Pushing CDS Capabilities Ahead

Clinical decision support (CDS) systems hold great promise in improving patient care and safety but only if organizations refine the CDS they are using, and clinicians then heed the suggestions the systems offer.

Studies on patient safety show that there is a great deal of room for improvement. For example, for every 1,000 patients coming in for outpatient care, there appear to be approximately 14 patients with life-threatening or serious adverse drug events. For every 1,000 prescriptions written, 40 have medical errors and for every 1,000 women with a marginally abnormal mammogram, 360 will not receive appropriate follow-up care.

Getting the Alert Level Right

While a number of strategies may be helpful in improving safety, CDS systems have a great deal of potential for preventing these types of errors in particular. One study1 I worked on looked at providers’ overrides of computerized drug-drug interaction (DDI) alerts in primary care. We found—as have others—that too many alerts result in providers ignoring many alerts, often to the extent that they may overlook even the clinically important ones.

Of almost 25,000 DDI alerts over the study period, 40 percent were accepted. Overall, 68.2 percent of the DDI alert overrides were considered appropriate. Among inappropriate overrides, the therapeutic combinations put patients at increased risk of several specific conditions. A small number of drugs and DDIs accounted for a disproportionate share of alert overrides.

Of the 121 appropriate alert overrides where the provider indicated they would “monitor as recommended,” a detailed chart review revealed that only 35.5 percent actually did. Providers sometimes reported that patients had already taken interacting medications together, despite no evidence to confirm this.

The most common coded reasons for overriding DDI alerts were “will monitor as recommended” (43.9 percent), “will adjust dose as recommended” (16.9 percent), and “patient has already tolerated combination” (15.7 percent). Providers chose the coded reason “other” in 19.7 percent of alert overrides, providing for reasons that the drug had been recommended by another healthcare provider, the patient already had been counseled not to take both drugs together or the patient was no longer taking one of the drugs listed as a potential cause of the interaction.

CPOE challenges

In another study2, we used a qualitative approach to observe clinical activities and capture the experiences of physicians, nurses, pharmacists and administrators at five community hospitals in Massachusetts that adopted computerized physician order entry (CPOE) in the past few years.

Analysis of observations and interviews resulted in findings about the CPOE implementation process in five domains: governance, preparation, support, perceptions and consequences. Successful institutions implemented clear organizational decision-making mechanisms that involved clinicians (governance). They anticipated the need for education and training of a wide range of users (preparation). These hospitals deployed ample human resources for live, in-person training and support during implementation. Successful implementation hinged on the ability of clinical leaders to address and manage perceptions and the fear of change. Implementation proceeded smoothly when institutions identified and anticipated the consequences of the change.

Enormous opportunity

Most EHR benefits come from CDS but, to date, most implementations are very limited. What organizations have implemented is a mere shadow of what it could be. There is an enormous opportunity for further improving care if organizations can put in place more CDS.

One infrastructure component key to optimizing CDS is an event monitor program which sits over a database. Ours at Brigham & Women’s goes through all data, and is linked with a rule editor which is simple enough that any clinician can write a rule, though governance is held centrally. Many vendor applications include similar infrastructure.

Most organizations could do a much better job of making information available at the time that providers really need it. Clinicians need certain reference information in a specific context, often in a very limited period of time, and tools like infobuttons make that feasible.

I also think we’re going to see a lot of decision support going to patients in the near future. The evidence shows that patients who are more engaged in their care take better care of themselves and use fewer resources. I think all organizations should be thinking about how they can deliver robust, unbiased information to their patients. That will help everybody.

CDS also will be more and more important in population health management efforts. Most of the decision support we’ve had so far has been very focused on the individual patient. From an accountable care perspective, organizations will absolutely have to have good population-level tools to be successful.

I also see much more forward-thinking decision support with more complex rules that take into account many patient characteristics. Leveraging analytics will be one of the most important things going forward. Eventually, we should be able to be more directive in certain situations. We’d like CDS systems to be thinking along with the provider, making suggestions in real time that are context-specific and finding things of interest and value to providers and patients.

EHRs, analytics, databases and our increasing experience with these tools are potentially transformative in the aggregate—we have to begin to realize this promise.

David W. Bates, MD, MSc, is senior vice president for quality and safety at Brigham and Women’s Hospital in Boston.

1. Slight SP, Seger DL, Nanji KC, Cho I, Maniam N, Dykes PC, Bates DW; Are We Heeding the Warning Signs? Examining Providers’ Overrides of Computerized Drug-Drug Interaction Alerts in Primary Care. PLoS One, Dec 26, 2013.
2. Simon SR, Keohane CA, Amato M, Coffey M, Cadet B, Zimlichman E, Bates DW; Lessons learned from implementation of computerized provider order entry in 5 community hospitals: a qualitative study. BMC Med Inform Decis Mak, June 24, 2013.