Clinical Decision Support: The Technology Is Now Equal to the Challenge

PDF File: 

HCT2_wp_bormel.pdf

QuadraMed Corporation

By collecting knowledge in real time, next-generation clinical decision support not onlyimproves workflow but directly enhances the diagnostic and therapeutic abilities of caregivers.

Health care is a topic that appears as a primary voter concern during virtually every major election year in the United States. Because Americans enjoy what is arguably the world’s finest system of care, in the past, concerns have centered largely upon its accessibility and cost. Seldom had the question of quality come to center stage except when an isolated medical mistake briefly made news headlines.

But the focus began to change in 1999, when the National Academy of Sciences’ Institute of Medicine (IOM) released its now famous report “To Err Is Human.” Although the issue of medical mistakes had been quietly discussed since at least 1955, the IOM report, noting that up to 98,000 adults treated in health care institutions each year were victims of medical mistakes resulting in death, grabbed the attention of the media and elected representatives at every level of government.

Subsequent papers and opinions published by health care industry analysts indicated that too little has been done by institutions and care providers to effectively deal with what is now considered by many to be a crisis situation. “Healthcare Predictions for 2003: Slow Evolution,” produced by the Gartner Group, provided an estimate that preventable medical errors will decrease by only 25 percent system wide by 2007.

In 2003, two important studies were released that underscored the need to apply next-generation technology for improved patient safety. RAND Health produced what has been called the largest and most comprehensive research project conducted to date of health care in the United States. According to the report, “The Quality of Health Care Delivered to Adults in the United States,” published by the New England Journal of Medicine, roughly 45 percent of adults do not receive the recommended health care they need.

For example, RAND found that patients with high blood pressure routinely received less than 65 percent of accepted recommended care, and that such poor control contributes to over 68,000 preventable deaths each year. It is important to note that the methodology employed to produce this study incorporated data from more diverse health problems across the entire nation than ever before, and included individuals with all types of insurance who were healthy, as well as patients who were ill. The only logical conclusion we can draw from the data is that far more adults suffer negative outcomes than originally estimated in the IOM report.

In part, the RAND authors reported that there is a need for technology to provide information on required measures of care as an integral part of the process to deliver that care. Further, the health care industry must focus on automating the entry and retrieval of key data in order to provide proper support for the clinical decision-making process.

Worthy of mention is a second report released in 2003 by the Institute for Healthcare Improvement. The publication, “Move Your Dot: Measuring, Evaluating, and Reducing Hospital Mortality Rates,” provides institutions with a guide to help them learn more about their performance relating to mortality, their relative position when compared to other hospitals, and to promote an understanding of how mortality rates can be improved.

Using a specific methodology, the paper depicts the mortality rates for 1,739 of the nation’s hospitals. The authors concluded that there is a “450 percent variation in a patient’s chance of dying, depending on which hospital he or she enters.” The paper also includes the following findings:

“Concurrent with the industry’s efforts to improve care, purchasers, and consumers are increasingly focusing on information about quality to choose providers. ‘Report cards’ detailing process and outcomes data are increasingly available at the health plan, hospital, and provider group levels.

“In addition to feeling the professional imperative to provide the highest quality care, hospital executives are also feeling competitive pressure to demonstrate quality by publishing outcomes data. A measure that carries significant emotional weight, as well as important implications about quality, is the hospital’s mortality rate.”

We can deduce from IHI’s findings that those institutions with a greater number of negative patient outcomes not only suffer under the weight of greater risk, but a level of public scrutiny that may ultimately threaten their very existence.

Clearly there is not simply a need to implement next-generation clinical decision support (CDS) by health care institutions; there is a critical requirement that such technology be deployed to improve patient safety and outcomes across the continuum of care.

Clinical Decision Support – The Timing and Technology Are Right

Simply defined, CDS is the systematic approach to collecting knowledge, and providing it in real time to clinicians as a direct aid to enhance diagnostic and therapeutic efforts.

CDS offers clinicians the four things they want most from information technology: a simple process to organize work; a convenient way to review information, make decisions, and enter orders; assistance to become better at their profession; and a dependable methodology to improve patient outcomes.

In the past, the core concept for making diagnostic and therapeutic decisions was based upon a clinician’s ability to identify the malady, review pertinent written materials, consult with peers, combine all the input with any personal knowledge of the patient, and then produce a customized plan for optimal care. This approach to health care is time-consuming, inconsistent, and potentially chaotic.

Certainly, the personal experience of each clinician is important, but when such experience is available in a common repository, along with alternative diagnoses; therapies (of both known and unknown efficacies) and their potential complications; plus a realtime patient history; clinicians can practice evidence-based medicine (EBM) using greater knowledge to manage individual patient needs. Invariably, the results will be fewer medical mistakes.

