Surgical infections have been a major cause of worry for doctors since time immemorial and though antibiotics have reduced the problem to a large extent, revolutionary artificial intelligence techniques have helped in reducing it by 74 percent. At University of Iowa Hospitals and Clinics doctors have been working for past three years to reduce surgical site infection leading to savings of nearly $1.2 million during this period. Artificial intelligence techniques are being used by hospitals and healthcare services providers like John Hopkins and Amazon Web Services for pain management and post-operative care.
The University of Iowa is working with Dash Analytics on a machine learning platform called Dash Analytics High Definition Care Platform (HDCP) that uses a proprietary design to provide valuable data and metrics to support decisions at critical treatment periods. The platform has helped to reduce risk of surgical infections, blood transfusions, brain failure and unrecognized sepsis.
To improve the success rate of surgeries it uses curated information from historical data to bring down the risk of infections. The model calculates infection risks and links it to preventions that the surgeon can take during suturing the surgery area to reduce infection.
The machine delivers information about surgery related infection risks along with recommendations to doctor and nurse through an interactive interface allowing the latter to record with the click if the surgeon actually followed the suggested recommendations. As only some patients require these interventions, selective use of this technology minimizes cost and risk of patients. These feats are possible only through artificial intelligence as integrating this kind of medical information may be possible for an experienced doctor but even they are not able to remember and apply appropriate treatment in all situations. Algorithms developed by machine learning can help to identify best treatment methods for individual patients consistently based on systematic assessment of past information.