Clinical OMICS

SEP-OCT 2018

Healthcare magazine for research scientists, labs, pathologists, hospitals, cancer centers, physicians and biopharma companies providing news articles, expert interviews and videos about molecular diagnostics in precision medicine

Issue link:

Contents of this Issue


Page 28 of 51 September/October 2018 Clinical OMICs 27 the data are stored within the electronic health record (EHR) can hinder access and utility. "One of the challenges for clinical genomics is the representation of the underlying data, which can be very diverse," said Frei- muth. "The way that clini- cal genomic test results are recorded and stored within the EHR has a dramatic impact on the ability of a decision support system to find that data, to process that data, and then to provide the clinician with recommen- dations that may impact a patient's care plan." "EHRs were not designed for genomic data, and most genomic test reports are not delivered in a way where the results of these tests are stored as discrete fields within the EHR," continued Freimuth. "That means the decision sup- port rule has a hard time finding and using that data in the context of a given clinical workflow." In response to this challenge, most EHRs are investing heavily in the development of additional modules that will better support genomics. Freimuth also noted that "our under- standing of genomics and our utilization of genomics in clinical practice is evolving so quickly that it is difficult for large sys- tems, like EHRs, to be able to keep up with where the field is and where it is going." Playing Matchmaker For some patients, like those with advanced stage cancer that hasn't responded to treatment, a clinical trial is their only treatment option. However, matching patients to the appropriate clin- ical trials can be complex and arduous for physicians and CDS systems too. "The website took a really big step to organize all of the tri- als," said Krevsky Elkin. "It's a really great resource, but it has its drawbacks." The information listed online for a clin- ical trial is unstructured, lacks standard- ized terminology, may be outdated, and can be updated or change enrollment status at any time. Altogether, this makes it difficult to bring up a comprehensive list of candidate tri- als for a given patient. Like trial statuses, patient statuses change, too. Patients progress through lines of therapy and stages of disease, and as they do they may become eligible or ineligible for a clinical trial based on exclusion and inclu- sion criteria. "It's always an appropriate option to enroll someone on a trial," said Louis Culot, business leader of Oncology Informatics at Philips Healthcare. "This [capability] needs to be integrated end-to-end in the platform because patients aren't static and trials aren't static." To provide truly individualized clinical trial recommen- dations, CDS tools need to pull in information from both and patient data from the EHR. Natural language processing, a form of artificial intelligence, can help computers make sense of free text data, such as that on and in the clinical notes of EHRs. "Natural language processing plays a large role in the enrichment of EHR-based data, both now and very likely for a long time in the future, because it allows researchers to process pre-text clinical notes that were made by providers and convert them into more discrete and codified represen- (continued on next page) Illumeo's Findings Presenter integrates patient in- formation and enables tailored output, changing how radiologists share clinical information. Robert Freimuth, assistant pro- fessor of biomedical informatics, Mayo Clinic

Articles in this issue

Links on this page

Archives of this issue

view archives of Clinical OMICS - SEP-OCT 2018