News Stay informed about the latest enterprise technology news and product updates.

How clinical decision support can improve medication management

Clinical decision support can prevent errors, particularly in dispensing medication. At a recent conference, three early adopters shared what works now, and looked to future uses.

BURLINGTON, Mass. -- While clinical decision support might be one requirement for meaningful use, the technology that will simplify the clinical data analysis CDS entails seems to be a few years off. Right now, hospitals are getting the most bang for the CDS technology buck -- and preventing errors in patient care -- by applying it to their medication management processes.

The lowest-hanging CDS fruit involves software that helps physicians verify that they're prescribing the right meds, three panelists agreed at the Massachusetts Health Data Consortium's HIT11 conference. Downstream, these tools also can help nurses confirm that the correct medications make it to the bedside and that patients are taking them at the right time. As interoperability between electronic health record (EHR) systems and other clinical applications improves, however, and as hospitals adopt more consistent data standards, other facets of health care could benefit from clinical decision support as well.

The panelists -- all physicians -- came from different vantage points:

  • Marylou Buyse, chief medical officer at Scott & White Health Plan, a payer serving 32 Texas counties.
  • Michael Lee, Dedham, Mass.-based pediatrician and director of informatics at Atrius Health, an 800-physician ambulatory group.
  • Dan Nigrin, senior vice president and CIO at Children's Hospital in Boston.

Let committee decide whether to turn off medication error alerts

Current e-prescribing modules in clinical decision support tools can signal major potential interactions between drugs a patient is taking, track a patient's drug allergies and warn physicians when they're about to prescribe a drug that will cause a reaction. The problem is they also overwhelm physicians with lesser warnings and alerts -- such as the risk of potential food interactions or drug ingredient warnings -- to the point where physicians ignore most or all of them. Yet for fear of liability issues, "no one wants to turn them off," Lee said.

Atrius Health has worked hard on controlling the proliferation of medication error alerts so that physicians get the most relevant ones, Lee said. The project started with polling physicians about which alerts were, in his words, "driving them crazy." The ambulatory physician group left some on, shut others off and shared the wealth by offloading still other alerts to staffers when appropriate.

"It's not easy [to turn off an alert]. … For really annoying alerts, we have a clinical panel, including the pharmacy, look at the research and debate if there's enough danger to leave it on," Lee said.

Children's Hospital has a similar process, one that takes into consideration an analysis of which alerts physicians ignore, Nigrin said.

Bar codes support care decisions, prevent medication errors

Downstream from the physician, nurses also can get into the act of using CDS systems at the bedside to prevent medication errors. Children's Hospital built a clinical decision support tool it calls PhedEx to track medications from the time they are prescribed to when the patient takes them, Nigrin said.

Driven by bar codes, which are scanned at various points along the way (when a medication is ordered, filled and delivered to the bedside, for example), PhedEx enhances adherence to medication prescriptions and helps detect errors along way, making sure the right meds get to the right patients at the right time.

In the eight months since the system went online, it has prevented 4,500 potentially significant (by Children's Hospital's own definition) medication errors, reducing them by half. Implementing computerized physician order entry and its automated drug reference tools did have some effect on reducing medication errors, but it was CPOE in conjunction with PhedEx that created a sustained decrease, Nigrin said.

"To us, that's a pretty significant and meaningful reduction," Nigrin said. "We know the patients feel the same way."

The automated system also solved a longstanding administrative problem, Nigrin said: patients and nurses calling the pharmacy to check on the progress of a prescription, and the pharmacists having to drop everything to determine the answer.

Data warehouse mining provides new clinical decision support tools

Like many large health care organizations, Atrius, Children's Hospital and Scott & White use a clinical data warehouse to store patient information. Mining the warehouse can offer a lot of insight about the patient population an organization serves, but the process comes with caveats.

Almost 50% of doctors now have EMRs. If only 50% of banks had ATMs, would you be a little bit concerned? Who'd go to a bank without an ATM?
Marylou Buysechief medical officer, Scott & White Health Plan

First, it's important to establish the relevancy of data mining projects, Lee said. It's just as important to understand how to pose queries, because, as in all computer systems, "crap in gets crap out."

Nigrin agreed, adding that clinical decision support projects involving data mining should start with determining the most relevant questions that need to be answered, and at which points in the clinical workflow they should be answered. The best way to do that? Polling the physicians who will be using the system.

More on this topic

  • How meaningful use could drive clinical data analytics adoption
  • Radiology on cutting edge of clinical decision support

What's typically the problem with data mining is that it takes an advanced data analyst to write the queries, Scott & White's Buyse said. That person usually gets busy once physicians and nurses start to see the value of getting answers to their very specific questions about patient outcomes.

The longer the hang time between when a query is made and data is returned -- especially when new ideas are being tested that require several rounds of updates from the data warehouse -- the less effective clinical decision support becomes.

The solution is to provide clinicians with simple-enough tools that they can post queries themselves and get immediate answers, Buyse said.

Future of clinical decision support improving with mass EHR adoption

Buyse divided the evolution of CDS tools into five stages -- first, as collector of reference information, then as documenter of decisions, physician helper, colleague, and finally, mentor. Right now most of health care is somewhere around the second stage, she figures.

For clinical decision support technology to hit the mentor stage, where it can teach physicians how to provide better care or at least provide real-time feedback at the point of care, Buyse offered the following roadmap:

  • Lab, radiology and EHR systems have to be more interoperable.
  • Health care providers and patients need transparent pricing data so they can make more informed choices.
  • Data security has to become smarter.
  • EHR vendors have to reduce prices to achieve true market saturation.
  • To facilitate data exchange, EHRs and personal health records, or PHRs, should be merged into single systems and connected across ambulatory and inpatient settings.
  • Data standards must evolve so all practitioners are speaking the same language and exchanging standard files that each other can read.

If all these things happen, it would open up clinical decision support for robust tools that can support much more than medication management, Buyse said.

"[Health] IT systems are 10 or 20 years behind other industries in standardizing, refining, even adopting it," Buyse said. "Almost 50% of doctors now have [electronic medical records]. If you think about it, if only 50% of banks had ATMs, would you be a little bit concerned? Wouldn't the banking industry sort of collapse? Who'd go to a bank without an ATM?"

Let us know what you think about the story; email Don Fluckinger, Features Writer.

Dig Deeper on Clinical decision support systems