Stacked Control Assessment (SCA) Approach for Production Optimization

Within the pharmaceutical industry, there exists combination products which combine aspects of medical devices and pharmaceuticals. For example, a Transdermal Drug Delivery System (TDDS) is a device that adheres to the patient’s skin and delivers active pharmaceutical ingredients (API) to the patient via passage through the skin.

In the process to manufacture a TDDS, much like most manufacturing processes, there are multiple unit operations with intermediate steps. The Stacked Control Assessment (SCA) approach consists of stacking the sources of variation for a particular finished product’s critical quality attribute (CQA), which is controlled by a product specification. In this article we will use the example of product potency as the CQA. Since product potency can change throughout a product’s shelf life, sources of variation to potency during shelf life must also be considered.

The SCA approach defines the largest potential sources of variation and represents a measure of process ability to meet specifications for each process step. The analysis also provides insight into the historical performance of raw material controls and provides visibility to which specification ranges can be tightened with the least risk to raw material specification nonconformance. This is a critical step in determining the overall control strategy. It encompasses all controls, both product and process related.

Specific to TDDS manufacturing, there are three main unit operations that provide sources of variation to product potency. These are the intermediate unit operations of Blending, Coating, and Pouching. Each of these unit operations have sources of variation for potency that will stack cumulatively in the finished product form, determining the overall stacked control for potency. Adding sources of variability throughout shelf life depicts the potential extent of potency variability at the end of product shelf life.

The SCA approach defines the largest potential sources of variation and represents a measure of process ability to meet specifications for each process step.

An example of SCA by potency variation stacking is provided to better understand the mechanism for the cumulative stacking approach. For simplicity, we will only describe the first unit operation equation in detail and later show the mathematical relationship accumulating the variation for each unit operation. An example for the sources of variation to blend potency is provided below in Figure 1.

Figure 1: SCA Equation for Blend Potency Blend Potency as % of Target = (API Assay * (API Weight (mg) /(API Weight (g) + Other Ingredient Weight (g))) / Blend Potency Target (mg/g)) * 100

The API concentration ratio represents the process variation caused by the addition of an API to the process kettle. We will assume for this exercise that the weights of all ingredients added were at target values making this ratio equal to 1. The raw material control in this equation is API assay. The acceptance criteria defined by the specification determines the potential variation allowed by this control. In this example, we will assume the allowable API assay acceptance criteria varies from 98-102% purity. According to the equation in figure 1, the ratio of API assay to target is utilized, therefore 0.98-1.02 are the potential inputs to the equation. In the case of the above equation, the raw material control of API assay has a direct 1:1 impact on blend potency if the correct ingredient weights are added to the blend. In other situations, there may be more than one raw material or process control affecting the CQA and these would accumulate mathematically to affect blend potency.

In the SCA, the minimum and maximum cumulative variation is calculated for a CQA. Each unit operation may have an intermediate specification in which the maximum and minimum cumulative variation can be compared. Ultimately, by continuing to accumulate the variation for all unit operations, these overall cumulative maximum and minimum variation values are compared to the finished product specification range (Figure 2). Further analysis of potency on stability would be conducted by stacking any degradant generation to the finished product equation.

Figure 2: SCA Equation for Finished Product Potency Finished Product Potency as % of Target = (Laminate Potency * Patch Area as % of Target) Where, Laminate Potency = Blend Potency * Coat Weight as % of Target

The difference between the potential stacked maximum and minimum values and the respective finished product upper and lower specification limits (USL and LSL respectively) can be calculated. This difference represents the potential for the product to not comply with specifications at any intermediate point and for the finished product. This analysis is displayed in Table 1.

Table 1: SCA Maximum and Minimum Variation Comparison to USL and LSL

Once all the data is input for the maximum and minimum variation ranges, a histogram depicting the largest contributors to potential variation via control ranges can be created to determine what sources of variation are the largest contributors (Figure 3):

Figure 3: Histogram for Sources of Variation

In parallel to the variation analysis using specification ranges, a historical maximum and minimum variation should be accumulated. This technique is beneficial as it allows visualization of the potential to reduce raw material specification ranges. The histogram, as shown in Figure 3, would provide the largest sources of variation first, which when coupled with the raw material historical performance to specification data, provides an effective mode to enhance the overall control strategy without increasing the chance of failing upstream raw material specifications. This assumes adequate historical data has been captured to accurately represent the true variation of that control. This is depicted in Table 2.

Table 2: SCA Maximum and Minimum Variation Comparison to USL and LSL with Historical Variation

Table 3 displays the final revised variation potentials after enhancing the control strategy using historical variation to tighten specification ranges. In this case coat weight and patch area controls were reduced to align closer with historical ranges.

Table 3: SCA Revised Variation Potential with Enhanced Control Strategy

In conclusion, SCA provides a low-risk approach for enhancement of control strategy that allows for greater ability to meet product specifications.