Google came with Death Predicting AI system for healthcare

By Smita Kumari |Email | Jun 19, 2018 | 5292 Views

Google, one of the leading company to harness the benefit of artificial and related technologies, developed a machine algorithm that uses patient past health records and data from various digital source and predict patient's probable death time.

The doctors and medical practitioner in healthcare or medical hospitals were using and trying to leverage huge amount of electronic health record and other data sources in real time to save human life. The current method of analysing and mining data is costly, cumbersome and time consuming. But the google predictive AI system can help doctors to analyse patient's health and predict more accurately when a patient will die after the admission in hospital. It also displays sources to reach conclusion about patients' death.

Google has identified the potential of AI neural network to learn and improve itself automatically by using huge data. The fact is highlighted in the research paper published by Google in journal Nature which shows a women death case suffering from breast cancer. The report says that a woman admitted in hospital struggling at the last stage of breast cancer even fluid already flogging into her lungs, she is viewed by two doctors and had a radiological scan. The computer analysed the health and estimated 9.3 % chance that the patient would die during her stay. The same condition was analysed by google new AI system that read around 175639 data points and estimated 19.9%death risk and she passed away in a matter of days. This shows that the AI algorithm could forecast how long will patient stay in hospital.
This new AI system will include the information like traffic, weather and other local news that may have impact on patient's health outcome. Also google are using many data sources such as google android app which collects data about mental decline and physical movements to track biological signals. To make AI predictions more accurate, google has used data from different sources for digital medical records such as University of California, San Francisco and the university of Chicago and still working to secure more partners so that the AI predictive solutions can be used by every hospital.
Earlier the companies were using data without their permission, but the recent study shows that Google with its other hospital partners ae concerned tat the patients health record must be used with the permission of patients. This may become more challenging when the data spread to local hospitals and huge medical networks.
Google has experiment many use cases to identify right direction to leverage the benefits of AI and now their efforts are in the direction to improve health care outcomes. The next move of google will be to implement this predictive model in clinical setting that can forecasts symptoms and diseases causing a hope or alarm in patients' health with greater level of accuracy.

Source: HOB