The Problem: On the one hand we’ve got statistics. On the other hand, we’ve got human lives. Because of the latter, our hearts tend to have zero tolerance for errors in clinical trial data. How do we find the right level of risk and quality control (QC)?
Let’s tempt the wrath of risk assessors everywhere and look at the issue. What’s the real purpose of QC? To be sure, it is NOT to achieve perfection. On that most of us can wholeheartedly agree, including the FDA.
The question is how to decide how much and which types of QC will help ensure the acceptable error margins we’ve set for ourselves.
That’s a whole other kettle of fish.
The Solution: What we strive to produce is a quality management system that works as a whole. Over time, we add and subtract QC tasks until the right balance emerges—one that satisfies both the risk assessment model and the passion for correct data.
At Veracity Logic, one of the steps in our total QC package is a little bit of bother known as “visual verify”. We’ve reached down into the data management archives, pulled it out, dusted it off, and propped it up on the mantel, remembering a time when it was axiomatic that any single-entry act in the double entry world of quality control had to be independently reviewed. One can almost hear the wooden wagon wheels creaking….
Every data change (DCR) and configuration change (CCR) we’re asked to make in an active IRT project system is executed by a qualified project team member and then subjected to an independent visual confirmation by another qualified member of the team. Successful verification is signed off by the reviewer and becomes part of the project record. Only then is the DCR/CCR regarded as complete.
One almost wants to hum the theme song from ‘Somewhere In Time’…