The unfortunate norm in today’s marketplace is “retrospective” CDS, which offers only after-the-fact analysis, typically when outcomes can no longer be affected. It takes an overly simplistic approach based upon improving workflow through engines derived from totally unrelated business functions.

But EBM-enabling technology, the obvious choice to improve patient safety, is currently available from a small percentage of health care information technology vendors. Therefore, vendor selection is critical, and to achieve full functionality, decision makers must not expect next-generation CDS systems to interface with diverse applications supplied by multiple vendors.

This “next-generation” technology implements “concurrent” CDS, which provides interactive decision-support capabilities “just-in-time,” during actual care delivery when patient outcomes can be positively impacted.

Using a concurrent methodology, real-time context-sensitive decision support includes diagnostic information (what is going on with the patient), merged with relevant therapeutic options (what the care practitioner can do about it) to produce integrated knowledge management, rather than the severely limited benefit of improved clinician workflow. Early attempts at CDS resulted in systems that were too slow, overly complex, interfering, or otherwise unusable, highlighting the critical requirement for next-generation, intuitive design which incorporates the unique needs of the health care industry. Such a design is characterized by simplicity, consistency, ranking, filtering, and local customization to produce a viable, integrated CDS system.

Therefore, a concurrent CDS system takes into consideration diagnostic needs in terms of what is happening with and to a patient at a particular moment in time, and the therapeutic alternatives the clinician has to affect a positive patient outcome, based upon patient- and context-specific information. This is “just-in-time learning.”

Primary Components of a Concurrent CDS System

• Decision-support capabilities are tightly linked to clinician workflow.
• A decision-support system is fully integrated with the reference information model of the care provider’s HIS (health information services) through fully controlled terminologies and vocabularies – from initial data capture through final clinical documentation.
• Context-sensitive display of all patient information, so that relevant content automatically appears at the right time and in the right places within clinical workflow applications – what you need to know, when you need it, in a way or format that you can readily use.
• Incorporation of real-time knowledge management capabilities through end-user authoring of local content, not just canned, commercially-developed content.
• Automatic generation of information sets, a mechanism of associating relevant patient information, potentially useful orders, references, guidelines, and rules with clinical conditions, observations, and orderable items, with “just-in-time” display of all relevant diagnostic and therapeutic information, so that mistakes can be avoided or caught before they effect a patient outcome.

Since people are fundamentally fallible through unintentional omission or mistakes, and can be overwhelmed by large amounts of data, next-generation CDS systems can provide a safety net for clinicians in even the most routine clinical situations.

Through all of these mechanisms, concurrent decision support is leading-edge knowledge management that greatly improves the probability for successful patient outcomes. That is the primary reason for accepting and implementing CDS – deployment of a Web-based, intuitive system that recognizes facts, anticipates the clinician’s needs, and provides these facts automatically to minimize mistakes.

In short, next-generation CDS offers the potential to not only improve workflow, but to literally revolutionize the decision- making process through real-time knowledge management.

CDS Push

The argument in favor of CDS seems pervasive, yet as evidenced by the Gartner projection, there is considerable “push” in some circles against adopting this technology. As is the case with most innovative approaches to problem solving, two primary areas of resistance exist: first from those traditionalists who by definition resist change, and then by those who find the technology imperfect.

Admittedly, there is an individual learning curve for every user of new technology. This takes time that traditionalists equate to “robbing” patients from needed personal attention (but ignores the proven fact that such personal attention, in many instances, produces negative outcomes due to a lack of sufficient knowledge). Another argument is that even the best next-generation technology is imperfect. True, but intelligent knowledge management technology is now significantly better than clinicians working as virtual islands unto themselves, who become aware of medical mistakes only after the fact.

The “best-of-breed” next-generation CDS systems are not based solely upon the thinking that produces pure workflow engines. That is, effective CDS does not, in most cases, reduce the time necessary to effect improved patient outcomes. Instead, pereffective CDS reduces medical mistakes by presenting a comprehensive base of objective knowledge and alternatives that clinicians can use on a just-in-time basis to improve care on an individual basis. Logically, this makes sense, but it also takes more time at the front end of treatment. Considering the alternatives, and the existing state of medical mistakes, implementing CDS technology is the common sense approach to risk reduction through improved patient outcomes.

We all need to understand that without the implementation of technology, too much information needs to come together and be assimilated in real time to ensure clinicians can provide safer care. The existing process is dependent upon information that comes from peers, institutions, third-party vendors, and personal experience – all of which are probably mutually inconsistent and impossible to merge in real time to benefit the patient. Plus, there is the need to decipher and separate the important from unimportant information within the context of treatment for an individual patient that must also be available in real time.

