Repeating the control: just more info, or just rolling the dice?

A recent question popped up in one of our recent webinars:

"It seems to me that repeating a QC may or may not have value, yes if it provides further troubleshooting data, and not if it is used to override a proper error flag. But to group the two together, without making that distinction seems inappropriate. Perhaps expand the survey question detail?"

This is a defense of repeating the control. And it comes by the defense honestly: isn't another control measurement more information? And isn't information a good thing? If we were running two controls, and then we get a 2s violation, then we repeat the control, now we have 3 measurements, and that is definitely more than 2. 

But what about the intentionality of that repeat? As the questioner notes, if the lab is simply repeating the control to get the "in" result, and override a proper error flag, that's not in pursuit of gaining additional information. That's just continuing to seek out the answer you want when you get the answer you don't.

One challenge to that is determining whether or not the repeat is only being done to override a previous out-of-controlvalue. Another challenge is determining whether or not the previous out-of-control value is "proper" which I take to mean, an actual error, not a false rejection. Since every sequence of repeating controls ends with an "in" value, it's hard to know if the previous values were proper.

If we could telepathically examine the lab tech who repeats the control, perhaps we can determine whether or not that control is being repeated with the correct intentions. If they did so with the desire to gain more information, then we can allow that repeat. If they did so with evil in their heart, seeking only an "in" so they can get on with their day, then that repeat is something we should not accept.

Unfortunately, neither the Levey-Jennings chart nor the Westgard Rules can read minds. So we have to make a judgment by a different approach.

It may not be possible to tease out the intentionality of a repeated control, nor the "appropriateness" of any particular control value. But we can separate out the practice of using the 1:2s from repeating the control. If we stop using the 1:2s rule, we will dramatically reduce the number of times we are tempted to repeat the control. 

A simpler solution than reading minds: just stop the 1:2s. Not for warning, not for repeating, and not for rejection.

If you switch from 1:2s with 2 controls to even the full Westgard Rules, you drop your false rejection rate by 66%. The P:fr of 1:2s with 2 controls is 9%, but full Westgard Rules with 2 controls is only 3%. 

Note: if you disagree with this, and decide to keep reading other articles until you find the answer you like the most, you might just be ignoring a proper answer.