We had more than 1,700 registrations, over 700 live attendees for our recent webinar on Hematology and Six Sigma Metrics (it's available on-demand now).
One of the side effects of a large registration, particularly an energized one, is a lot of questions. With that came dozens of questions during the live session, and over 200 questions submitted during registration.
I've never met someone willing to read a single article with 200 questions and 200, so we're going to break this down into more digestible bites. One or a few at a time, which we'll post online every few days.
Our most frequent question was, "Are Sigma metrics applicable to hematology?" Analytical Sigma metrics for biochemistry analytes gets discussed a lot, but not so much hematology.
To answer this question, let's cover some familiar ground:
Do hematology parameters have performance specifications, quality requirements, analytical goals?
Do hematology methods exhibit imprecision and bias?
Since the answer to these questions is Yes, and it is, then we can certainly calculate an analytical Sigma metric. The unique nature of hematology testing will be reflected in the quality requirements. For example, CLIA 2025 goals for the main hematology parameters were significantly reduced and have become quite demanding.
The other unique aspect of hematology is in the controls themselves. The shelf life of a hematology control is not like that of biochemistry, with perhaps years of life. Hematology controls typically last months, not years.
What's worse, hematology controls tend to die less gracefully than other controls. They trend up or down as they approach their expiration date.
Where this problem really hits hard is the 10:x rule (or 6:x, 8:x, 9:x etc). It's highly possible that a control near the end of its life will slip to one side of the mean, triggering a Westgard Rule. But that rule violation is a sign of the control, not the method.
In adapting QC to hematology, then, we'd like to use the 10:x rule as little as possible. That means we need methods at 4 Sigma or higher. Westgard Sigma Rules tell us that 4 Sigma and better methods don't need 10:x rules at all. Thus, we'd be able to ignore those violations and if the control is degrading, that's not going to bother us.
That's easy to say, "We need methods that are better than 4 Sigma." But getting method performance on all hematology parameters to be 4 Sigma or higher is a heavy lift. I might say it's not possible on today's market, with today's goals and today's instruments. There are some parameters just not good enough yet.
With the right instrument, in the right lab, however, getting a majority of hematology parameters above 4 Sigma is possible. That's where we are now.