Do we need doctors in the future? National Health Service medical director Sir Bruce Keogh believes that doctors will be replaced by computers, since artificial intelligence allows computers to start diagnosing patients with more precision than what was previously possible (Campell, D. 2017). Especially in radiology, IT comes with groundbreaking innovations. By analyzing the huge datasets collected from patients and their symptoms, machine learning can help us to improve diagnoses, choose the right treatment and to execute this treatment (Jacobs, F. 2016).
So what will the advantages of using computers for diagnosing patients?
Diagnostic imaging such as X-ray, CT scan, MRI scan and echography are of high importance for diagnostic measures and treatments. Radiologists analyze the MRI-images with bare eyes and are able to quickly and adequately diagnose the outcome. However, analyses made by human eye are an estimation, there are not quantitative (Jacobs, J. 2016). Due to technological innovations that enable computers to read and analyze diagnostic imaging, analysis can be made quantitative. Artificial intelligence decreases the dependency on incomplete knowledge or memory of the doctor, which decreases human mistakes and makes diagnoses more trustworthy (Vize, R. 2017). A great example of this is the artificial intelligence machine IBM’s Watson that has been used in the hope to diagnose a 60-year-old woman, since doctors weren’t able to figure it out. The machine went through 20 million research papers about cancer and within 10 minutes came up with the correct diagnose as well as suggesting a new treatment for an uncommon form of leukemia (Ng, A. 2016).
Another advantage of artificial intelligence in diagnosing is that small and subtle changes over time are very hard to observe by human eye. When observations are quantitative, the smallest changes will be noticed which enables doctors to adapt daily patient treatment (Jacobs, F. 2016).
Will machines replace radiologists in the future?
The revolution from bare eye analysis to computer-assisted diagnosis is groundbreaking (Jacobs, F. 2016). As a consequence of artificial intelligence systems, will the practicing radiologist be redundant? According to Van Buchem this will not be the case in the upcoming decennia, because there will always be the need for an expert with sufficient knowledge to analyze the complex information (Jacobs, F. 2016). The systems will be more advanced for computer-assisted diagnoses, and not for complete computer-diagnoses. The founder of digital healthcare company Babylon stated; ‘there is no solution which can fundamentally cut the costs of healthcare as long as we are reliant on humans’ (Vize, R. 2017). Since the need for an expert will not decrease, unfortunately healthcare cost will not decrease in the near future.
We can conclude that due to artificial intelligence, diagnoses can be quantitative, which increases accuracy of diagnoses and decreases doctor’s errors. Small changes in the status of the patient can be noticed, by which the doctor can decide to adapt the treatment in an earlier stage than he would be able to do without this quantitative knowledge. The role of doctors will be refined, not replaced. As a consequence, it is unlikely that healthcare cost will decrease in the near future.
References
Campbell, D. (2017). Patients’ illnesses could soon be diagnosed by AI, NHS leaders say. [online] the Guardian. Available at: https://www.theguardian.com/society/2017/sep/12/patients-illnesses-could-soon-be-diagnosed-by-ai-nhs-leaders-say [Accessed 27 Sep. 2017].
Jacobs, F. (2017). Kunstmatige intelligentie verandert beroep van de radioloog – SmartHealth. [online] SmartHealth. Available at: http://www.smarthealth.nl/2016/11/28/philips-kunstmatige-intelligentie-radiologie/ [Accessed 27 Sep. 2017].
Ng, A. and Ng, A. (2017). IBM’s Watson gives proper diagnosis after doctors were stumped. [online] NY Daily News. Available at: http://www.nydailynews.com/news/world/ibm-watson-proper-diagnosis-doctors-stumped-article-1.2741857 [Accessed 27 Sep. 2017].
Vize, R. (2017). Technology could redefine doctor-patient relationship | Richard Vize. [online] the Guardian. Available at: https://www.theguardian.com/healthcare-network/2017/mar/11/artificial-intelligence-nhs-doctor-patient-relationship [Accessed 27 Sep. 2017].
Thank you for your interesting post, Charlotte!
I also believe that artificial intelligence has many advantages for medical imaging. When it comes to healthcare, quality is of the essence as we are dealing with people’s lives. The interesting thing about implementing AI technology in healthcare is that it will actually make it more human. I think that AI technology will assist radiologists in diagnoses, instead of replacing them. This is because radiologists’ and other medical practitioners’ knowledge and skills will still be needed to verify the output given by AI and to attend to other matters such as face-to-face contact with the patient.
