By Jonathan Teich, M.D.
To achieve national goals for enhancements in quality, patient safety and cost control, while qualifying for incentives under the American Recovery and Reinvestment Act (ARRA), healthcare professionals in practices and organizations of all sizes should adopt and use electronic health records (EHRs) backed by sound clinical decision support (CDS), with particular attention to performance on evidence-based clinical guidelines. The practical aspects of how to complete the adoption process can be elusive. In the past, many organizations achieved major advances toward these goals using CDS, while others struggled and did not recognize the full anticipated benefits.
The industry already knows that it needs effective, adoptable CDS to fulfill government expectations and mandates, to realize the promise of information technology for improved quality and safety and to report on performance quality metrics in a meaningful way. The key is turning critical quality guidelines into CDS interventions that meet clinicians' needs and fulfill the multiple criteria of usability, value, cost effectiveness and ease of implementation.
The Agency for Healthcare Research and Quality has outlined an ideal loop of interaction among quality organizations, guideline developers, CDS producers, EHR vendors, providers, patients and reporting agencies. All are critical participants in making effective CDS adoption, impact and measurement a reality.
The loop effectively begins at the top with quality-focused organizations that determine, collect and disseminate clinical recommendations and guidelines based on scientific evidence. These organizations include specialty societies such as the American College of Cardiology and American Heart Association, as well as consortia such as HHS's national patient safety task force and the National Committee on Quality Assurance. Some quality organizations overlap their functions with the next component in the loop - the guideline developers who convert prose recommendations into more quantitative and actionable specifics, including data-set specifics to help define qualitative terms such as "congestive heart failure" or "patient takes a beta blocker," and numerators and denominators for performance measurement.
Following guideline developers in the CDS loop are CDS producers and content providers. CDS producers take specifically defined measures and turn them into practical forms that can be used by providers, such as order sets, reference guides, alerts, care plans and analytic tools. These tools are then provided as self-contained packages or built for incorporation into EHRs, allowing providers to implement quality guidelines within the day-to-day clinical work flow.
If a quality organization recommends that pneumococcal vaccination should be given to all diabetic patients, for example, a guideline developer might translate that into an equation such as "all patients who have a charge code, procedure code or a valid exception corresponding to pneumococcal vaccine, divided by all patients who have a recent ICD-9 charge code starting with 250." A CDS producer might then develop a form or alert that is triggered when a patient has such an ICD-9 code in the record and has not yet had one of the events in the numerator. That form or alert provides relevant information to the provider and might include a button for ordering the vaccine.
The EHR vendors that follow CDS developers should ensure that CDS interventions get integrated into the CDS components of their systems, so providers can easily access and use the clinical information. An important issue is how CDS producers will develop interventions that EHR systems can quickly and seamlessly accept, to respond to new guidelines.
The rise of the H1N1 virus, for example, has created a tidal wave of recommendations on patient screening and treatment. In an ideal scenario, CDS producers would be able to drop these interventions into EHR systems as quickly as they are formed and approved, so providers could make more-rapid, accurate treatment decisions for patients.
Even if quality organizations, guideline developers, CDS producers and EHR vendors do their job, providers and patients should still use CDS in everyday practice. Clinicians should ultimately view CDS as a partner and tool capable of enhancing their ability to care for patients. This is why CDS should fit into the clinical work flow, add ease, convenience and cost savings to practice operations, and improve patient care quality and satisfaction.
Clinician users of CDS can also benefit by generating reports that help them qualify for reimbursement and incentives from payers and employer programs. Next-generation standard CDS reporting can also provide feedback to CDS producers about how effective the intervention has been in real practice, as well as feedback to quality organizations and guideline developers about the value and impact of the guideline itself.
Clinicians' feedback on the quality and reliability of CDS systems will help producers modify and develop systems so they meet the needs of clinicians and patients. Meanwhile, reporting agencies will be able to develop feedback on individual guidelines, encouraging the kind of optimization that makes guidelines more practical and useable.
Several organizations and activities have already attempted to close the loop that connects clinical guideline development to practical use of CDS and measurement of both performance and effectiveness. The Roadmap for National Action on CDS, released by the American Medical Informatics Assn. as a work product for HHS, helped to galvanize the discussion around the promise and potential of standard methods for guideline translation and CDS sharing. Research projects at Duke University, the Partners Healthcare System-led CDS Consortium, Yale's GLIDES project and the Morningside Initiative are among the current efforts in knowledge translation and CDS standardization.
Also involved in the CDS movement is the National Quality Forum (NQF), which has recently developed its own information technology division. The NQF has commissioned the Health IT expert panel to establish a consistent data set that can be used for a wide variety of guidelines. Having such a well-known and limited data set would allow EHR vendors and others in the loop to standardize the meaning of their data elements more easily.
The NQF has also spearheaded the formation of the National Priorities Partnership. Working to improve quality through the elimination of "harm, waste and disparities," the partnership identifies quality guidelines most in need of immediate attention by the healthcare industry.
Among the key questions to be considered: How can CDS interventions be put into a format that would facilitate rapid integration by varied EHR systems? How can one design a repository, or a standard format that can be used by many repositories or services, that can store guidelines and effective CDS interventions, so that providers could select individual interventions or guideline packages to download into their EHR systems? How will these recommendations be triaged, vetted and authorized so that providers can have trust in the interventions they download? All of these questions must be answered to get to a place where providers can make use of each others' experience, rather than each provider re-inventing the wheel for each guideline.
In the end, publishing and incorporating a new treatment innovation into an EHR should be as easy as incorporating a newly invented drug into a pharmacy. Ideally, the more than 5,000 hospitals in the United States would no longer be burdened by each having to develop its own CDS elements. Instead, hospitals would leverage information from a pool of past experience, both in guideline-to-CDS translation and in implementing new, effective CDS interventions. A collaborative effort involving quality organizations, guideline developers, CDS producers, EHR vendors, providers, patients and reporting agencies could go a long way toward putting improved quality and safety within reach of every healthcare provider.
Jonathan Teich, M.D., is chief medical informatics officer for Elsevier. He is a practicing emergency physician and an assistant professor of medicine at Harvard.
For more information about Elsevier, click here.