Blog

Blog: AI is getting smarter and creepier, and it can even predict when a person will die

http://bit.ly/2R99OI8


AI algorithms are proving to be an effective solution in predicting death.

There’s been an explosion of breakthroughs in the field of artificial intelligence (AI) over the past few years. The evidence for this can be found in almost every industry, and there’s no doubt that the rise of AI will continue to disrupt existing sectors in the future. AI has already proved its usefulness in automating dull and mundane tasks in industries such as retail, finance, and even construction. However, AI algorithms and software could so much more. In fact, such as system could be capable of predicting premature death. This may seem frightening, but for the healthcare sector, it could do wonders.

University of Nottingham’s research

Researchers have already made multiple discoveries in this field of AI. For instance, a team from the University of Nottingham created an AI-based system capable of predicting premature death. Thanks to machine learning algorithms, the system performed better than standard prediction models led by human experts. The system used data from over half a million people in the UK, between the ages of 40 and 69. Besides health information, it also analyzed people’s demographic data and lifestyle choices. It even assessed their daily fruit, vegetable, and meat consumption to predict the risk of early death.

This AI system could help in the fight against serious diseases and play a major role in preventative healthcare. According to one of the researchers, Professor Joe Kai, “There is currently intense interest in the potential to use AI or machine-learning to better predict health outcomes. In some situations we may find it helps, in others it may not. In this particular case, we have shown that with careful tuning, these algorithms can usefully improve prediction.”

Capabilities of Google’s AI technology

Another solution is Google’s predictive algorithm that’s been tested among hospital patients to predict early death, discharge, and hospital readmission, and it showed higher accuracy than traditional clinical predictive systems and methods.

As part of the study, Google applied its algorithm to a patient who suffered from metastatic breast cancer. The algorithm gave the patient a 19.9% chance of dying in the hospital, while the hospital’s augmented Early Warning Score provided a 9.3% estimate. To come up with the prediction, Google’s AI analyzed 175,639 data points from the patient’s health records, including her vitals and medical history. In less than 2 weeks, that same patient died, and Google’s tech outperformed the hospital’s traditional prediction model. During the entire study, the AI system analyzed over 46 billion data points from 114,003 patients. An innovation such as this one could help hospitals prioritize patient care and adjust treatment plans before medical emergencies even occur.

LogitBoost machine learning algorithm

The European Cardiology Society conducted a similar study. Presented at the 2019 International Conference on Nuclear Cardiology and Cardiac CT, the study found that AI could serve as an efficient method for predicting heart attacks and death. The researchers used a machine learning algorithm called LogitBoost to analyze cardiac imaging data from 950 patients from the Turku PET Center in Finland. All patients had chest pain and had undergone various cardiac scans over the course of six years. The team involved in the study claims that LogitBoost showed 90% accuracy when it comes to detecting a patient’s risk of suffering a heart attack or premature death in the future. Since doctors and other healthcare professionals already collect a lot of patient information, integrating such data with machine learning could accurately predict patient outcomes and risks and help doctors personalize treatments.

Gero’s AI app

Though LogitBoost could simplify the way doctors predict patient outcomes, relying on an AI-based algorithm that harnesses “the activity tracking from smartphones and smartwatches to estimate your lifespan” could be a more practical solution. For this reason, the Russian startup Gero teamed up with scientists from the Moscow Institute of Physics and Technology (MIPT) to train an algorithm to determine life expectancy.

The algorithm analyzes people’s step counts, sleeping habits, and how often they switch between active and inactive periods. The company even developed a beta version of a mobile app that predicts the user’s lifespan. The technology, however, doesn’t give a full picture of one’s health, because it doesn’t use data such as genetics or diet. Once combined with clinical analysis, the app would give a more accurate estimate. And despite it not being ready for medical application, the app shows potential, as it could contribute to the development of more effective anti-aging treatments, as well as mitigate risks in insurance and pension planning.

Final thoughts

More and more industries today are being influenced by AI and machine learning, and healthcare is no exception. In fact, it’s one of the industries that could benefit the most from AI, which will undoubtedly transform the future of healthcare. Thanks to smart algorithms, doctors would be able to accurately predict patient outcomes and even the risk of premature death. Such valuable information would help healthcare professionals to better personalize treatments and improve care delivery before it’s too late.

Author: Richard van Hooijdonk

International keynote speaker, trendwatcher and futurist Richard van Hooijdonk offers inspiring lectures on how technology impacts the way we live, work and do business. Over 420,000 people have already attended his renowned inspiration sessions, in the Netherlands as well as abroad. He works together with RTL television and presents the weekly radio program ‘Mindshift’ on BNR news radio. Van Hooijdonk is also a guest lecturer at Nyenrode and Erasmus Universities.

Richard van Hooijdonk

Sources

https://thetrendsnext.com/automation-3d-printed-organs-immersive-tech-the-future-of-healthcare-is-tech-powered/

https://www.richardvanhooijdonk.com/en/blog/ai-could-help-people-decide-for-how-long-to-keep-loved-ones-on-life-support-should-we-trust-it/

Campbell, Patrick, 
https://www.mdmag.com/medical-news/machine-learning-boasts-90-accuracy-rate-for-predicting-heart-attack-death.

Fingas, Jon, 
https://www.engadget.com/2018/03/30/ai-predicts-your-lifespan-using-activity-tracking-apps/.

Kaser, Rachel, 
https://thenextweb.com/artificial-intelligence/2018/06/19/google-ai-predictions-death-hospitals/.

Park, Andrea, 
https://www.beckershospitalreview.com/artificial-intelligence/logitboost-algorithm-predicts-heart-attacks-death-with-superhuman-accuracy.html.

Tangermann, Victor,
https://futurism.com/googles-ai-predict-when-patient-die.

University of Nottingham, https://www.sciencedaily.com/releases/2019/03/190327142032.htm.

Zavyalova, Victoria, https://www.rbth.com/science-and-tech/328014-neural-networks-predict-human-longevity.

Source: Artificial Intelligence on Medium

Leave a Reply

Your email address will not be published. Required fields are marked *

Back To Top