A new artificial intelligence program developed by the Google researchers can predict a person’s death with up to 95% accuracy. This type of AI can help hospitals save money and apply their resources more effectively by increasing the number of lives which can be saved. Google’s hospital AI uses the same technology which is used in many other projects by the company. Google has trained its AI to evaluate patient conditions based on thousands of factors available in patient’s health records.
By using that information, the software can predict that a patient will die while in the hospital and how long the patient has to live. This information will give doctors and other hospital staff a lot of valuable information about the patient, and which patient needs the most attention and have more chances to make a successful recovery. During the trial run of this software at the two hospitals in the US, the Google AI was 95% and 93% accurate in its prediction. This is because of the number of variables which are used by Google’s AI. The AI analyzes more than 100,000 factors to make its predictions as compared to only a few dozen or less for other models.
Machine-learning technology plays a major role in making this analysis possible. The credit also goes to the recently introduced Fast Healthcare Interoperability Resources (FIHR) standard. It allowed hospitals and other healthcare providers to share patient data in an accessible format. Using this standard, researchers can include a lot of information and analyze borderline unreadable freehand notes from doctors as well.