Children's Hospital of Philadelphia researchers created an algorithm that can mine through EHR data to pinpoint pediatric oncology patients that would be a good fit for potential clinical studies, according to a study published in Pediatric Blood and Cancer.
The research team analyzed EHR data from PEDSnet, a national research network for clinical pediatrics, collected between 2011 and 2016. Patient information was extracted from three hospitals: CHOP, Aurora-based Children's Hospital Colorado and Seattle Children's Hospital and included data on diagnoses, procedures, medications, lab results and specialty providers.
The researchers developed a set of clinical features from the EHR to help establish the search algorithm. The criteria the team included were if the patient had at least three visits to pediatric hematologist-oncologists, at least one leukemia or lymphoma diagnosis and at least three specialist visits, two diagnostic codes and at least two administrations of chemotherapy. The algorithm was then based off a final group of 1,825 study participants that matched all the required criteria.
After performing an analysis of the study group's full medical records, the algorithm displayed 100 percent sensitivity and 99 percent to 100 percent specificity in accurately identifying the patients as having pediatric leukemia or lymphoma.
While additional research and studies should be done to alter the algorithm to meet specific needs of the study, lead study author Charles Phillips, MD, a pediatric oncologist at CHOP, concluded that the "algorithm can accurately and efficiently narrow down the number of medical charts researchers need to review to identify a patient cohort for subsequent clinical studies," according to a news release.