In the first nine posts in this series, we looked at the history of Quality, with a big emphasis on its roots in manufacturing. In post 10, fittingly, we saw the culmination of this history in ICH's Q10 guideline on the Pharmaceutical Quality System, adopted as the gold standard for manufacturing environments. In post 11 we saw which of these quality system elements were and weren't specified in GCP regulations and guidelines.
Today we're going to compare GMP and GCP environments to set the stage for this question: Is a GMP-style Quality System "fit for purpose" for a GCP environment? We've summarized the differences between GMP and GCP before on this blog and in various presentations as "GMP is from Mars, GCP is from Venus." Let's take a fresh look:
- Products. The output of a GMP manufacturing process is a physical product that meets standards for safety, identity, strength, quality, and purity. Each unit must fall within strict tolerance limits for each standard; ideally, each unit of IMP is identical to the next. The outputs of a GCP clinical trial are data that give an accurate reflection of the safety and efficacy of the product under the given conditions. Data must meet standards for data integrity, defined in E6 R3 as attributability, legibility, contemporaneousness, originality, accuracy, completeness, security, and reliability, but data points typically fall within broad tolerance limits. Variation in data is expected; identical data points would be suspect.
- Facilities. Manufacturing occurs in a facility, or sequentially in a relatively small number of facilities. Large clinical trials take place in hundreds of facilities (investigational sites, laboratories, data centers, sponsor offices, and vendor offices) simultaneously.
- Equipment. Manufacturing utilizes machines that are designed for particular purposes; calibrated and maintained to standards; and operated by skilled, trained workers. In a clinical trial, the machines are people--people who are not professional clinical trial subjects, people with health conditions, people with real lives that interfere with protocol schedules, people who can withdraw consent at any time. These people produce blood, urine, vital signs, and answers to questions that produce the data for the clinical trial.
- Processes. A manufacturing process utilizes a relatively small number of procedures and methods that are repeated exactly to produce products to exact specifications. A clinical trial has hundreds of processes, some of which are repeated iteratively and others that occur linearly. Each of the hundreds of facilities mentioned above has its own different set of processes; furthermore, each clinical study has its own set of study-specific processes that co-exist with the clinical site, sponsor, CRO, and vendor standard operating procedures.
- Vendors. A manufacturing process may have one supplier for drug substance, one for drug product, and a third for packaging and labeling. Each of those suppliers receives supplies from yet another chain of vendors. In clinical trials, the use of vendors is more like a service web than a supply chain. Clinical trial vendors work interdependently, rather than serially, continuously modifying their procedures and communicating their activities to avoid redundancy and ensure that nothing falls through the cracks.
- Oversight. A sponsor oversees manufacturing by reviewing and approving batch records, deviations, and CAPAs. A sponsor oversees a clinical trial by training vendors and site staff; attending hundreds of meetings; facilitating decisions on how deviations will be followed up; reviewing and approving study-specific procedures and documents; performing User Acceptance Testing on systems; reviewing metrics; and sampling work product (e.g., reviewing monitoring visit reports or essential document packages). In clinical trials, the sponsor works side-by-side with many of its vendors, performing some of the same activities (e.g., co-monitoring, project management, and data review) as part of oversight. (Imagine a sponsor walking into a manufacturing facility and pitching in with the labeling process; in clinical trials, this is routine.)
- Change. In both GMP and GCP, changes must be carefully managed, but in GCP, change is nearly constant. Protocol amendments are frequent; each amendment may affect study-specific plans, data acquisition systems, informed consent forms, and other study-specific documents. Regulatory approvals are slow, and teams frequently find they have not even finished implementing one change when a new change is introduced.
- Deviations. In manufacturing, deviations are an unfortunate reality; in clinical trials, deviations are nearly constant. Every day, patients miss visits; site staff fail to collect data; data are inconsistent. A trial with no deviations would be suspect due to the fact that our machines are human beings.
That's a lot of differences! Next we'll look at how these differences impact the effectiveness of the ICH Q10-model quality system in GCP.
