Will Physicians Become Obsolete?

Will that target market soon become obsolete in the choice of medication and other services?



Maybe not entirely, but from the work we are doing using AI to create data-driven patient treatment decisions for specialist physicians, it is now becoming clear that at least 80% of physician work, if not more, will become obsolete in the future, and potentially the near future, with the advent of the tricorder already in trials this month.

Current approaches to diagnosis

Currently, a large amount of patient diagnosis is done by ‘Dr Google’. Although it is far from perfect, given most conditions have a large variety of symptoms in common, other work in the area of accurate diagnosis is rapidly increasing.

Let’s consider how specialists diagnose at present. Firstly, they get patient history, then the physicians may order some tests based on their memory of what they learned as well as a large amount of drug advertising and PR. The impact is that the treatment is inferior to what it could be, as humans are fallible, and patients often die unnecessarily as a result.

One study in the US (by Johns Hopkins) found that 40,500 patients die each year in ICU as a result of misdiagnosis. To put this in perspective, this is equivalent to the number of deaths from breast cancer. However, this is not to place blame on the physicians as they are in a difficult situation with hundreds of thousands of different conditions that have similar symptoms, at least 8, 000 ultra rare conditions that most doctors would never be exposed to, plus thousands of updates in each medical area every week. It is physically impossible for physicians to stay up-to-date in most areas, but especially in rapidly progressing ones such as oncology.

A study in Oncology pitted AI diagnosis against a panel of four leading oncologists wherein both sides examined patient scans and made a diagnosis around tumor progression. The AI outperformed the human diagnoses by a significant margin.

Just to clarify, given some people have been calling some things ‘AI’ and they are not. This study was, of course, using real Artificial Intelligence – not simply expert medical systems that are clinical pathway and expert clinical support systems which occasionally erroneously label themselves as ‘artificial intelligence’ because it is a buzz word. The expert clinical support systems are very rule-based and simplistic and cannot take in the complexity that is required in today’s oncology work with complex patient profiles, in comparison to what a real Artificial Intelligence (a combination of machine learning, deep learning and evolutionary computation) can do.

A lot of what physicians currently do - including testing, diagnosis and treatment decisions - can actually be done better by active data collection and collation, sensors, and AI analytics. Physicians are meant to consume all those data points and consider it in the context of the latest medical literature and the patient history, and make decisions in the best interest of the patients’ health outcome.

It actually is physically impossible for physicians to consume all the latest information and integrate it constantly. For example, how many cardiologists can digest all of the latest 5000+ articles and research updates on oncology on a weekly basis and still do their job? It is an impossible task.

Is Artificial Intelligence the future of diagnosis?

So, is Artificial Intelligence the future of diagnosis? Artificial Intelligence can pull data from trillions of data points in a second, analyze all relevant factors and come up with a far more accurate conclusion than a human brain is capable of, as the studies in this area have already shown. Now that we are in the era of personalized medicine, utilizing far more complex models with thousands of baseline and multi-omic data points, up-to-date data and data capture are critical to make the most advanced treatment decisions.

Of note is that Eularis are currently working on an Artificial Intelligence driven clinical platform for oncologists to do the following:

• Review all relevant, authoritative medical literature in oncology and collate it,
• Review all clinical trial data in the space and collate it with results, biomarkers and more
• Add in local country and hospital treatment protocols,
• Link this to the electronic patient records (longitudinal data) to examine scan images, all blood and diagnostic test results (including genetic testing) along with treatment and patient outcomes,
• Collate all that constantly updated data to allow the Artificial Intelligence to identify what treatment will have the best possible outcome for a specific unique combination of factors for a specific patient

We are doing this in oncology currently but could equally take this wider into other spaces in the future.

New technologies will allow physicians to work faster and improve their work as all the data can be taken into account – something that is simply not possible currently with all the constant increases in data. In the future, technology will replace the diagnosis component of medicine, which means that fewer doctors will be needed.

All diagnosis and treatment plans will be Artificial Intelligence powered; the physicians will provide the care. These systems will take time to perfect but they are already pretty good. Soon, like all technology, no doubt, they will become commonplace and cheap, and all people will have diagnostic systems at home. Think back to the days when computers were big room-sized objects, and Bill Gates is remembered for saying that 640K was more than anyone would ever need. Now we have more than that on our phones, let alone our computers.

The future is almost here

Inspired by the Tricorder in Star Trek, which was how the Starship doctor performed his medical examinations (by moving the device over the patient’s body and getting a full reading of everything going on), this technology is almost a reality now. The Qualcomm Tricorder XPRIZE is a competition to develop this device with a $10 million prize for the best device – which will be a handheld device like the original in Star Trek. The first user tests started in September 2016. They are around capturing key health metrics, and at this early stage are expected to diagnose 13 health conditions. The winner will be announced in the next few months. This device signals the beginning of a new era of healthcare diagnosis. The breadth and range of conditions will, no doubt, be added to regularly with more - and improved - features.

Conclusions

We are living in a very interesting era where technology is changing a lot of areas in medicine. This has implications not only for physicians, patients and health outcomes, but also for pharmaceutical companies and their traditional ways of marketing and selling products. In pharma, we have all heard about ‘lip service’, ‘patient-centricity’ and ‘beyond the pill’. However, the challenges in our environment now are forcing these changes more rapidly than those in pharma may even be aware.


For more information on these topics and how you can get involved in the next generation of medicine, contact Eularis at: http://www.eularis.com