts with ownership from the local clinical leadership.

One common challenge for many organizations is migrating from an existing technology platform, with many older systems unable to support some of the complexities that EBM order sets entail. An example of EBM functionality that adds value to the system but is complex is the ability to support nesting orders that trigger the opening of secondary fields. Tools such as Boolean logic, sliding scale algorithms and dependency-based order sentences all work on paper but can be very difficult to program. Order sets must be built that match the technology's capability--neither too complex for the IT interface tools nor unnecessarily simple such that the power of the system is wasted.

Another key need is a careful upgrade and replacement strategy. Hospital executives must take a long-term view and critically evaluate the underlying architecture when selecting a system. They must consider the impact that different system components have on one another. For example, the physician order entry, the pharmacy and the electronic medication administration modules must share an integrated, functional architecture. A simple interfaced solution will not allow the functional integration necessary to support rules and alerts. The rules (if/then logic, for example) require a level of data analysis and interpretation that must occur in a single logical database. It cannot be spread into multiple data files while retaining the ability to build the necessary algorithms. For example, choosing a pharmacy system without careful consideration of the required related systems (e.g., order entry and nursing documentation) can unintentionally lock a hospital into the chosen vendor's other products.

Building a Scenario-based Model of Care

A tiered, scenario-based model of care effectively creates depth and complexity in a clinical information system. This kind of complex adaptive modeling applied to evidence-based medicine represents the next generation of development.

A recent clinical information systems implementation in New Zealand utilized this approach. Participants began with a clinical diagnosis and built patient-specific scenarios around it. These scenarios represented the spectrum of likely clinical events and allowed much better matching of the needs of an individual patient.

For example, in the diagnosis of asthma, the project team literally built dozens of scenarios reflecting the many clinical experiences that a population of patients with this diagnosis could have. Instead of one asthma pathway, they used EBM to come up with more than 20 different order sets that would match the clinical experience of almost any patient with asthma. They identified measurable patient characteristics, such as how quickly the patient could exhale, along with fever, blood count, current medications and other charted measures, and used this information to present a recommended set of orders that could be applied consistently and accurately. The algorithm was complex, but hidden from the busy front-line care provider. The result was a set of orders, recommendations and alerts that much more closely matched the needs of the specific patient--and that could change along with the patient's condition.

Designing Effective Rules and Alerts

Designing rules and alerts effectively also helps maximize the potential advantages of evidence-based medicine. A well-designed system helps support a physician's ability to make appropriate clinical decisions and comply with the guidelines built into the system. But if the alerts are not focused properly, or pop up too often and are considered obvious or annoying, their value is diminished. As a natural human behavior, clinicians will start to automatically dismiss all of the alerts, rather than taking the time to distinguish those that are helpful from those that are merely "noise." The key is to involve care providers during the design process in finding that "sweet spot"--not too much or too little--recognizing that the balance point will evolve as user sophistication grows.

Usefulness of the alerts is a key consideration. A warning about a patient's medication allergy that also provides a recommended alternative medication--and a single mouse click to order the drug to replace the initial order--can save time for clinicians while enhancing the quality of patient care. The medical staff is much more likely to recognize the value of such alerts and, consequently, to use them.

Even when high quality evidence is built into a sophisticated technology platform, it is not an effective solution without acceptance and adoption by front line providers. Working with active leadership from the facility's medical staff will best ensure that order sets, rules, alerts and other evidence-based guidelines built into a clinical information system are properly used. Clinical ownership is necessary in both design and implementation for the end product to be well used.

A clinical information system implementation reflects the broader organization's challenges. If the medical staff and hospital have an aligned and harmonious relationship, they are likely to see that reflected in the implementation of an EBM project. An organization that has turbulence between staff and hospital, or within the medical staff, often sees that turbulence translated to such a project. In fact, the system implementation shines a bright light on the underlying cracks and flaws of an organization. It may be challenged not because of bad technology or EBM tools, but because of underlying organizational problems.

