To err is human

·3 min read

The chief radiologist for Eastern Health says any large-scale review of diagnostic images would likely find examples where a fresh pair of eyes yields a different interpretation than that of the original radiologist.

And that’s the case whether the equipment used was up to standard or not.

Dr. Angela Pickles admitted Friday that no one can say for sure whether the handful of discrepancies being found among thousands of mammograms under review would have come to light no matter what.

“When is the last time you picked up a book that has been proofread I don’t know how many times by somebody and you will still find mistakes in it?” she said. “We are only human, so we are going to find mistakes.”

All four of the province’s health authorities are reviewing more than 13,000 mammograms because they discovered they were read on three-megapixel monitors, rather than the required five-megapixel screens.

Three of the authorities — Eastern, Western and Labrador-Grenfell — announced Thursday that six scans have been targeted for followup out of more than 2,000 read so far.

Last week, Central Health said it singled out nine scans out of more than 3,000.

But Pickles says that doesn’t necessarily mean those scans were incorrectly interpreted.

“First of all, when we look at discrepancies, some of this is differences of opinion,” she said in phone interview. “For instance, if I was reading your chest X-ray and I say, ‘Hmm, is that like a little early pneumonia?’ I might suggest you get a followup in a month’s time just to see, whereas someone else might say, ‘No, I think that’s pneumonia. I’m just going to call it,’ and say you should get antibiotics.’”

Someone else might say it was just a bit of old scarring, she added.

While the field of diagnostics is increasingly precise, it’s not perfect.

“If you were to review 10,000 of my chest X-rays that I have read over the past 25 years, you would find mistakes that I have made. I’m only human. I try my best not to make mistakes, but things happen. There are computer errors, things don’t load properly, I get interrupted and sidetracked with other things in this department, and I make mistakes. And we are going to find those, just by the sheer nature of having a second look at anything.”

Five-megapixel monitors were chosen as the minimum standard for digital mammograms in order to correspond to the higher resolution used for film-based mammograms.

But Pickles said there has never been an objective assessment as to whether the different screen resolutions make a difference.

“There’s no non-biased studies done. A lot of the studies done on this are industry driven. And those five-megapixel monitors are more expensive than a three-megapixel monitor,” she said.

“There has never actually been a study that says ‘A is better than B and you will miss things on a three-megapixel monitor.’”

The difference between them is not detectable to the human eye.

On a five-megapixel screen, the pixel pitch — the distance from the centre of one pixel to the other — is only .165 mm. They’re .21 mm apart on a three-megapixel screen. But the human eye can only detect items that are at least .25 mm apart.

Still, Pickles says a tiny calcification might show up on one and not the other, depending on what magnification is used by the viewer.

She said it would be extremely useful if the current review also served as a study on screen resolution, but the resources to do so would be too time-consuming.

The authorities are still investigating why substandard screens were used.

Peter Jackson, Local Journalism Initiative Reporter, The Telegram