It’s difficult for me to remember what things were like before we could simply look up information on the Internet. I remember working on a research project which required booking time with a librarian (not the one in the fantasy movies), who would assist me in the use of a primordial version of PubMed (a database of article citations). After an hour of time [which cost a significant amount of money], I was presented with a paper copy of the list of references to journal articles that matched my areas of interest.
The purpose of the librarian was not simply to type onto the keyboard and tear off the list of results. The librarian was there to help you break your request down into keywords and concepts in order to optimize the quality of your search results. If you were interested in “cancer”, but did not specify the right search words, the list of citations that were returned by the computer could very well be useless.
I remember a number of times when the citation on my printed list was to a book on the topic of interest. I imagined myself sitting and reading two or three lengthy technical tomes of information that may or may not have assisted me in finding the specific answer I was looking for. The process was of course lengthy and painful. Sometimes, the point of getting an initial list of journal articles from the librarian was primarily to look at their list of citations, in order to direct you further down the rabbit hole.
The obvious question at that time was, how could one be sure that all of the related research was uncovered. The answer was that definitely, all of the pertinent research was not recovered. This explains why present day reviews of old research may find fundamental flaws in summary discussions. The authors of the old journal articles, simply did not know of certain research that had been done at the time. One of the reasons for missing pertinent other research was that it could have been “hidden” in the journals of other specialties.
Not long ago, the power of present day computerized learning systems was demonstrated by their ability to find links to research that apparently had nothing to do with the primary area of interest. I remember reading about one cancer researcher who was particularly interested in a given enzyme. As it turned out, botanists had been working on the same enzyme and had in fact already discovered a great deal about it. Because the primary cancer researcher never thought to look at journals from other fields of research, important findings were missed. The IBM deep learning system called Watson, demonstrated its ability to find these “hidden in plain view” parallel research studies. As my late brother once said, the cure for cancer may very well come out of a lab studying yeast. But without an IBM Watson type search engine, critical data could remain lost in the stacks of journals that cancer researchers never reviewed.
What will be the role of physicians in a future where computers will be capable of automatically capturing all pertinent clinical information and then discerning the best treatment for a patient’s complaints? Many people have already discussed this at length, and understandably, many doctors insist that computers will never be able to replace human physicians. I don’t pretend to know what the future will bring. I am willing to say that the role of physicians will fundamentally change within the next two decades.
I would humbly suggest that the new role of physicians, for at least some significant period of time in the future, will be equivalent to the librarians of yesteryear. Until the computer-human interface is so perfect that any human being in any language can easily retrieve whatever information is necessary, there will be a need for a medical professional to assist people in extracting information about their ills.
There may very well be a significant period of time during which the primary role of physicians will be to guide people through the use of specialized clinical systems that ultimately make the diagnosis. These systems, however, will function far better if a specially trained “doctor-librarian” guides the patient through the proper use of the system. So it may very well be that the future physician will do the equivalent of today’s doctors, i.e., he or she will ask about the patient’s condition, run certain tests and examine certain physical findings. But the future doctor will also review computer collected clinical information and then sit with the patient in front of the computer to hone down through impossibly huge medical databases, to find the most appropriate diagnosis and treatment.
Just as with the old-time librarian that I described above, the computer will be doing the heavy lifting. In the past, the librarian typed in certain keywords and then guided the user towards improving the search. Future physicians may very possibly have an equivalent role. These future doctors will still need to study anatomy and physiology, and they will need to understand pharmacology and x-ray technology. But all of this knowledge will be necessary in order to optimize the output from the intelligent search engine they are using, rather than diagnosing the patient directly from personal memory or by using a specific EHR.
Will the role of physicians ever be entirely eliminated? I personally think that this is likely. At the same time, I do think that it will take at least a couple more decades until human physician involvement is made totally unnecessary by advanced computer systems. The only real advice I can give to the first year medical student of today is to stay flexible, stay up-to-date, and to be ready for a new world that is presently unimaginable. This would have been good advice for me to receive in my first year of medical school, and I still think it is good advice today.
Thanks for listening.