HCLF: The evolution, future of decision support

CHICAGO—Healthcare is among the most information intensive fields and most physicians acknowledge that some decision support would be useful to them, said Edward Shortliffe, MD, PhD, professor of biomedical informatics and senior advisor to the executive vice provost for health solutions at Arizona State University, speaking at the Healthcare Leadership Forum on Nov. 14.

The first published recognition of the role of statistics in medicine occurred in the 1950s. Since the 1980s, there has been explosive interest artificial intelligence, Shortliffe said.

There is tremendous variation in practice, even among the so-called experts, he said. “We need to understand better how experts use personal heuristics—their experience, data and knowledge to arrive at decisions. The better we understand, the better we can teach and make it an explicit part of education and training and improve decision support systems to make them more useful to clinicians.”

Patient safety and error reduction are major motivators for the use of decision support and there is increasing incorporation of decision supports functions in commercial products, he said. It’s not enough, however, to use EHRs simply to provide information back to clinicians caring for a patient. “There’s a bigger cycle that needs supporting” which includes research pertinent to the diseases patients have, public health, disease registries and biomedical and clinical research. “That’s been the dream for a long time but with paper records there was never a way to integrate what’s happening in the clinical setting.”

The key challenges facing the success of decision support include identifying the context-specific information needs of clinicians; modeling patients and the care process; integration of systems—interoperability is a big barrier to doing this well, he said; terminology translation; and user education about the strengths and limitations of these types of tools.

Integration with routine workflow is critical to increased use of decision support, Shortliffe said. Transparency helps to assure acceptance and the internet is great facilitator of integration. That does not avoid, however, the need for standardized terms and data-sharing protocols. “We’re experiencing new challenges in providing locally-derived decision support through collaborations with vendors of EHR systems.” The healthcare industry, he said, must also take on the responsibility of making decision support part of the culture and the way people are trained.

The Healthcare Leadership Forum was sponsored by ClinicalKey and presented by Clinical Innovation + Technology.