Technology such as CDS enables clinicians to readily assimilate the system in short order, with minimal effect on their already acute workloads. The fact is, such systems mimic established “ways of doing things,” including word-of-mouth and paper-based processes, but provide data just in time to promote better decisions and greater clinician knowledge. Next-generation CDS systems emulate the “comfort zone” of the traditional verbal and paper environments, while improving upon them by streamlining the decision-making process.

Employing change management to ensure acceptance and assimilation from push detractors of CDS is mandatory. In his book, Managing At The Speed Of Change, Daryl R. Conner described change management as creating awareness, then acceptance, and finally, commitment. This thinking enables end users of CDS to refine, and perhaps even redefine the process to ensure the user experience, technology, and content are acceptable from concept through implementation. CDS proponents are perfectly justified in a measured degree of “shove” when it comes to critical “push” against intelligent systems.

The second argument is that CDS systems are imperfect, and are nothing more than an attempt to further restrict physician freedom in diagnostic and therapeutic processes. The first part of this argument is true. No system, in fact, is perfect. However, it is clear from all available sources that existing manual systems produce an unacceptable number of medical mistakes, thereby subjecting the health care system to greater scrutiny. Therefore, contrary to the second part of this argument, a CDS provides the technological foundation for clinicians to provide consistently high quality care, not to restrict them. CDS is very consistent with the positive aspect of medical culture, which values continual improvements in clinical practices.

The thesis on CDS design is that user experience, technical performance, and content management are the three competencies that must be done well individually, and must hang tightly together.

Conclusion

The concept of CDS is not new to the health care industry. It has been available in one form or another for more than 25 years.

During the past two years, systems have rapidly been refined to meet a primary goal of providing clinicians with reliable information in a user-friendly format that intuitively supplies the right facts at precisely the right time. CDS gives clinicians immediate access to knowledge management.

Chief among the challenges to successfully implementing a CDS is its acceptance and assurance that the patient data contained within the system are correct. No matter how well a system is conceived, the validity of its data and its assimilation by clinicians determine its worth. CDS developers have effectively addressed these challenges, and system implementations are gaining momentum. But with less than 5 percent of health care institutions having implemented a CPOE application, much less a total CDS, the industry has a long way to go to achieve a more reasonable level of patient safety.

The importance of CDS as a critical component of EBM is not disputed, nor is its ability to operate in a highly intuitive manner. But no matter how sophisticated such systems may become, even the most optimistic developers recognize that the value of each clinician’s experience, perception, and common sense, are and will remain, the most important factors affecting the quality of care as well as patients’ perception of that care.

However, the tools of technology that are now readily available are proven to help care providers make better choices, lower their risk, and raise their level of expertise. In the November 2003 report, “Patient Safety, Achieving A New Standard of Care,” IOM has taken the position that the health care industry must develop a culture of safety that stresses error prevention by adopting technology for decision support and inclusive patient safety data systems.

Simply reporting medical mistakes does nothing to solve the existing problem. Prevention is the cure, and next-generation technology is equal to the challenge. The level of patient safety will increase dramatically when CDS becomes ubiquitous as a tool for improved outcomes.

Endnotes

1. Capgemini Ernst & Young. “Health Care’s Top 10 Business Issues and Impacts for 2003,” 2003
2. Institute for Healthcare Improvement. “Move Your Dot.” Sir Brian Jarman, M.D., Senior Fellow, IHI; Tom Nolan, Ph.D., Senior Fellow, IHI; Roger Resat, M.D., Fellow, IHI, et al – 2003
3. Institute of Medicine.“To Err is Human: Building a Safer Health System,” 1999.
4. HIMSS. “A Technological Approach to Enhancing Patient Safety,” Kathleen Covert Kimmel, Joyce Sensmeier – 2002.
5. IMIA Yearbook 2003. “Leveraging IT to Improve Patient Safety,” Marion J. Ball, David E. Garets, Thomas J. Hardler – February 2003.
6. JAMA. “Effects of Computer-Based Clinical Decision Support Systems on Physician Performance and Patient Outcomes,” Dereck L. Hunt, R. Brian Haynes, Seven E. Hanna, Kristina Smith – October 1998.
7. RAND Health.“The Quality of Health Care Delivered to Adults in the United States,” E. A. McGlynn, S. M. Asch, J. Adams, J. Keesey, J. Hicks, A. DeCristofaro, E. A. Kerr, New England Journal of Medicine – 2003.
8. The Gartner Group. “Impact of Computer-Based Record Systems on Medical Error Reduction,” Barry Hieb – April 2002.
9. The Gartner Group. “Computer-Based Patient Record Systems: Addressing the Continuum of Care and Caregivers,” Thomas Handler, M.D. – April 2002.

QuadraMed is a registered trademark of QuadraMed Corporation. Intelligent Care Sets is a trademark of QuadraMed Corporation. All other trademarks are the property of their respective owners.