Sepsis is a potentially life-threatening condition that is notoriously difficult to diagnose. National and international sepsis awareness campaigns such as the surviving sepsis campaign (SSC) raise awareness of sepsis by promoting adherence to predefined, time-sensitive bundles for the management of sepsis in order to improve outcomes. However, the timely execution of these life-saving bundles is not always guaranteed.
Not surprisingly, companies like bioMérieux, Merck & Co., and Intensix are developing technologies to more easily make intensive care practitioners aware of the presence of sepsis, which needs timely and aggressive interventions — most notably antibiotics and fluid resuscitation — to ensure patient survival. The Centers for Medicare and Medicaid Services enforces strict quality control measures (SEP-1) that require hospitals in the US to comply with the adherence and documentation of 141 specific actions and variables upon the treatment of sepsis patients.
Even though many physicians dread electronic alert systems due to their notorious ability to generate irrelevant warnings — potentially concealing a life-threatening condition amid a sea of innocuous pieces of information — they are welcoming the development of electronic alert systems for sepsis to alleviate the increasing legal and monetary pressure associated with non-compliance to sepsis’ very complicated treatment algorithm.
On Feb. 24, bioMérieux, a French biotech-company specialized in in vitro diagnostics, received FDA clearance for its automated pro-calcitonin (PCT) assessment assay VIDAS BRAHMS to guide the initiation of antibiotic treatment in patients with high levels of PCT, which is commonly observed in patients with bacterial septicemia. While this tool has great promise to identify sepsis patients with high risk of mortality and is able to reduce the antibiotic burden by discontinuation of antibiotics if the infection is already cleared, GlobalData notes that the low sensitivity for this system, particularly in cases of fungal sepsis, will still require guidance by an experienced emergency care physician.
On Feb. 16, Merck & Co. announced the launch of its electronic sepsis treatment guidance system, ILUM Health Solutions. The validation of the ILUM system was performed in a general ward and surgical hospital setting at the East Jefferson General Hospital in the US, where the automated clinical decision support solution guided the implementation of the 3h- and 6h-SSC treatment bundles. The implementation of ILUM resulted in an overall increase in sepsis diagnosis rates by 129 sepsis and septic shock patients compared with the previous year. While these results are ambiguous in terms of different incidence rates during the study period, the guided decision support system did show a reduction in hospital stay (6.81 versus 7.11 days, p=0.05) and fewer patients in need of ventilator support (5 versus 154 patients) compared with the previous year.
On Feb. 19, Intensix, an Israeli biotech company that specializes in big data mining, presented positive results from its five-month sepsis screening algorithm. The electronic early sepsis warning alert uses a five-year history of International Classification of Diseases, Ninth Revision (ICD-9) and Tenth Revision (ICD-10) billing data for the machine-learning algorithm, and processes real-time patient data for real-time mapping. In the real-life application, Intensix’s algorithm was compared with physicians’ ability to detect sepsis and septic shock patients in 782 patients from the Mayo clinic in the US. The machine-learning algorithm showed results comparable to manual review and had a sensitivity of 90.5 percent (95 percent CI, 85.4–94.3 percent) and a specificity of 88.5 percent (95 percent CI, 85.7–91.0 percent). GlobalData anticipates that the sensitivity and specificity will further improve with input from sequential organ failure assessment data as suggested, with the implementation of the new sepsis disease definition—where sepsis is defined as organ dysfunction due to infection—in the machine-learning algorithm.
With the introduction of pleiotropic automated alert systems for sepsis, GlobalData is very optimistic about the future early diagnosis and treatment of sepsis and septic shock patients. However, there remains plenty of opportunity to improve the management of sepsis, as experts interviewed by GlobalData stated that current guidelines tend to describe which avenues not to pursue, rather than providing detailed treatment regimens such as choice of antibiotic, appropriate type of resuscitation fluid, and applied volume of fluids, as well as the use of vasopressors.