> For the few patients whose oxygen saturation measurements were between 85% and 93%, the difference between the smartphone estimate and the ABG device varied b

> For the few patients whose oxygen saturation measurements were between 85% and 93%, the difference between the smartphone estimate and the ABG device varied between -5.5% and +2.5%. In other words, the readings become less accurate as the patient becomes more hypoxic. That's accurate enough to tell whether you have serious Covid in many cases, especially if your baseline reading is 99% like mine. "No evidence" has killed a lot of people in this pandemic by discouraging practices that work but haven't been RCT'd. "Should not be trusted" because something has error bars, even when they're smaller than the dangerous range, could kill even more. It's good to have the information out there that this is an inferior way to check, but I already assumed that was true just because it was only a phone app. "Possibly-inaccurate and free" is the choice I want to make with my health. I have low risk and it leaves more oximeters for others.