Prediction of postoperative infections

Wound infections after surgery is a major problem which oftentimes necessitate additional invasive treatment, lowering the quality of life of patients, leading to increased costs. Early or even timely recognition of infections fails in the current system. The project team's vision is to disrupt the current post-operative work processes and protocols in hospitals by providing a unique software tool that uses machine learning (ML) algorithms to enable a shift from a diagnostic and responsive system towards a predictive and proactive system. This tool to predict post-surgery infections is based on electronic health records (EHR) data and will be accurate, transportable, robust, scalable, easy-to-use and integratable into existing hospital workflows and IT infrastructures. This can potentially significantly lower the number of infections, hospital readmissions and length of stay of a patient, leading to lower costs, a relieve on the workload of hospital staff and a higher quality of life for patients.

The project has developed an international connection and has contributed knowhow and specialized methodology to support PERISCOPE. In particular, the project has developed a software pilot and new methodology for predicting post-surgery infections only based on blood samples from the electronic health records. The results obtained thus far are very promising. A Disclosure of Invention (DOFI) is about to be filed based on these results.

Infections hit one in four patients, yearly more than 10 million in Europe, after their surgery. On average only after day 5 the the infection is diagnosed with a maximum accuracy of 70% (through biomarkers, vital signs, medication use, past medical history and pre-op risk score) and treatment started. Infections more than double hospital stay and oftentimes necessitate additional invasive treatment, lowering the patient’s quality of life. The cost of treating a post-operative infection is on average €10,000 per patient. Early or even timely recognition of infections fails in the current system. The project team's vision is to disrupt the current post-operative work processes and protocols in hospitals by providing a unique software tool that uses machine learning (ML) algorithms to enable a shift from a diagnostic and responsive system towards a predictive and proactive system. This tool to predict post-surgery infections is based on electronic health records (EHR) data and will be accurate, transportable, robust, scalable, easy-to-use and integratable into existing hospital workflows and IT infrastructures. This can potentially significantly lower the number of infections, hospital readmissions and length of stay (LoS) of a patient, leading to lower costs, a relieve on the workload of hospital staff and a higher quality of life for patients.