Artificial Intelligence in Healthcare: Babylon Health & IBM Watson Take the Lead



A person becomes a doctor by graduating from medical school, but becomes a good doctor only after years of experience. What if we could give doctors the intuition that comes from years of experience on their first day? Several companies believe they can help every doctor achieve the same outcomes as top experts with tools powered by artificial intelligence (AI). AI is a term used for computer programs that can do things that we think of as “human”. Scientists use AI to automate complex tasks like driving a car or stock trading. AI has also been used in the development of new drugs.

More recently, scientists have begun to train computers to provide better healthcare. In this article, we highlight two leaders in this space, Babylon Health and IBM Watson.

AI for the patient and provider:

Babylon Health wants everyone with a smartphone to have access to affordable healthcare. They believe an app that offers instant diagnosis is the key. As their CEO, Ali Parsa, explained to the Telegraph:
“[Medical professionals] are the most expensive part of healthcare. And the second… is timing… [By] the time [most diseases] present their symptoms a £10 problem has become a £1,000 solution.”
Babylon Health believes they can drop both of those costs. Today, Babylon Health offers a free app that makes it simple for users to track their health and consult their AI-powered chatbot. For a fee, users can video-chat with top doctors who can access that user’s health records and a set of proprietary AI-powered tools that Babylon Health claims can improve treatment quality. By tracking the vitals, treatments, and outcomes across a broad user base, Babylon health has tapped an incredibly valuable dataset. This dataset makes it scalable to continually improve their AI’s performance alongside customers’ health.

IBM Watson for Oncology has a narrower focus: improving the outcomes of cancer treatments. IBM believes they can give every medical professional treating cancer the same insight that doctors at top cancer research centers have. IBM has partnered with specialists at Memorial Sloan Kettering to train their computers with a wealth of medical records and research. Launched in 2016, Watson supports doctors with patient-specific recommendations from cutting-edge treatments in a fraction of the time. According to Deborah DiSanzo, the General Manager of IBM Watson Health, Watson for Oncology had already been used in the treatment of 16,000 patients by the third quarter of 2017. With computers handling the analysis, doctors can focus on what humans excel at: treating the emotional distress of a patient fighting cancer.

Data for artificial intelligence is food for thought:

Both IBM Watson and Babylon Health agree: doctors can deliver better treatment by learning from the results of other patients. AI can learn from historical data and forecast how a patient’s ailment would respond to treatment options. Both companies are using machine learning, a technique that has become synonymous with AI in recent years. Machine learning is an automated technique used by a computer to teach itself to make decisions using training data. Training data is the fuel of machine learning, as described by Andrew Ng of Stanford University.

Babylon Health and IBM Watson have both designed systems that generate this “fuel” from their users. As they attract more users, they will generate better insights. This network effect is a virtuous circle where the product becomes better as it adds more users. The downside of products with network effects is that they are notoriously difficult to kick-start. Just think how hard it is to get the first few members for a dating website.

Babylon Health and IBM Watson have each partnered with established players to overcome this challenge and get the fuel they need to prime the motor. Babylon Health is bootstrapping their product with help from a UK NHS partnership. The UK NHS is seeking ways to mitigate their doctor shortage, and will trial Babylon’s chatbot for six months in North Central London, an area covering 1.2 million citizens. IBM Watson is partnering with Memorial Sloan Kettering to help train Watson on the wealth of clinical information and medical expertise that the center is known for.

Regulatory risk: A potential challenge:

With AI-powered healthcare products showing so much promise, one might expect regulation to pass quickly through the FDA. However, the FDA is currently struggling. As the Wall Street Journal puts it:
“How on earth are you going to regulate software that learns?”
Current regulations lack standards to assess the safety and efficacy of AI systems, which the FDA has attempted to address by providing guidance for assessing AI systems. The first guidance classifies AI systems as “general wellness products”, which are loosely regulated as they present low risk to users. The second guidance justifies the use of real-world evidence to assess the performance of AI systems. Lastly, the guidance clarifies the rules for the adaptive design in clinical trials, which would be widely used in assessing the operating characteristics of AI systems.

Despite these challenges, things are looking bullish for AI-powered healthcare. Babylon Health and IBM are only two of many new projects that are extending the reach of healthcare by reinforcing the parts that don’t scale: doctors. While each of these companies has their own view of the future, they all agree that AI will let our limited medical professionals bring the best treatments to the greatest number of people. Especially when the best treatment is taking action before we get sick.

Comments

Post a Comment

Popular posts from this blog

Why Obesity Is Still a Problem and How the Gut Microbiome Factors in

Image
Have you ever wondered why two people can eat the same amount of food, but weigh differently? This may be due to the gut microbiome, the taxa of microorganisms residing in the gut and their genes. Many people have tried to find correlations (and remedies) to obesity, a health problem that affects 13 percent of the global population. Studies from chili peppers to the amount of sleep, to vitamin C, have all tried to find probable cause and prevention of obesity. But there is another factor. Research of microbiomes might reveal a new mechanism for weight gain, beyond diet, lifestyle and genetics.  How the Gut Microbiome Shapes Obesity The gut microbiome is the taxa of microorganisms residing in the gut (the microbiota) and their genes. The gut microbiota consists of trillions of microbes, including bacteria, archaea, viruses and single-celled eukaryotes. Current estimates reveal around 1,000 distinct bacterial species in the gut, with the most common phyla being Bacteroidetes an...
Read more

Microfibrillated cellulose: A new material with high potential in the packaging industry

Image
The cellulose polymer is a naturally occurring linear polymer made of repeating units of glucose. The single polymers are stacked together forming fibrils, with these fibrils stacking together again to form the cellulose fiber structure that is present in nature.  This makes for a very interesting supramolecular structure that consists of both crystalline and amorphous regions.  Microfibrillated cellulose (MFC) is obtained through a fibrillation process of the cellulose fibers. Using mechanical shearing, the cellulose fibers are separated into a three dimensional network of microfibrils with a large surface area. The obtained fibrils are much smaller in diameter compared to the original fibers, and can form a network or a web-like structure as seen in the image below. Scanning Electrone Microscope image of WEIDMANN microfibrillated cellulose WMFC Q_standard. Image courtesy of  WEIDMANN Fiber Technology How is this material different ...
Read more

What are some of the latest waste-to-energy technologies available?

Image
Introduction: Landfills emit by-products like methane, dioxins and leachate (a toxic liquid that is formed when waste breaks down in the landfill and filters through waste), which, when left untreated, can leach into the soil, contaminating water sources, plants and even food. Waste-to-Energy (WtE) technologies that process non-renewable waste can reduce environmental and health damages, all the while generating sustainable energy. Types of waste: At present, waste is classified as the one of the following: Municipal solid waste (MSW) Process waste Medical waste Agricultural Waste Figure 1: Utility scale plants with different feeds The World Energy Council (2016) reports that according to the current rate of waste generation, global waste is estimated to reach 6 million tonnes/day by 2025. So, increasing utility scale WtE plants using MSW or agricultural waste would be a constructive way to deal with waste, as evident from figu...
Read more