The University of Pennsylvania is using big data to prevent cancer patients from needing to go to the emergency room by developing alerts that prompt physicians to intervene earlier and provide service at a clinic or at home.
Oncologists and data scientists are digging into a trove of patient data to generate a formula that will predict ER visits among oncology patients, according to a blog post from Penn Medicine. Data generated from lab tests, radiology visits and patient-reported symptoms could all play a role in identifying patients that are at risk for and ER visit. The provider has also partnered with Independence Blue Cross to access a more complete data set of ER visits throughout the community.
By generating a predictive algorithm, physicians can intervene before the ER visit occurs. At Penn, those patients can be siphoned to the organization’s new Oncology Evaluation Center, created specifically for this purpose.
One notable challenge: Some of the most useful information for cancer patients is housed within doctor’s notes, making it difficult to pull in uniform data points. However, even if an algorithm doesn’t predict an ER visit, it may yield other useful information and keep oncologists informed about a patient’s current health status.
RELATED: Predictive analytics fuel Johns Hopkins 'command center'
“We hope to be able to help patients avoid being admitted to the hospital,” Evans said. “But even in this case, when admission is required, it’s controlled," Tracey Evans, M.D., an associate professor of Clinical Medicine in Penn’s Abramson Cancer Center, said in the post. "This also keeps the patients with doctors they already know, which provides a more comfortable environment.”
Studies show the vast majority of patients utilize the ER for conditions that could be treated in less expensive settings, and some states have taken aim at ER super users for that reason. Researchers have previously used predictive analytics to prevent overcrowding in the ER and predict hospital readmissions.