Secondly, you mention that it is unlikely that healthcare costs will decrease because (human) doctors will not be replaced. I do believe that this is possible, due to the use of innovative technologies in healthcare that will enable remote healthcare solutions. The Mercy Virtual Care Center nearby St. Louis, uses technology, such as remote monitoring devices and video interfaces, to help patients. By doing so, they don’t need to invest in hospital beds, because they assume that patients are taken care of in their homes.
Another way digital technologies, such as AI, can possibly reduce healthcare costs is by reducing the number of human errors made. As you mention in your post, medical practitioners analyze images with their bare eye, making it vulnerable to human errors. If, in the future, AI technologies can provide accurate analysis, this will reduce the number of errors, subsequently reducing costs.
References
https://itpeernetwork.intel.com/making-healthcare-human-artificial-intelligence/
https://www.christenseninstitute.org/blog/will-ai-really-lower-the-cost-of-healthcare-a-lesson-from-a-hospital-without-beds/
Interesting article Charlotte. Although i do agree that many processes, (especially imaging-recognition) will be replaced by machines, there are numerous qualitative aspects that come into play when doctors make decisions regarding treatment. It will be interesting to see what role internet healthcare websites such as WebMD will have going forward when helping patients with smaller ailments.
Hi Charlotte,
Thank you for the amazing post! This is a very interesting topic to me since I almost chose to study medicine.
I am a believer in AI playing a more and more important role in diagnosing diseases and even performing surgery or curing diseases. There are solid points of AI’s advantage in diagnostic in the blog, which got me very excited.
Here I would also like to mention an advantage of AI over a doctor in performing surgery and curing patient – as long as it’s well designed, it could easily overcome the problem that most doctors only have experience in particular regions, which limits their knowledge and capability of dealing with patients of other races. I could give a typical example. A year ago my mother, who is a Chinese, needed a minor surgery when she lived in Germany. The anaesthetist, only having experience dealing with westerners, omitted the fact that Asians’ tolerance for medicines is lower than whites. My mother ended up staying in ICP for a week. AI could help avoid such issues. Also, a senior doctor is typically an expertise in a particular field while has limited understanding of the others. AI, however, could combine knowledge of different medical fields and solve a sophisticated case with multiple diseases.
While some challenges that AI might face on its way of occupying medical field should still be noticed. For instance, similar to technology in IT related industry, knowledge of medicine develops and renews dramatically fast. A doctor needs to keep studying and taking exams to remain updated with new theories. Similarly, high costs of constantly updating the AI system should be expected. The updates could be minor such as adjustment of diagnostic, or could be major such as discovery of new diseases that causes incompatibility of old medicine use, leading to sophisticated changes in treatment plans. In a more sophisticated case, it could take much more effort to educate programmers for the adjustments. Another challenge would be educating talents with knowledge in both programming and medicines, just like BIMers.
Yet, the benefits of introducing AI in hospitals are definitely worth the effort to solve the upcoming challenge. So despite of all the challenges, I still believe AI is the trend in medical field.
I believe that with the years to come, automation will become higher and higher with ever so increasing pace due to AI’s ability to learn faster and more efficient and as hardware is becoming more and more powerful. Automation has already taken a huge part of the factory jobs that were once done by humans (for example: welding robot arms). On the other hand, complex duties, requiring years of experience and practice are less endangered in the short run, thus I believe that doctors are not expected to be replaced by machines in the next 20-30 years. On the other hand, AI-enabled instruments to help detect certain health issues are coming as a part of the doctor’s tools for examining patients for sure.
Interesting post, Charlotte!
Many exaggerate the role of AI in substituting human jobs, but examples like the one you give prove the opposite. It is incredible how computer assisted medical services are creating better diagnosis for patients and observations of a condition overtime are becoming more accurate. The examples of the comments above regarding the decreasing costs of healthcare thanks to remote monitoring devices and better suitability to different regions around the world are also very interesting.
What I think is an essential point regarding the topic, however, is the relationship that doctors have with their patients. In my opinion, nearly all types of healthcare services rely on an established base of trust between the doctor and the patient. AI cannot substitute doctors since machines cannot feel true compassion, emotions and understanding. At least not for now. Who knows, perhaps someday very far in the future we will not be able to distinguish robots physiologically and mentally from a real humans.