Sepsis onset, mortality rates predicted by EHR information

When patients are admitted to a hospital for a procedure, the expectation is that only the condition in question will require treatment. Try as they might, however, physicians cannot guarantee that complications will not arise from either issues during surgery or even the patient's own immunological response to the treatment.

One of the most difficult-to-treat complications that arises from even routine procedures is sepsis. The National Institute of General Medical Sciences explained sepsis as a severe overreaction of the body's immune system to an infection. Instead of neutralizing the harmful bacteria, the agents released by the body may continue to attack the tissues surrounding the original site of infection. From there, serious inflammation can occur that could lead to a loss of circulation and eventual organ failure.

According to the NIGMS, about 750,000 people develop sepsis in U.S. hospitals every year. Moreover, anywhere from 28 to 50 percent of those with the condition die as a result. At its high water mark of 375,000, sepsis is responsible for more deaths per annum than those caused by prostate cancer, breast cancer and AIDS combined.

However, a recent study from researchers at the University of California, Davis and published in the Journal of the American Medical Informatics Association may provide physicians with a way to detect signs of sepsis in the body before it progresses to a fatal state. Through the use of patient information stored in electronic health record systems, the threat and mortality rates from sepsis may be drastically reduced.

Analyzing patient data
The UC Davis study was co-authored by Tim Albertson, chair of the Department of Internal Medicine, and Hien Nguyen, associate professor of internal medicine and medical director of EHR. Albertson and Nguyen reviewed the data stored in the EHRs of 741 patients who had developed some form of sepsis in 2010 during a stay at the UC Davis Medical Center. 

Their findings pointed to several key parameters – blood pressure, respiratory rate, body temperature and white blood cell count – that could repeatedly predict the early onset of sepsis in patients who showed elevated levels in these areas.

Also, the study found that increased mortality risks from sepsis could be predicted by keeping track of elevated respiratory rate, blood pressure and lactate levels. Lactate is an acid that organs deprived of circulation release into the bloodstream, and as sepsis progresses to a state where inflammation is causing lack of blood flow, lactate levels are key in determining the state of the condition.

Sepsis poses a unique problem for physicians because not only is it difficult to definitively identify, but the treatments for it include broad antibiotic medications that may have negative health effects. For a patient already in recovery, this may worsen his or her condition if sepsis is not present.

With Albertson and Nguyen's findings, however, physicians may be able to use EHR systems to more accurately predict the condition.

EHR use and sepsis diagnoses
Illias Tagkopoulos, an assistant professor of computer science at UC Davis and a senior author of the study, emphasized the need for an accurate and timely way to diagnose sepsis in all areas where medicine is practiced.

"Rather than using a 'gut-level' approach in an uncertain situation, physicians can instead use a decision-making tool that 'learns' from patient histories to identify health status and probable outcomes," Tagkopoulos said in a statement. "Another benefit of the sepsis predictor is that it is based on routine measures, so it can be used anywhere – on the battlefield or in a rural hospital in a third-world country."