Finding what you want to find!

“Because I was looking for it”, says Sherlock Holmes, half-patiently, in the middle of an intense fight with Lord Blackwood. He was referring to the poison-tipped fine-glass needle.

While this story had a happy ending, looking for what you want need not necessarily be the perfect way ahead. The neurosurgical team at Tufts had a tough lesson to learn.

They found what they were looking for and pretty much shut out everything else.

While the neurosurgeon had asked for omnipaque, the pharmacist had informed the nurse about the unavailability of omnipaque, and had instead supplied them with two bottles of MD-76, which clearly was contraindicated for spinal infusion. The serious warning saying “NOT FOR SPINAL INJECTION” had slipped through multiple barriers not because nobody was looking for it. Rather because everyone was looking at the positive signs such as being handed over by a trustworthy pharmacist or nurse.

You expect them to be correct. What you expect to happen will happen because most of the time it actually does.

This is not the first time serious implications have been served up as a result of confirmation bias. Definitely not going to be the last time. Numerous studies similar to this by Mendel et al have shown that physicians looking for positive evidence find positive evidence and tend to ignore non-conforming signs and symptoms, leading to misdiagnosis. Diagnostic errors by their sheer magnitude in terms patients per year (close to 12 million per year in the US) are a huge drain.

One possible mechanism to prevent confirmation bias is to have better designed software and protocol-based diagnostic methods. Although this might seem to promote cookbook medicine, there is nothing sexy about making a diagnostic error due to a cognitive bias in the name of promoting individual flair in medicine. Atul Gawande’s essay titled “Big Med” makes a compelling argument for protocol-based medicine. However, the quantum of false equivalency involved in comparing fast-food chains that produce standardized food items to medicine whose menu varies as much as the people it serves is arguable.

References:

1. http://www.bostonglobe.com/lifestyle/health-wellness/2014/08/30/surgical-error-tufts-medical-center-prompts-widespread-changes-operating-rooms/8JFgKrx6FIvLLT6Uja2bkJ/story.html

 

2. http://journals.cambridge.org/action/displayAbstract?fromPage=online&aid=8415519&fileId=S0033291711000808

 

3. http://www.newyorker.com/magazine/2012/08/13/big-med