As a first step toward helping clinicians and patients understand genetic health information, electronic health records should at a minimum have a way to represent or ingest family history data. But speakers at the July 21 HL7 Genomics Policy Conference in Washington, D.C., said a huge chasm exists between the value of family pedigrees and how widely they are used, and they put at least some of the blame on the capabilities of EHRs.
Howard Levy, M.D., Ph.D., assistant professor of medicine at the Johns Hopkins School of Medicine, said, “It is a tragedy that in 2015 we are talking about how poorly EHR vendors are doing with family history.” He added that he was frustrated with the current version of Epic that Johns Hopkins is using because it doesn’t support representations of pedigree and how people are related. “We need EHR vendors to do better,” he added.
Meg Doerr, principal scientist with the Seattle-based Sage Bionetworks, said studies have shown that the way family history is currently collected by providers leads to data that is not actionable. It is unstructured data and therefore unavailable to clinical decision support tools. She said there are 20 different stand-alone family history software tools available, such as Hughes RiskApps or Duke’s MeTree, 18 of which store the patient data in separate data warehouses rather than in an EHR. Sixteen of those 18 are ready to pull and push data to EHRs, she said, but none has done it yet. “Nobody is doing it because of the lack of capability on the EHR side,” she said. It is difficult to find a place to put that structured family history in the EHR.
Doerr, who previously worked at the Genomics Medicine Institute at the Cleveland Clinic, added that EHRs could also be a source of much-needed detailed data to populate the family histories. “We want to make sure data transfer is as frictionless as possible,” she said, but when pulling data out of a health system’s EHR or clinical data warehouse, there is often a tug of war about who owns that data. “If we are using it to care for patients, maybe we are all on the same team,” she said.
Another challenge with having patients enter family history information is that there is no standardized lexicon that is patient-centered and also maps to the established medical lexicon, Doerr said. “That is a huge gap and critically important in gathering data from patients.”
She also noted that only a handful of family history tools have patient permissions that clearly spell out ‘I am giving data and I understand how it will be used and I approve.’ Most patients expect their data to be used for clinical care and to drive discovery and cure diseases, Doerr said, “but we don’t ask them, though.” She said legislation would likely change the rules soon, and require asking for permission about the ways health systems collect and use that data. “We can’t hide,” she said. “Patients want to share data. We need to create the structures to make it happen.”
Joy Larsen Haidle, president of the National Society of Genetic Counselors, described the process she goes through doing a cancer risk assessment with a patient. First, in the face-to-face meeting with the patient, she draws out the family tree by hand. Then she has to enter the data into a program that calculates the likelihood of cancer, then re-enter it into an EHR family history table, then summarize it in a consult note. “The lack of electronic tools that speak to each other creates this redundancy,” she said. It increases the time she spends on each patient instead of allowing her to see more people. “We are asking for it to become a priority in the EHR,” Larsen Haidle said. “Integration is critical if we are going to impact health. These silos make it difficult to get the optimal outcome.”