TED Conversations

Neema Aggarwal

Electrical Engineering Student, The Cooper Union for the Advance,

This conversation is closed.

Do we rely too heavily on technology for medical diagnosis?

This week in my bioelectricity class, we discussed electric fields that can be measured on the body (eg the brain, skin, eyes) and the ability to interpret signals for diagnosis, lie detection etc. Currently, there exists clinical decision support systems (CDSS), which are interactive computer software systems designed to aid doctors with medical decisions. Various test results and other data from the patient is inputted into a CDSS which then processes it and provides a list of possible diagnosis and options for treatment. The problem is, like all machines, they can often make crucial mistakes.

Dr. Lisa Sanders, a physician at Yale School of Medicine, and technical adviser for the popular TV show, House, wrote a book called “Every Patient Tells a Story” dealing with the uncertainty doctors face when analyzing a patient’s symptoms. Sanders says that misdiagnoses account for as much as 17% of medical errors. She discusses how despite the many technological advances made recently, sometimes these diagnostic tools are to blame. Relying too heavily on machines and lab results can result in symptoms being missed. Or on the other hand, sometimes exam results are normal; blood tests, electrocardiograms, CT scans, all may suggest a healthy body even when that is not the case. It can take a trained, experienced eye to notice small details in the patients’ behavior to unravel the mysteries of an unknown illness. Sanders states, “For all the data they collect, machines lack important components for diagnosis. They cannot hear a patient’s story, touch a patient’s skin, or look into a patient’s eyes.”

My question is, have we become too dependent on machines and technology to make medical decisions for us? Have doctors been lured into a false sense of security by allowing tools like CDSS to provide answers? How can the value of intuition which comes only from experience be balanced with technology without being lost? Can machines ever be a good enough substitute for doctors?


Showing single comment thread. View the full conversation.

  • thumb
    Mar 30 2013: Hey Neema!

    This question reminds me of this time last year, when my modern physics professor bemoaned the fact that we (the students) were of "the wolfram alpha generation." He said we had grown to depend on websites like wolfram alpha to solve integrals for us, and so our intuition wasn't where it should be, for engineering students who have to use calculus every day. Of course he was absolutely correct, and I think it applies here, as well. I think it's important to note that the machines are doing the jobs they were programmed to do, but also that there is a risk in putting too much faith in them. Doctors who use diagnostic machines on a daily basis can lose their intuition, and those that have trained only alongside the machines might not trust their intuition at all.

    On the other hand, when a diagnosis is needed, it's needed as fast as possible. Misdiagnosis can happen whether it's coming from a person or from a machine, and the more resources we have, the better. Unfortunately, I've never worked in a hospital, so I can't really say the extent to which doctors actually do depend solely on the diagnosis from the machines, but I'd have to imagine that it's balanced with their own diagnoses and hunches.
    • thumb
      Mar 31 2013: Hi Hindi,
      I do agree the the ideal situation would be a perfect balance between machine diagnoses and a doctor's own hunches. And you're right that misdiagnosis can happen from a person as well as the machine. I wonder if there was some way to ensure that the balance remained without one side being weighted more heavily than it should?

Showing single comment thread. View the full conversation.