Understanding the Context in the Clinical Text

Analysis of 'big-data' and a patient's clinical information to improve the quality of his medical treatment stands at the basis of the software program developed by the OpisoftCare Corporation. The company, that focuses on improving the capability of accurate real-time clinical decision making, was established in 2014 by three entrepreneurs: Gideon Israeli, Erez Peleg, and Amit May-Dan. "OpisoftCare's development is a system for analyzing all the patient's clinical data, both structured and unstructured, in real-time and for providing a warning regarding preventable dangerous clinical situations during his hospitalization, such as hospital-acquired infections or readmission and repeat surgery," explains the CEO, Amit May-Dan. "This development has been made possible by a grant we received from the R&D fund of the Israel Innovation Authority."

"A fighter pilot needs to process dozens, even hundreds, of pieces of information in real-time to decide his next move," May-Dan illustrates. "T succeed, he makes use of sophisticated systems that analyze the situation and assist him in making the correct decision. A doctor is in a similar situtaton to a fighter pilot, when in the hospital. Doctors examine thousands of pieces of data and need to make the optimal decision for their patients. The vast scope of data makes it difficult for the medical team to decide correctly in real-time. A comprehensive computerization process began five years ago throughout all hospitals in the United States called 'Unified Medical Record' – a single system that gathers information from all systems about all patients but does not assist in decision making. That is where OpisoftCare comes in. Just as the pilot's systems provide him with recommendations, so too OpisoftCare gathers data from all hospital systems in real-time, thereby enabling the clinician to make an informed decision about a particular patient and to warn of problems long before the medical team is aware of their existence."

For example, OpisoftCare's software can 'read' the clinical data and provide a warning, by crosschecking the structured and unstructured information in real-time, that one of the hospitalized patients is at risk of a particular infection. This data already exists, but is generally in textual form, because clinicians record visitation summaries and test results in writing. Naturally, this leads to a lack of uniformity – each hospital has a different database system which lacks other data components. Our technology, however, creates uniformity between all the hospitals and supplements the missing information. We are the only technology platform that enables the real-time creation of all necessary data for the different hospitals", emphasizes May-Dan.

"In practice, one of the innovative components that we have developed, with the assistance of the grant from the Israel Innovation Authority, is an engine that understands the context within a clinical text. For example, when we search for pneumonia, in most cases the medical text does not even mention the word 'lungs,' but rather, will provide terms such as 'Filtrate of the Upper Left Lobe.' In other cases, the information will be irrelevant, such as evidence of a previous lung infection. It is important to understand that the concept is not only language-biased but also attuned to culture. In South Carolina for example, 'fainting' is termed 'falling.' Our system understands these nuances and knows to analyze the clinical significance of these medical descriptions in real-time", he explains.

Clinical Text

Analyzing Risks and Complex Situations in Real-Time

"A large number of studies are continually conducted in the medical world, but hospitals' accessibility to this knowledge is very complicated and not always practical", mentions May-Dan. "Our technology enables the utilization of this vast knowledge from studies already conducted, in order to solve various medical problems, without the need to go back to the beginning. Our approach is that instead of bringing a research team to the hospital, our innovative technology give them access to all existing knowledge accumulated over time and also provides the knowledge that is lacking so the doctor can treat the patient."

"The first issue we dealt with was hospital-acquired infections. These infections are a serious and sometimes deadly problem. In the United States alone, approximately one and a half million hospitalized people acquire infections every year, and more than 150 thousand of them die from that same infection. The annual average financial cost to the hospitals in the United States stands at USD 6 million.

Our focus is on the American market, where hospitals there have an immense financial interest in collaborating with us because the average cost of a patient who acquires an infection in hospital can reach tens of thousands of dollars, and on average, that patient will stay ten days longer in the hospital, something for which the hospital receives no refund.

One example of a hospital-acquired infection is pneumonia for patients attached to respiratory machines. In this context, OpisoftCare analyzes the data in the intensive care unit and warns the doctors of patients at high risk to develop that kind of pneumonia, which patients are at low risk and when the risk level changes. The program also continuously informs the doctor which of the patients has already developed pneumonia and how quickly it is possible to wean off the respirator in order to lower the risk", says May-Dan.

OpisoftCare's software for reducing infections was first implemented at the Sheba Medical Center in Tel Hashomer and was subsequently installed at University Health Louisiana – a large academic hospital in the United States. At the beginning of the year, the company won a tender for the creation of a national system for the reduction of infections that will include all hospitals nationwide.

"Our plan is to continue development. Every three months we insert an additional content area into our program", May-Dan says. "We began with hospital-acquired infections and readmissions and today our program also analyzes other medical situations: repeat surgery, resistant bacteria, chronic diseases and other complex situations that require real-time analysis of vast data in order to make clinical decisions."

"Our business model runs in parallel the concept and the technological model. We have a subsidiary company in the United States that is working with a senior and experienced American team on the business development and sales to the American market. This company is establishing the infrastructure for operations in the American market. Furthermore, we are close to the completion of strategic processes to partner with the largest corporations in the United States, a step that will enable us rapid growth in that market", summarizes May-Dan.