A team at the Lucile Packard Children’s Hospital has adopted a new approach to storing patients’ electronic medical records (EMR) called the Stanford Translational Research Integrated Database Environment (STRIDE), according to a perspective paper published in the New England Journal of Medicine.
This new method allows doctors to compare their current patients with previous patients with similar symptoms and factors, and helps to determine the best step to take when issuing a treatment, according to the team. Physicians are no longer solely reliant on past experiences or forced to sift through the hundreds of pages contained in a chart for a single patient, said Christopher Longhurst, chief medical information officer at the Children’s hospital and a co-author of the paper.
“Wouldn’t it be great if computers stored the data and helped you understand how to help other patients?” said Longhurst, who is also an architect of the Children’s hospital EMR system. He added that doctors are now able to help current patients by looking at other patients that have been admitted with similar symptoms.
Longhurst said that a current challenge for doctors is that published evidence is available for fewer than 20 percent of decisions made, which forces doctors to rely instead on their own experience and anecdotal advice from their peers. This is especially true in pediatrics because less information is published in medical research literature in that area, he added.
According to the paper, this new method of storing and accessing information was put to the test when a 13-year-old girl was admitted to the hospital with inflammation of both the kidneys and pancreas as a result of lupus — a rare condition that did not have any clinical trials or medical literature to offer up a best course of action.
Jennifer Frankovich, who led the medical team, said that because of the girl’s risk for thrombosis — a formation of a blood clot inside a blood vessel that obstructs the flow of blood — the team realized that the girl would require anticoagulation drugs to prevent blood clots. She added, however, that the team thought the use of drugs could lead to internal bleeding, and was unable to find studies pertaining to the girl’s situation and related risks.
The team then turned to STRIDE and the hospital’s electronic medical record to view a cohort of pediatric patients with the girl’s condition that had been treated by clinicians at the Children’s hospital between Oct. 2004 and July 2009, according to Frankovich. Of the 98 patients in the cohort, 10 developed thrombosis.
After a statistical analysis of the data — which took less than four hours — the team was able to determine that the risk for blood clots was much higher in patients with inflammation in the kidneys and pancreas, and decided to give the patient anticoagulants within 24 hours of admission, according to the paper.
“I envision that this is going to be the future of all medicine — that most institutions are going to use electronic records,” Frankovich said. “We are working toward making these records accessible for clinical research. It is a major change in how we process medicine.”
Frankovich also said that because a doctor cannot look at a patient’s chart without permission, the new platform is working to anonymize patient records by removing all personal information — such as names and birthdates — with the hope of increasing the number of records that can be referenced in patient care.
“However, [STRIDE] doesn’t supplement the need for medical trials,” she said. “Obtaining the information doesn’t tell you if a certain treatment is better or worse. You still have to interpret the information with caution.”
Longhurst also referenced Patients Like Me, a social networking health site founded in Silicon Valley, which enables patients with unusual diseases to log in and share their data and outcomes with other patients. He said systems like STRIDE and Patients Like Me help to generate a large warehouse of data and req anduire the active participation of patients, but allow everyone to better understand medical information.
“Shouldn’t data act like that?” he asked.