One reason health IT appears to be falling short on its promise of improving patient safety is that it's difficult to translate clinical guidelines into electronic form for use in clinical decision support systems. 

A study published in the journal Applied Clinical Informatics in July illustrates the problem. Researchers from Indiana University School of Medicine examined what one author calls the "authoritative guidelines" for pediatrics in the U.S., and found that just 21 percent of 245 specific care recommendations were "actionable," a prerequisite for computerization.

"Actionable recommendation statements are both decidable (i.e., every condition is described clearly enough so that reasonable practitioners would agree on the clinical circumstances for which the recommendation should be applied) and executable (i.e., the recommended action [what to do] is stated specifically and unambiguously)," according to the study.

"The guidelines are not written so they can unambiguously be written into decision support rules," its author, Stephen M. Downs, MD, director of children's health services research at IU School of Medicine, and former director of the IU-Regenstrief Institute Biomedical Informatics Research Training Program, tells Healthcare IT News.

"We all kind of knew this was a problem," Downs says. However, he adds, "the magnitude surprised me a little bit."

The researchers reviewed, "Bright Futures: Guidelines for Health Supervision of Infants, Children and Adolescents, 3rd edition," published by the American Academy of Pediatrics, to see how adaptable the guidelines were to EHRs. The first edition dates to 1994. "When they started, I don't think they had electronic records in mind," Downs says. 

AAP clearly does recognize the importance of EHRs now, but Downs says it will take time to update the Bright Futures guidelines to make them more compatible with clinical decision support – and with pediatricians' workflows. 

"The current version of Bright Futures is not organized to easily implement this capability in a computer," the study argues. "Instead, recommendations are listed according to what should happen at each of the 21 predefined health supervision visits. A problem with this format is that it assumes that the child will be seen for health supervision at each age and that previous visits have been completed."

The fact that guidelines are organized by visit means there is a lot of repetition of recommended interventions and patient/parent instructions, and the researchers only found 245 discrete recommendations out of more than 2,100 guidelines. "Substantial modifications to Bright Futures will be required before it can be effectively implemented in computerized clinical decision support," they wrote.

To fit with CDS, guidelines must meet three basic criteria, according to Downs. They must contain solid medical evidence, must consider benefits vs. harms of each intervention for a given patient and they must be stated clearly. "If you can make it clear enough for a computer to understand, it's also easier for a human to understand," Downs says.

As a result of the research, Downs is now under contract to develop an electronic version of Bright Futures for AAP. He says he is working to set Bright Futures-based CDS as a Web service that first will be tested and interfaced with a Cerner EHR at Indiana University Health. Downs hopes to have a functioning test in IU clinics in the first half of 2015.

"If it works well, the obvious next step is to see if it can work with other vendors' EHRs," Downs says.