The sheer complexity of the number of steps (and, often multiple venues) involved often leads to interactive, but not integrated SOPs between several groups (jurisdictions). For example, when a warehouse operator secures a fiber drum of raw materials for a QC technician to take samples, which group is responsible for noting the sample has been moved from production? QC? Warehouse operator? QA? Further, where is the document stored and who then reviews it? Another question begs asking: If the operator is to (representatively) sample the finished product or intermediates, and it is stored in multiple containers, does he or she need (yet another) SOP on how and where to sample? And, for final assay, how representative are, say, the 20, 30 or more tablets taken from 20 to 30 fiber drums?
More modest Pharma companies tend not to have the deep capital reserves necessary to keep up with advances in control and automation technology. The first attempts at placing measurement devices along the process stream (in the name of PAT) were intended merely to take measurements without using these data to control process. In other words, PAT was not achieved. The process analyzers merely churned out more numbers to store for later analysis. Figure 1 shows that the work flow changes somewhat, but is barely improved when instruments are added, but not integrated.
LET’S GET INTEGRATED
As process steps become integrated using a central controller, where the intermediate product flows from step one to step two as a smooth flow (based on instrument readings), two things happen. First, the information flow now resembles Figure 2, where all data are both archived and acted upon instantaneously. Second, the operator is taken out of much of the process. Since data is automatically logged, there are no transcription errors and fewer problems for QA to worry about (or write yet another SOP to manage it). When the intermediate product is measured automatically, samples need not be taken, logged, assayed, etc., at points where each step is susceptible to errors and requires more SOPs — ones likely to be misunderstood or ignored. The smaller number of SOPs and operator interventions will almost immediately cut down on the need for complicated written procedures.
An integrated PAT system allows materials to flow from the dispensary through the processing steps to the packaging line without being taken to a storage facility. This will not only increase speed of production, but generate more process data as well. We now know and can prove that step 1 is good before the material moves to step 2. The difference is that several days/weeks have been reduced to several seconds. The results are automatically stored (taking decisions out of human hands) and, more to the point the company is following cGMPs’ recommendation of proper (my emphasis) in-process testing and a statistically significant number of final dosage forms (not the 20 out of 5,000,000 units as most now assay).
As any relatively aware person in the Pharma industry understands, there are consequences for producing and selling poor-quality products. However, such consequences are usually financial and seldom directly affect managers, the ones setting a poor example for those he or she supervises. Unless and until penalties reach further down the corporate ladder, the game of “beating the odds” is likely to continue. When, for example, there are six or seven state troopers patrolling 200-300 miles of an interstate, most people play the odds and speed, assuming they won’t get caught. Some Pharma executives may be playing the same game — if the penalties get high enough, the incentives to cheat diminish accordingly.
Spoiler alert: The end game for PAT (and of course, QbD) is continuous manufacturing (CM). Under the paradigm of CM, production is performed on small, development-sized mini-batches, each being controlled to be exactly the same as the last (see Figure 3). Compare that to the “traditional” manner of product development through which production is a step-by-step development, with each step scaling up in size from development through pilot plant to production.
The complexity and number of people involved almost assures there will be paperwork errors, if not actual procedural errors. When the development, pilot and production size batches are the same, the chance for quality and compliance errors are greatly diminished. When lot-after-lot can be manufactured the same every time, the product, itself, can be continuously improved and consistently produced.
From a training standpoint, a constant level of equipment will make it easier for operators to become proficient in their jobs and not have so many procedures to follow (or misread or outright ignore). With fewer types of equipment to learn to operate and clean, fewer mistakes will be made, leading, again, to higher standards of quality. So if we are aiming for the highest quality and tightest compliance in the Pharma industry, CM will be the end game.