Are you ready to consult a computer?

Posted by on in Clinical Decision Support
  • Font size: Larger Smaller
  • Hits: 1429
  • Print

Please raise your hand if you ever felt that you would never use a computer program or app to diagnose a patient during a physical exam.

The healthcare sector is notorious for being slow when it comes to adopting new technologies. Even for some patients it could be scary if they know that their trusted physician is relying on a computer program to make a diagnosis.

The main arguments are that computers are lacking consciousness, human intuition and most of all, instinct. While physicians are naturally good at matching patterns, logic and knowledge with their instinct.

Where are we in the human vs. machine race today?

Where do you get a second opinion?

A recent article in the NY Times introduced Dr. Gurpreet Dhaliwal, 39, a self-effacing associate professor of clinical medicine at the University of California, San Francisco. He is considered one of the most skillful clinical diagnosticians in practice today.

The question they’re addressing in the article is: “What can Dr. Dhaliwal do that a computer cannot?” Will a computer ever successfully stand in for a skill that is based not simply on a vast fund of knowledge but also on more intangible factors like intuition?

The short answer is YES, but only to assist the clinician in making a better decisions.

The fact is computers are good at crunching numbers, fast. The advantages are, computer applications can relatively easily be multiplied across clinics, bringing virtual specialists to the corners of the earth. And, computers are cheaper than doctors.

Up till now clinical decision support systems in the healthcare sector were only available to the selected few.

  • In the 1970s, researchers at the University of Pittsburgh developed software to diagnose complex problems in general internal medicine, resulting in a commercial product called Quick Medical Reference.
  • In the 1980’s Massachusetts General Hospital developed DXplain, a program that provides a ranked list of clinical diagnoses from a set of symptoms and laboratory data.
  • In recent years, IBM is working on Watson for Healthcare.

With the advances made in cloud computing over the last few years, we are seeing a boom in this sector.

An article in The Economist uses the example of the California Healthcare Foundation (CHCF) that uses Artificial Intelligent (AI) technology to examine retinas at a lower cost and more quickly.

They used thousands of images of retinas, both diseased and healthy, to train computer algorithms. Benjamin Graham, a statistician at the University of Warwick, in Britain, was able to develop an algorithm in 5 weeks that agreed with a doctor’s opinion 85% of the time. Diagnosing diabetic retinopathy earlier and at a better outcome.

SensiCardiac used the same approach. Training algorithms, on thousands of heart sounds, to be able to distinguish between pathological and innocent heart murmurs with a sensitivity and specificity of respectively 87% and 90%. Outperforming most cardiologists. Making it possible to diagnose a heart defect before it becomes life-threatening, quickly.

Dr. Dhaliwal concludes, “... not ever arriving at a final diagnosis is fine. Being O.K. with uncertainty is a big part of medicine.”

For most clinicians, a computer is an added uncertainty. It (uncertainty) is a part of our everyday life, but digital technologies do have the ability to reduce the risks and possible liabilities.

Are you exploring ways to use clinical decision support technologies in your practice? Have a look at how SensiCardiac can assist you in diagnosing structural heart defects.

Last modified on
Tagged in: auscultation heart
Cron Job Starts