Healthcare AI Disruption

Sharing is caring: Facebook | Twitter | LinkedIn | Google+

Healthcare AI is the new buzzword, with it floating around as a subject of great interest. Personal and business AI is ubiquitous now, revolutionizing all fields related to technology. Be it the humanoid robots which have the capacity to replace workers in dangerous working environments or AI programs saving time for lawyers on redundant clerical work. Healthcare has been a field which was generally opaque to the advancements in computer science. And now, it is poised to reap the benefits of AI in the near future.

Healthcare AI has enormous implications. It could free thousands of hours for doctors from paperwork to patient work. Physicians spend a lot of time on redundant treatment, and AI could help prevent that. It gives them the leverage to attend patients with unique symptoms and complex ailments. This applies in particular to developing countries like India or Brazil, where the number of doctors to patients ratio is very low. Having an AI assistant could help in diagnosing patients faster and with more accuracy. This, in turn, could improve the efficiency of the doctors in charge.

Google's DeepMind on healthcare AI

The tech giants of our times are locked in a tussle to unleash their respective artificial intelligence platforms. This is happening across different verticals of healthcare. Google acquired the British AI tech company DeepMind and is pioneering research in the field of healthcare. They are concentrating on visual intelligence and neuroscience at the moment.

Currently, DeepMind has been running tests over millions of healthcare datasets from across the world. It works on understanding specific diseases by creating patterns in the big data. Google has entered into a partnership with England's National Health Service (NHS)((Google’s DeepMind agrees to a new deal to share NHS patient data)). Through this five-year deal, NHS has agreed to give access to DeepMind to develop and deploy its healthcare application.

The application being designed is proposed to be a virtual assistant to the doctor. It helps analyze the patient's health condition and alerts the physician when there is a spike or an aberration in the processed data. The ailments that are first being worked on are kidney stone problems through visual data. In the course of the five-year period, the range of challenges tackled is expected to be expanded. Issues like blood poisoning and even blindness through diabetes complications are in the AI queue.

IBM Watson's take on healthcare AI innovation

IBM Watson is taking virtual assistantship to higher levels. It uses patient history data, his economic standing and health fragility conditions for analysis. Through that, Watson can suggest the best way of treatment for his ailment. This is done over comparisons of patients with similar problems, and the process is charted out for the doctor to verify it himself.

IBM believes that this can improve the way medication is administered. Improvements in medication administration could also mean that treatments would be swifter and more efficient than before.

Apart from direct intervention, IBM is trying to work with the pharmaceutical industry in discovering new drugs. The research processes that go into developing drugs are usually very complex.((5 amazing ways IBM Watson is transforming healthcare)) The average time taken for developing a single pill is around 12 years. Watson is expected to help in streamlining this process, by accelerating the whole drug discovery chain. It can also go a bit further and understand the type of people who would respond better to the new drug. This could thereby improve the results of clinical trials.

Intelligent care robots powered by AI

Critical emergency care has always had a problem with the supply and demand. This predicament is further complicated by the shortage of doctors on ground zero. Scientists believe that AI-powered robots could be a stop gap solution in here. Robots could be used to provide preliminary medication and wound cleansing in a case of injury.

Healthcare has always been slow on the uptake of technology. Intelligent machines have the potential in improving the chances of medical help being available 24x7.((Robots in health care could lead to a doctorless hospital)) This could assist in increased adoption of healthcare AI in the near future. AI works along with these machines in not just assisting the doctors, but also in making sure the doctors tend to the most critical patients first. AI utilizes the "lean processing" methods used on the industrial floor, to improve efficiency in hospitals.

Challenges faced by AI in healthcare

One of the principal issues encountered in introducing technology to healthcare is the privacy issues that tag along with it. In the US and Europe, many legal complexities arise when there's a need for procuring patient data. The recent hacking episodes happening around the world has been a major cause for worry. The concern is that private medical records could be wrongfully used and even sold illegally.

Artificial intelligence works only as well as the amount of data that it has in its database. Consider the case of speech recognition on Google Android. Google's AI works on millions of data points, and this gives it the accuracy we see in the final application. But this process is dented in the case of healthcare, simply because of the difficulty in procuring relevant annotated data. This brings forth the question - can AI be as accurate as a certified medical practitioner? New hybrid techniques are being developed in deep learning right now. They are believed to be accurate with a fraction of labeled data along with a lot of unlabeled data. ((Three Challenges for Artificial Intelligence in Medicine))Only time can tell how effective these techniques turn out to be.

Another challenge is that artificial intelligence takes out the human factor from the equation. This, though not foreseen as a big issue, could deter patients from getting treated by a machine. Till date, AI hasn't developed to an extent where it can understand the behavioral psychology of the patients. This could be a problem in the case of a conservative population, needing a human touch to the treatment.

Apart from technical challenges, there is also the challenge of wading through a lot of regulation and red-tapes. Healthcare being highly monitored takes a long time to implement new ideas in the field. This also includes technological advancements through artificial intelligence. For instance, it took Google's DeepMind a few years to get through their landmark partnership with NHS. This though shouldn't be a great cause for concern. A lot of medical institutions are opening up to the possibility of AI. This is primarily because they want to lead the next wave of healthcare innovation. It also signals their stature in the medical circles.

There also are the instances of deploying the new technology in the actual healthcare circles. Historically, adoption of new practices has always been slow in this field. This is owing to the tremendous amounts of regulations it faces from the governing bodies. There also is the challenge of cashing in on the technology, which is quite a puzzle for hospitals at the moment. This is because there is a lot of uncertainty in the business models that govern a self-sustainable AI interface. But this isn't expected to push back innovation in any tangible way.

The business equation for healthcare AI

Like mentioned above, privacy breaches in the healthcare industry is becoming a major issue. But there is a huge possibility of circumventing this problem by reinventing the way data is held. Right from the dawn of Bitcoin, the technique of blockchain has been held in high regard. A lot of startups and AI engineers are getting behind the idea of blockchain in healthcare. It offers a lot more flexibility when compared to conventional encryption software. This is made available through better data protection and distribution. The healthcare industry is poised to adopt it soon and thus; blockchain is a good place to look into for investing your time and resources.

Procuring data, although is difficult in the West, is a tad bit easier in the East. Baidu, the Chinese search engine giant, has realized this. It is working along with a lot of medical institutions in China. Through them, Baidu procures millions of annotated patient data to help to build its healthcare AI systems. Businesses in the West should try looking towards the East for medical data. It is beneficial due to the relaxed regulations over there. This could fasten the processes of utilizing healthcare AI, into reality soon.

Putting weight behind startups which work on healthcare AI could help further the cause as well. Startups range from creating AI  to diagnose if someone has skin cancer to using AI to understand our brains. Startups are an easy way to gain access to the field. It also helps, since they can scale up and pivot faster than a conventional corporation could.

Original article: Healthcare AI Disruption at AI in Companies / Machian Future