Creating a knowledge infrastructure for the ‘learning health system’
They spoke with 30 healthcare organizations about how they keep up with updates such as CDC Zika guidelines in terms of CDS. Many said they don’t know how to implement a change management process, Muthu said. One community hospital with 500 providers and a chief medical information officer told them that although the hospital has had CDS capability for several years, it has only built six pieces of CDS since the EHR went live.
Muthu and Tobias looked at how to break up the elements in the CDS support lifecycle so that the public health groups can identify formal definitions and logic conditions, and the providers can identify workflow scenarios and convert that to their platform. In other words, the public health institution authors and maintains CDS knowledge artifacts and the provider integrates it into workflow.
PHRASE works as a clinical decision support integrator sitting between the public health agency and the provider organization’s electronic health record, Tobias said. The prototype was created using CHOP’s customized clinical decision support within its Epic Systems EHR and alerts from Centers of Disease Control and the Philadelphia Department of Public Health.
As the clinical encounter is going on, the PHRASE engine sits between the EHR and the public health data repository in the cloud. Using FHIR requests against the EHR data, it checks to see if any of the criteria that would spur a public health recommendation are met, and if so the CDS alert pops up inthe EHR.
Another CDS-related project is ramping up to try to speed the process of getting the findings from federally funded patient-centered outcomes research into point-of-care clinical decision support. The Patient-Centered Clinical Decision Support Learning Network is a 4-year project launched in April 2016 with funding from AHRQ to catalyze dissemination of patient-centered evidence and practices via clinical decision support to improve care and outcomes.
In a recent online meeting of the learning network, Chris Moesel, a principal computer science engineer at Mitre Corp. described work on an AHRQ-funded project called CDS Connect. Launched in September 2016, it is being designed as a repository of computable CDS artifacts based on evidence-based standards of care.
“We are embarking on a CDS repository to host and share CDS artifacts,” Moesel told the meeting. “We want it to be easy to navigate and search, and pleasant to use.” He said people can go to the repository, find the CDS they are looking for based on a guideline and keyword, understand what it is or does and what is required to implement it. Then they could download it and integrate it into their system.
The first artifacts the researchers are working on involve cholesterol management and statin use based on published guidelines. CDS Connect is getting ready to kick off a pilot of the approach with the Alliance of Chicago, which deploys cutting-edge health IT in safety net settings. The Alliance will pilot the CDS Connect artifacts using its GE Centricity EHR.
Scaling up a learning health system
Getting back to the Knowledge Grid concept being developed at the University of Michigan, Allen Flynn stressed that the concept supports CDS modules, but it extends beyond that. “CDS is one end point of how you can apply computable knowledge, but not the only one,” he said. “You can have behavioral apps for patients or analytics that run in the background to do population health scans and screens.”
I asked Flynn how that all tied into the learning health system concept. He said that researchers at Michigan believe that a learning health system has profound analytical capability and the capability to bring knowledge to practice. They are working on the infrastructure for knowledge to practice, recognizing that the infrastructure for data to knowledge is already very well on its way. “We believe this Knowledge Grid is needed to bring about a learning health system at scale,” he said.