The hospital – at least the intensive care section of it – is a dangerous place. As many as 30% of the people who go in do not come out, and half of those that do emerge with a slew of problems that require additional treatment – high blood pressure, ambulatory issues, digestive problems, and more, often requiring extensive and expensive time in rehab.

Intensive care patients are especially vulnerable to problems such as kidney and liver failure, cardiac arrest, shock, organ failure, and — perhaps the greatest danger of all — sepsis, which is one of the leading causes of death of ICU patients, both while they are in the hospital and when they get out.

All this, despite the large amount of resources expended on the ICU. In a typical hospital, the intensive care unit is usually the hospital’s smallest – making up about 9% of its beds – but taking up 31% of a hospital’s budget, according to Gal Salomon, co-founder of Israeli big data medical research company Intensix. “It also has the highest ratio of doctors and nurses to patients – as many as one of each for each two or three patients.”

Why would this be? With all those resources, it would be expected that the ICU should have the best patient results in a hospital, not the worst.

The truth is, said Salomon, doctors and researchers have no idea why the ICU is the most dangerous place in the hospital. But a study by Intensix that has been going on for the past eight years sheds some light on what the causes of what those dangers might be.

Even more important, a new study that the company just initiated will utilize its database of over 600 billion pieces of patient data that it gathered in its first study to analyze the conditions of patients in real time, providing alerts to doctors and nurses of impending medical problem like sepsis, allowing them to intervene before the problem gets out of control.

Gal Salomon (Courtesy)

Gal Salomon (Courtesy)

Fever is a good example of such problems, said Salomon. “In septic patients – where harmful bacteria manages to get into the body, causing a severe infection – fever rises and blood pressure drops. But that could be due to any number of causes. It’s only much later on that it becomes clear that sepsis is the cause – and by then it is often too late to do anything, because at the later stages, deterioration can be rapid.”

“Our system correlates data from patients and tries to understand the factors at work – so using our example, we hope to determine which factors are at play that have led to septic conditions in patients, so that doctors can intervene and use the proper protocol early on,” said Salomon.

The data being used by Intensix for its correlative study comes from the medical data collected on some 8,000 patients at the Sourasky (Ichilov) Medical Center in Tel Aviv, as well as several American hospitals, between 2007 and 2014.

The anonymized data was cleaned, filtered and analyzed using advanced mathematical algorithms to create models for early predictions of deterioration and/or complications for patients in the ICU. “Accurate models for predicting the clinical outcomes will enable the medical staff to improve care and provide tailored treatment to each patient based on their individual needs,” said Salomon. “This capability is absolutely vital in intensive care units. Early detection, even within hours of possible complications, can actually save the patient from death and significantly reduce morbidity that could influence the clinical outcome of the patient long after they are discharged from the ICU.”

In the coming weeks, the system installed at the hospital by Intensix will commence its second phase of prospective study. With the models developed based on the original data, the system, using bedside devices, sensors, and electronic health records, will keep a constant eye on a patients’ vital signs, environmental elements, medical condition, and other factors to determine if they need extra help – and what kind of help they need.

At this point, said Salomon, the system cannot isolate specific causes for sepsis and other conditions. He believes, however, that the correlative aspect of the system will significantly lower the casualty rate in ICUs. “Eventually, however, we believe we will be able to narrow down the factors that cause specific conditions to arise in patients,” he added.

For hospitals, the system is a true boon, said Prof. Idit Matot, Head of the Anesthesia, Pain and Intensive Care Division at the Tel Aviv Sourasky Medical Center.

“Patients in intensive care units are the most complex, and therefore they are the most actively monitored in the hospital. A system that recognizes the development of complications or deterioration, hours before they occur, will enable effective assessments and early treatment that will actually reduce mortality and morbidity,” Prof. Matot said.

According to Izik Itzhakov, VP Business Development at Intensix, “More than 5 million patients are hospitalized in intensive care units in the United States annually, and up to approximately 30% of these patients will not survive. Early diagnosis enables the improvement of quality care, the increase in efficiency that could translate to reduction in morbidity and mortality, the earlier release of patients, and most significantly, the minimization of long-term damages that may affect patients long after they are discharged.”