Vinod Khosla, a prominent Silicon Valley investor, says that Doctors will soon be a thing of the past, replaced by computer algorithms:

Eventually, we won’t need the average doctor and will have much better and cheaper care for 90-99% of our medical needs. We will still need to leverage the top 10 or 20% of doctors (at least for the next two decades) to help that bionic software get better at diagnosis. So a world mostly without doctors (at least average ones) is not only not reasonable, but also more likely than not. There will be exceptions, and plenty of stories around these exceptions, but what I am talking about will most likely be the rule and doctors may be the exception rather than the other way around.

I think that a lot of the work that doctors do is actually more about solving complex, fuzzy problems and organising care in a way that prioritises emotional issues and causes the least distress for patients. I think that computers will be much more helpful that they are at the moment and will prevent many of the errors that are currently routine, but patients will still need a sympathetic, responsible, knowledgeable person to guide them through difficult and distressing healthcare decisions.

What do you think?

Views: 1564

Reply to This

Replies to This Discussion

Hi Chris,

If a doctor's main role is to prioritise emotional issues and cause the least distress for patients, i.e. a sympathetic, responsible, knowledgeable person to guide them through difficult and distressing healthcare decisions, then my question is, "How much of a 6 year medical degree teaches a doctor how to deal with people?".

When I went through Knox College at Otago University (half the college was med students), I saw very bright young people swotting endlessly up on pathology, and very little on how to deal and interact with people.

I think Vinod Khosla's point is that much of that memory-driven pathology can year-by-year be better performed by a computer, and then interpreted through a doctor who can then help the patient decide on and keep to to the best course of treatment. I agree with you on that.

Yet that is still not what doctors do - they still largely rely on their own memory for pathology, and then tell the patient what to do. i.e. "adhere and comply". The problem is that most folk don't think in terms of "adherance and compliance". 

Realistically, the issue is, "When will medicine change the now 100-year-old system for training doctors to be more human-centric, now that we have increasingly smart computers to help them crunch background factors, logic and probabilities related to pathology?"

There are significant roles which Doctors do like critical thinking and strategic patient care. Men have better abilities to deal with human beings rather than algorithm.  Algorithm is only a tool that would help doctors make a better decision when treating illness of patients. But not a substitute of a doctor that would relate to a human being. 

Not so fast.

“Health data breaches up 97 percent in 2011” by Healthcare IT News Senior Editor Diana Manos was posted yesterday.


"Health data breaches in the U.S. increased 97 percent in 2011 over the year before, according to a new report by Redspin, a leading provider of IT security assessments.” Daniel W. Berger, Redspin's president and CEO said, “Information security data breach in healthcare has reached epidemic proportions – the problem is widespread and accelerating.”

Even before the latest bad news about fading patient trust in digital, in dentistry, EHRs are not only more dangerous than paper dental records, but they are more expensive as well and do nothing to improve care. Good luck getting a computer to extract a tooth.

D. Kellus Pruitt DDS

There is an old saying in medicine and something I learned about first in my residency - "Patients don't care how much you know until they know that you care". Yes, technology has its limitations and can't replace the human component. I see computer algorithms being widely used already in the form of CDS software or EMR features. Watson may be too extreme!

As much as I don't like a computer directly guiding the care of a client, I think they could be very useful. People like people. If there was a computer that spit out a diagnosis and a care plan to someone who would then be told to the client. I don't believe this is a genuine process. There are so many grey areas when it comes to diagnoses that I don't believe a machine could be solely responsible in place of a physician.

Using a machine for this takes too much out of the human factor. Although humans do make errors, I believe the machine could be used as a tool and reference, but not a relience. A machine cannot assess a client, take its vital signs, hear lung/heart sounds, or judge degrees of parts of their assessment. Overall, not a good idea to let robots run our health care.

I agree that it is not a good idea to let computers run our healthcare.  Personally I do not want to go to a physician's office to simply have a machine give me a diagnosis.  If I am sick enough to go to the doctor than I want to talk to an actual person and have someone who knows what their talking about assess me.  Computers can't not provide the genuiness that is required by health care workers.   Also, many diseases have numerous signs and symptoms.  A simple stomach ache can be a symptom of many different conditions and a machine could not account for all of these. 