Ease of Use

The first key to any successful clinical information system is ease of use at the point of care. The functionality of screen navigation, rules and alerts, and information entry must be designed as clearly and simply as possible. Some essential points to consider include:

* Minimize the distance to information by displaying the most-needed data at the most immediate level. Always limit the number of mouse clicks and windows to access necessary input screens or information.

* Use color, font size and other visual differentiators to emphasize the most critical elements of information in a display (e.g., red abnormal lab values, flashing high priority messages, gray for completed fields).

* Use screen real estate carefully. Too much data compression on a screen causes information overload and increases the chance of the user missing essential elements. Show only what is critical to the functional purpose of an interface page. Blank areas and vendor logos not only consume valuable space; they also waste the users' time.

* Use tiered logic to drive use of the interface and nature of the data at every level. Remember that data can be entered via one interface but displayed in an entirely different format, layout, sequence and structure.

* Be careful about using legacy screen design elements to increase the comfort of users accustomed to old systems. Doing so may block the ability to use tiered logic, hyperlinked text and other, more efficient means of presenting data. Think about the 1000th time the user accesses the system, not just the first time. If training clinicians on a new user interface is inevitable, hospitals should be willing to address this need sooner rather than later.

* A properly designed interface must be flexible in handling different types and levels of users. Make sure that access and presentation of data are defined by the role of the user.

Accessibility and Training

Many organizations underestimate the number of necessary workstations. They need to plan for peak volume times, not just the average use rate. When hospitals install an inadequate number of terminals, user frustration typically ensues. No matter how fast the network is, clinicians will complain if they have to wait in line to use the system.

In planning system capacity, hospital executives need to analyze use dynamically and implement to match as much as possible. The realities of limited budgets and limited counter space in a busy nursing unit factor into the equation. Some organizations have chosen flat screen monitors mounted on the wall and accessed via cordless keyboards, hand-held PDAs and small, full-function laptops connected to a secure, wireless network to augment traditional desk mounted workstations.

Whatever the solution, comprehensive training teaches physicians, nurses and clinicians how to use the system, and also indicates how processes and workflows can be changed and optimized. The most successful training efforts utilize real-life, real-time, everyday scenarios. Using care unit-based "super users" to coach and answer questions in real time is at the core of many successful go-lives. While group sessions are important, the implementation team should also take advantage, whenever possible, of departmental meetings that the staff are already attending to train them in a live, interactive way. The system design and implementation team should walk through care of a typical patient with clinicians. The better trained the clinicians are, the more productive they will be when using the system.

Technology enabled, evidence-based medicine can provide safe, high-quality care in a rapidly changing environment by bridging best practices and information technology to create a system for care delivery that meets patient and provider needs. As the body of evidence-based knowledge grows and develops, and advancing technology gains in sophistication, the best care will be provided, without question, by organizations that understand the necessary investment.

EBM at Work: Knowledge Driven Care

Indiana University Hospital in Indianapolis, a Clarian Health partner, is a tertiary-care referral hospital with nationally and internationally recognized programs in specialized care. In partnership with Capgemini, the hospital recently implemented a program called Knowledge Driven Care. The program combines evidence-based medicine with clinical process redesign by using an advanced clinical information system.

Says Dr. Herb Cushing of Indiana University Hospital, "We're building a system that displays the proper prompts, guidelines and reminders to physicians when they are entering orders about their patients. This information is based on the medical literature and provides preventive care screening guidelines for primary care physicians. Our doctors developed all of our own order sets, a substantial effort that has already brought us more in line with best practices." The focused approach of Knowledge Driven Care will drive superior clinical outcomes, service excellence, operational efficiency and financial performance.

Adds Dr. Cushing, "Our end goal is to make sure that physicians can view medical information about patients at the time that they're seeing and taking care of those patients. We are able to display information about all our patients at all of our sites, including Methodist Hospital, Riley Hospital for Children, and a new hospital, Clarian West. That's a huge benefit and helps us deliver terrific patient care."

For more information about healthcare-related services from Capgemini, www.rsleads.com/506ht-203

Manuel Lowenhaupt, M.D., is a vice president at Capgemini Health and the national practice leader for clinical transformation. Contact him at manuel.lowenhaupt@capgemini.com.