In our most recent post on the future of GCP QA, we looked at the differences between manufacturing and clinical research environments. Today, we're going to start to examine how the elements of the Pharmaceutical Quality System that support the GMP environment don't always fit the GCP environment. First up for our consideration: SOPs.
As we saw in our last post, the outputs of a GMP manufacturing process are physical products that meet standards for safety, identity, strength, quality, and purity. Unsurprisingly, processes for producing drugs that can meet these standards are strictly defined and closely detailed. In this environment, an SOP is like a recipe: It includes all the information a user needs to complete a task, organized in different ways to facilitate retrieval of information - a list of ingredients at the top, followed by a list of chronological steps for creating the dish.
In GCP, we've internalized this concept, but because of the variability inherent in clinical trials, it doesn't always serve us well. A good GCP procedure is less like a recipe for creating soup in an industrial kitchen and more like the way my Italian family makes wedding soup. The recipe might be written down to help a clueless son or daughter reproduce it on their own for the first time, but it would look something like this:
Boil chicken with celery tops and some onion. Skim off fat to make stock. Make small meatballs. Simmer chicken and meatballs in broth with celery, escarole, spinach, and green onion.
There are a lot of details missing from that recipe. Quantities, for one. Cooking times. The whole meatball subprocess is barely addressed. But because that son or daughter had watched wedding soup being made at every Thanksgiving and Christmas for decades, they would be able to figure it out - with a few phone calls home (i.e., oversight).
We could write down all the detailed steps, but...have you ever made wedding soup? It takes days, there are dozens of steps, and every year right in the middle of the thousand tiny meatballs, I weep a bit. Even if we did write it down, we could never actually comply with it, because each branch of the family has their own spin. Some add pastina, some thicken the soup with breadcrumbs, some use carrots. There are different ratios of escarole to spinach. We would need a whole chapter on various methods of making, sizing, and cooking the thousand tiny meatballs. And why do any more work when that three-sentence recipe does a perfectly good job of conveying the essence of wedding soup for this user group?
In GCP, we've internalized the "industrial recipe" approach to SOPs. Even when we insist that our SOPs are going to be "high level" and "fit for purpose," they have too much information, because we are afraid of deviating from the template that was developed for the manufacturing floor. We commonly see Scope sections that reiterate the purpose; Responsibilities sections that duplicate the procedure (but not exactly or in chronological order!), Definitions that repeat guidances, and Reference sections that are half a page long but always manage to leave out some regional regulation or guideline. We spend too much time writing SOPs, too much time reviewing them, too much time training on them, and too much time tracking and correcting deviations when we're out of compliance with all that detail.
Meanwhile, as we lavish time and attention strictly controlling and painstakingly detailing our SOPs, study teams are standing off to the side writing dozens of study-specific plans. Even when teams do a stellar job of producing these plans, the sheer number of them and number of revisions during the study increases the chance of an error or inconsistency. They are frequently redundant with other plans, contradictory, ambiguous, and non-sequential. Study-specific plans are the first-line support documents for the study team. Errors, ambiguities, and inconsistencies in these documents carry real risks, but they do not get nearly the focus that SOPs do.
It's like we spent all our time writing a detailed recipe for wedding soup, which we make primarily at Thanksgiving and Christmas, but no one thought to teach their kids how to make pasta, which is the main dish every Sunday.
In short, SOPs are important, but study-specific plans present a greater risk to non-compliance, and our efforts should be commensurate with this risk.
In our next blog post, we'll look at some ways that GCP QA can do that.