Something else that a machine could not do is save a life.  If a person comes into the emergency room and is having a heart attack, there would not be enough time for a computer to diagnosis the individual.  Doctors need to be able to think quick on their feet and change direction at a moments notice.  A computer could never do this.  Computer can assist but they can never replace doctors or any other medical personnel

I can see the value in enhanced technology for basic needs, perhaps an automated Urgent Care center for basic diagnosis, but an overemphasis on technology could ultimately be a hindrance.  The current focus on health embodies more than simply physical health deviations like the common cold, it is inclusive of interpersonal dynamics and emotional/behavioral well being.  This would prove very difficult to account for with an algorithmic diagnostic system.  I am unsure of how much the technology has progressed, but detecting subtle nonverbal cues and applying the proper context to them has to be nearly impossible to program.  I do concede that possibilities in technology are expanding at an exponential rate every day and who really knows what we will be able to achieve with it. 


Maybe this article will help.It says almost everything about how I where I stand on this topic.

I believe that eventually doctors will be a way of the past. With the way healthcare is going, I feel that nurse practitioners will take charge of patient care. Nurse practitioners are more cost effective for insurance companies than a visit to a doctor. Many people are choosing to use self help methods instead of seeing a doctor due to heavy doctor bills. I do not completely agree with computers taking over most of the health profession. Computers are just a tool. We need more subjective data and observation than what the computer can provide. Computers will be helpful, but nursing is more than technology.

Doctors can never be replaced, I think. All the technologies and programs around are to supplement them to be more effective and productive in their career. Diagnosis and treatment of a condition is complex and cannot be totally taken care by rule based or case based systems. I see, in terms of ideal health care environment both the doctors and information systems are inseparable.

It would be helpful if more MDs would contribute to this discussion, to clarify and expand upon what skills they are trained in, and which of these skills cannot easily be transferred to computer algorithms.  But even those of us who have been treated by a physician can come up with important skills that don't seem to be compatible with the capabilities of computer algorithms alone:

--Make care coordination decisions
--Personalize treatment plans
--Perform physical exams and other health assessments using physical senses
--Recommend appropriate followup providers/specialists
--Instill patient trust and confidence
--Make decisions to order/perform additional diagnostic tests
--For chronically ill patients, determine frequency of tests aimed at monitoring of disease    indicators/health status
--Verify accuracy of/sign off on clinical documentation 
--Administer certain treatments
--Perform surgeries
--Interpret medical images
--Coordinate rapid decision making with interventional procedures in life-threatening emergencies

--Ensure required data is captured/stored for further analysis (how far are we from having patients talk to computers, and algorithms that reliably convert the verbal input into data formats that support good clinical data analyses?)

Not being a doctor, I probably missed a number of other categories of very important physician skills that do not make a good fit with the capabilities of computer algorithms.  I believe these algorithms will have greatest benefit in helping doctors infer or deduce a diagnosis - with reasonable speed and high accuracy - when there is an overwhelmingly large amount of data that must be evaluated.  Another high-benefit area involves reconciliation of proposed treatment components (medications, procedures) with each other and patient-specific information (e.g. allergies) to minimize life-threatening outcomes (e.g. a drug-drug interaction or drug-allergy interaction). 

Computer algorithms can only use data that is within their reach.  We cannot assume that all of the important data that can be used to guide diagnosis and treatment decisions are within the computer algorithms' reach (inputted/stored). People involved in the patient's care can arrive at this additional relevant information, about the the patient, that may not be easy to store for algorithm access.  




Advertise on the Health Informatics Forum


Advertise on the Health Informatics Forum

New Media Medicine

This website is published by New Media Medicine Ltd.

New Media Medicine develops new media solutions for the healthcare industry.

As well as the Health Informatics Forum, we operate a discussion forum for medical students and doctors:

We also operate the eLearning Talk social network for elearning developers.

© 2014   Created by Chris Paton.   Powered by

Badges  |  Report an Issue  |  Terms of Service