Blog. Why my favorite doctor is a robot
In the Netherlands, 10% of patients are faced with an incorrect or late medical diagnosis and annually 38,000 people suffer from unnecessary medical damage (source: www.gezondheidsnet.nl). Problems with the heart, lungs and abdomen are particularly difficult to recognize. If you are not treated, receive the wrong treatment or the treatment is not executed properly, the consequences can be fatal.
It isn’t however my intention to call the expertise and commitment of doctors into question, or to criticize the quality of health care in the Netherlands. This is a call to make much more use of robots, big data and artificial intelligence, as this could significantly improve the quality of health care. The same applies to the quality of service in a number of other businesses, but more on that later.
The principle is very simple. There are so many diseases and so many different treatments, that you can’t realistically expect an individual doctor to know all of them in detail. That’s why doctors specialize and work in teams. But what if they were able to not only consult a number of colleagues, but also be supported by a computer that knew everything that was ever published in any medical book or journal anywhere in the world?
Watson and Dr. Foster
The amount of available medical data doubles every five years. This will increase even faster in the coming years as more people start to measure their vital bodily functions continuously – including heart rate, blood pressure and glucose – using portable devices (wearables). Processing and analysis of this data is no longer possible without computers.
Two examples of companies that collect data worldwide on diseases, diagnoses and treatment methods, are Watson and Dr. Foster. Watson is IBM’s cognitive computer. Watson works differently than Google. If you have a question, you can enter a few keywords into Google and you will get a list of very many, relevant publications that you can study further. Watson offers a much more advanced service.
First of all, you can communicate with Watson using natural language. You ask your question as specifically as possible using ordinary sentences. Watson understands the question and comes back with possible responses including the corresponding evidence and a confidence level for each response.
As a doctor, you consider the possibilities offered and decide if you need to carry out more tests or if you are ready to make a definitive diagnosis. It is also important to note that Watson is not a static database, but it is self-learning. Watson will therefore, over the course of the time, be able to answer more questions, with ever increasing accuracy.
Dr. Foster is also interesting. This British company maintains an international database with data on mortality, complications, re-treatments, duration of treatment and other variables relating to over 250 diagnoses. Worldwide, 40 hospitals are participating, including university hospitals in the Netherlands. By benchmarking data and sharing knowledge, the quality of health care improves and global standards for the treatment of certain diseases emerge.
It is evident that the large-scale collection and analysis of data improves the quality of health care. There are also unintended consequences. The central database shows which doctors and hospitals are performing well and which less so. Furthermore, it’s no longer possible to keep this information secret. Transparency has become the norm.
The result is that patients will become more critical in selecting their doctors. Doctors with good outcomes will be in demand, and doctors who perform below average, will rapidly lose patients. The winner takes all is thus manifested in health care and applies to both individual physicians and healthcare institutions. The result is further concentration, not directed top-down by the government, but driven by patients who demand the best health care possible.
The use of computers and robots in health care is not restricted to diagnostics. There are also robots that perform surgery autonomously (without a surgeon), error-free and at a fraction of the cost. In nursing we see robots that help people get out of bed and wash patients’ hair. And there are humanoid robots that can teach a gymnastics class to the elderly. New robots have the ability to learn and they can interact with humans thanks to their cognitive capabilities. For true empathy you need real people, but robots are beginning to learn that too.
Artificial intelligence may also improve the quality of service in other businesses. Think of tax advice, financial planning and legal assistance. Would you prefer advice from someone who relies on years of experience or someone who gets support from an intelligent system? In all knowledge-intensive occupations robots could offer added value. Not to replace the professional, but to enrich their services and improve their effectiveness.