Skip to content Skip to navigation

Perception and Reality: Clinical Decision Support

September 25, 2013
| Reprints

Full disclosure: I am not a doctor.

Okay, with that truth bomb out of the way, I feel safe to go ahead and write about the subject of clinical decision support (CDS) openly.  I understand there is a considerable amount of skepticism and annoyance when it comes to patient safety alerts, alert fatigue, and clinical decision support systems within the electronic health record (EHR). The terms are intermittently linked and often it’s not in a positive light for the whole notion of CDS.

For instance, take this study from June issue of Pediatrics, which looked at the case of a young child being treated at Stanford University’s Lucile Packard Children’s Hospital. I won’t get into the details, but thanks in part to unnecessary clinical allergy overrides that came from alert fatigue, the child died (there were other complications).  The study’s authors said that it was dangerous to burden clinicians with unnecessary alerts and that many of the systems didn’t include up-to-the-minute clinical information

There have been other studies deriding CDS tools because they weren’t properly integrated into the clinical workflow or because certain systems weren’t properly customized to get rid of drug–drug interaction (DDI) alerts that were overridden 90 percent of the time. Heck, there was even a study, from researchers at the University of Missouri, which found that even patients said using CDS was a turnoff in a physician. All this leads me to believe that the association with alert fatigue has led many providers to roll their collective eyes at the notion of current CDS systems.

Quite simply, I don’t think idea of CDS is winning any popularity contests these days.

There is, however, one physician who is a big-time advocate for CDS, and his name is Jonathan Teich, M.D. Dr. Teich is the CMIO at Elsevier, Philadelphia-based healthcare information and software provider, and a practicing emergency physician at Brigham & Women’s Hospital in Boston.

Knowing his advocacy for CDS (he wrote a book on the subject), in a recent interview with Teich, I asked him about the CDS-related critiques, particularly when it comes to alert fatigue and integration into the clinical workflow. Here’s how he responded, almost verbatim, because I think it’s an important defense of an area that seems to faces misconceptions:

We need to realize that CDS involves far more than alerts.  CDS embraces many other methods, including intelligent data displays and intelligent knowledge delivery.  To that point, CDS can offer clinicians insight into the questions we tend to ask most:  What do I really need to know about this patient?  What do I need to do next?  Am I taking the best, most appropriate action right now? Over my last 500 patient encounters? 

These can be delivered in many workflow-friendly ways that help, rather than hinder, workflow; we spend a lot of time in the Improving Outcomes with CDS book describing just how to make that happen.  Providers can benefit from more insight into how different CDS methods compare in terms of information quality and completeness, speed of access, timeliness and usability.  Clinicians also need to be aware of how well these methods support them in making accurate, evidence-based diagnostic, treatment and prevention decisions, and how they help clinicians accomplish tasks, including data entry, review and assessment. 

In an answer to the next question, Dr. Teich talks about the potential CDS plays in “liberating” clinicians. He said CDS is “not ‘telling a physician how to practice,’ but making it easier for the physician to do the practice they already want to do.”

I recently read about a real-life example of where CDS was used in this way, as a warning system for missed doses of prescribed medicine. A team from the Universities of Leicester and Birmingham implemented an electronic system to trigger a series of emails if, for example, a patient missed more than two doses of antibiotics. If it continued, emails were sent upwards through the different layers of the hospital.  This system led to a 'substantial and sustained' reduction in rates of missed or delayed doses of medicines,’ reported the study’s authors.

This hospital isn’t the only one who has found the benefits of CDS to be critical in improving clinical outcomes. A KLAS study of 140 providers found that 79 percent say CDS surveillance systems have led to mostly positive clinical outcomes. It’s unclear which of these providers have customized their tools, an important element I think to CDS implementation. 

So despite the problems with CDS, problems that no doubt exist and have to be worked on at an organizational level, the reality is that it’s going nowhere. Inevitably, CDS will evolve forward to become a set of tools that physicians really need, can use, and embrace.  

As always, I want to hear your thoughts. Feel free to leave comments below or respond to me on Twitter by following me at @HCI_GPerna.



As a member of healthcare I.T. leadership I have been on the receiving end of many complaints from physicians about how healthcare I.T. is not making their jobs easier. My response has always been to agree, but to point out that it generally is not about making their job easier, it's about improving the quality of care and patient safety. Ease of use and workflow optimization *will* evolve in these products as vital functionality becomes common commodity and vendors are forced to search for new competitive advantages. The same sort of evolution will occur at the level of CDS. But it will occur most rapidly if CDS is broadly used. Plus in the meantime, as Gabriel has pointed out, tremendous benefit can be realized in spite of the negatives.

Exactly! You hit the nail on the head Danny. Thanks for the comments.


Hi Gabe,
Great to read Dr. Teich's comments about misconceptions. I am working on a story for the November-December issue that looks at some of the progress on shareable decision support knowledge that providers are starting to consume as a service. The complex standards and vocabulary work on those efforts are still under way, but the potential is great to reduce the cost of every health system having to reinvent the wheel to develop its own knowledge objects.

Thanks for the comments David. Agreed on the potential of the technology in really changing the landscape. In a few years, I think all of these criticisms of CDS will look a bit misguided.