What’s the word for that crusted plug that clogs the nozzle on the ketchup squeeze bottle? The one that resists the even flow of the tomatoey goodness until it can’t no more and then suddenly BAM! everything is red except the target meatloaf, and the counter and back splash look like the Tate-LaBianca murders. Equally hard to clean up from, too.
Thankfully the forecast calls for snow. Listen to me, grateful for snow. What gives? Well, it’s not because I’m a skier, I’m not, I’m a terrible skier. Closer to pathetic really. I want the insulation and light reflection snow provides. Especially the latter. We have a feeble sun this time of year and absent snow the dark ground absorbs all the light for itself. Making morning walks with Lucy less joyful.
And the world needs more joy, not less.
Here’s something: researchers claim they can detect colorectal cancer with a breath test. My dissertation research involved the study of human colon cancer. (Side note: If the malls are closed, you’re out of books to read, the chores are done, and especially if your normal sleep aid is failing you, please click the “About” link over there on the right, and look for the link to where you can read my dissertation).
Anyhoo, interesting article. Except I believe they understated the predictive accuracy of their test, unless they have more data than what was shared in the article, or a non-standard formula for accuracy.
The breath test analysis correctly identified 32 of the 37 patients with colorectal cancer and incorrectly diagnosed cancer in seven of the 41 healthy patients. Overall, the breath test had an accuracy rate of 76% in identifying patients with cancer.
In other words, 32 true positives (TP), 34 true negatives (TN), 7 false positives (FP), and 5 false negatives (FN).
Predictive accuracy = (TP + TN)/(TP + FP +FN + TN) = 84.6%
Some commenters to the article wondered if “shitty” smelling breath would produce a true positive. Ha ha, very funny.
RKN – With news, yesterday, that dogs seem to be able to detect lung cancer in people, what do you think of the possibility of dogs begin able to detect colorectal cancer in people?
If cancer detection is the result of a sniff test, so to speak, wouldn’t dogs be rather more ideal sniffers/detectors of cancer than a machine which is programmed by men? What are your thoughts?
Can’t rule dogs out, but in terms of a real useful test I’d be concerned about specificity and sensitivity, and thus their predictive accuracy.
All that coffee and bladder stones must be a tough whiff for Lucy.
I’d be concerned about specificity and sensitivity, and thus their predictive accuracy.
Understandable concerns, RKN. With that conceded, though, as far as I am aware, when dogs are trained for drug sniffing, or bomb sniffing, etc., they are being trained in specificity, with good rates of success. As to dogs’ sensitivity, research has already proven dogs’ exceptional abilities in olfactory sensitivity, they’ve good noses, which, I think, would be a plus in the the plus column in regards to predictive accuracy. Just thinking out loud.