Implementation best practices: Launching clinical decision support
Healthcare CIOs and their IT staff members can be left to their own devices when it comes to installing certain types of systems, but not clinical decision support systems.
Implementation of CDS demands great rigor and teamwork with clinicians and others within the provider organization. And it requires integration with other types of systems, usually including the all-important electronic health record.
Here, three experts in clinical decision support technology offer their implementation best practices for these systems, giving healthcare CIOs some tips to help them achieve success with decision support technology.
Determine the burden
Kody Hansen, a research director at KLAS, often hears from chief medical information officers that they do not adequately address implementation burden with their vendor partners up front.
“In retrospect, they often realize that significant internal IT resources, and or several informatics/clinical FTEs are required – especially when implementing clinical decision support functionality from EHR vendors,” Hansen said. “You’ll want to assess how much clinical expertise your vendor can bring to the table. One simple question posed during the procurement process can potentially save you hundreds of hours in the long run: What level of customization does this vendor allow?”
A healthcare CIO needs to understand the role vendors have played with previous customers in developing content and customizing algorithms, he added. Offering a highly customizable or configurable system is not necessarily a bad thing, but do vendors demonstrate strong commitments to certain paradigms across clinical use-cases that are central to one’s focus, Hansen asked.
“Some vendors tout they’ll put very little burden on your IT shop, while others employ full teams of clinicians and former users to support new implementations,” Hansen said. “You’ll want to share with a potential vendor partner the limits of your resources for this project – for instance, in terms of measurable IT hours, clinician expertise and dedicated FTEs – to determine whether they can counterbalance your constraints.”
Patient records are key to clinical decision support technology. In fact, the patient record is the “atom” of these systems, said W. Edward Reynolds, chief technology officer and executive vice president at Pepid, a CDS vendor.
“While not technical, the structure and specifications that define the patient record are critical,” Reynolds explained. “Demand a standard and make sure your vendors all use the same standard or standards as related to technical data specs. HL7 for example. Choose your vendors wisely.”
The ability to move and share and be data-agnostic is important to the organization’s future, he added. This allows one to move from one system to another or bring in new apps, processes and concepts all the while leveraging the data and data structures one has in place today, he said.
Another best practice is full integration, Reynolds said.
"A beneficial example of real-time clinical decision support is the usage of pharmacy consults for patients that have complex discharge plans."
Bill Kotraba, Information Builders
“Full integration is needed to support the uninterrupted workflow of the clinicians and medical staff,” he stated. “Without this, your system will never be totally effective. It is completely understandable and absolutely acceptable that not all components of the clinical decision support landscape will be fully integrated at the onset.
“But an organization can achieve goals in the short term without full integration by using jump-out applications and linking out to other apps and information,” he added. “However, you must fully integrate at some point to be successful.”
Clinical decision support systems are complex, and they can bring a lot to the table. But it might be important for healthcare CIOs not to have an all or nothing approach when it comes to implementing these systems.
“Ask yourself, ‘What does good enough mean for our organization,’” said Hansen of KLAS. “Those with high expectations are often disappointed when the end product – functionality or integration – doesn’t measure up. Ultimately, we must choose how to focus the value of our clinical decision support systems. Is your organization willing to invest to make content, rules logic and integration top priorities? If not, you will likely not optimize to your satisfaction.”
However, it is important not to conceive clinical decision support as an all or nothing program, Hansen added; the difference between added value and added noise lies primarily in one’s organization’s philosophy.
Healthcare CIOs, Hansen said, will want to consider the following:
- Are you implementing clinical decision support out of necessity or out of a recognition of the long-term potential of clinical decision support? Meeting regulatory or certification mandates is a lower-level motivation, and many systems exist today to meet these needs (both custom and outsourced).
- If striving for the ideals of clinical decision support – reducing alert fatigue, subtle prompts that support end-user workflows, better surveillance for adverse events, promoting evidence-based decision making – is the goal, then you may find that you must allocate more resources to do so.
- Many philosophies exist for accomplishing and presenting decision support recommendations at the point of care. You’ll want to explore which vendor partners share your values and ambitions. This process should not be underestimated.
- Your investment should reflect the extent to which you’re convinced that clinical decision support will become increasingly more important in a value-based care environment.
- How else do you expect precision medicine, AI and consumerism will manifest in the patient/provider relationship?
Bill Kotraba, vice president, healthcare solutions and strategy, at Information Builders, a technology vendor that markets decision support, master patient index, integration, analytics and other offerings, said several factors are important to CDS implementation and efficacy – especially the availability of discreet and person-specific data, timeliness of the information and the ability to present at the point of care.
“In order to be really useful, the coordinated employment of three technologies in an intelligent manner should be considered – real time data management, predictive algorithms and point of care deployment,” he contended. “Some clinical decision support systems use one or two of these technologies, but few use all three in a synchronous fashion.”
When combined seamlessly, clinicians can reliably use clinical decision support at the time and place of care – for example, deployed display within an EHR or via mobile technology, he added.
“A beneficial example of real-time clinical decision support is the usage of pharmacy consults for patients that have complex discharge plans,” he said. “Hospitals do not staff enough pharmacists to consult for each discharge, but with clinical decision support that combines all three technology aspects, they can reliably make sure consults are given to the patients that need them – in a timely manner that does not slow the discharge process.”
Follow three objectives
Healthcare CIOs would be wise when implementing clinical decision support technology to follow the path of the three clinical decision support objectives, said Reynolds of Pepid.
“Right person/right user interface, right time/right systems, right information/right clinical decision support content,” Reynolds explained. “Technology is big. Big data is big. Remember your core objectives and your users and insure you are meeting their needs. Faster is not always better. Better is not always faster. It’s not always the technology, it’s the content or delivery of that content that is paramount.”
Delivering incomplete or inaccurate data quickly and perfectly is inadequate as compared to delivering the right content not so quickly, he added.
On another front, a key implementation best practice is to choose sustainable and scalable technology and content, he said.
“Clinical decision support, while not new, is changing,” Reynolds said. “There are ‘levels’ of clinical decision support that range from multiple-page, reference-like content to actual point of care decision support. AI has entered the field of play and will impact your future decisions as related. You will need to think outside the box a little when it comes to forecasting where you are going.”
First, he advised, align what one requires in the short term; next, forecast what one will require in the not-so-distant future.
“Next, become overly familiar with the ‘projected’ changing landscape,” he said. “This is the tricky part. Then project where you are going and how, five-plus years from now. Now, roll all this back into your roadmap. And remember, you have to apply current regulations and current health information standards across the entire landscape.”
Health IT implementation best practices
This 20-feature series examines in-depth what it takes to deploy today's most necessary technology and tools.