»
LARGE MOLECULE
» Figure 1. SSL Topical Product: Process Mapping Example.
and how an FMECA should be performed. Instead, discussion is only provided to interpret the information in Table 3.
In this case, S, O, and D were expressed numerically on a scale from 1 (low risk) to 10 (high risk). The RPN is a product of S, O, and D, also expressed in numerically similar low, medium, and high scales. The FMECA shown above is not necessarily representative of an initial risk assessment for a mixing process. If this were an initial risk assessment for which the investigator had little understanding, knowledge, and lack of experimental data, the risk scores may be higher. In that situation, the investigator would perform a risk assessment to the best of his or her ability, perform experiments, and gather data and information to minimize the risk and repeat the risk assessment exercise until the risk is controlled to an acceptable level. It is paramount that a sponsor clearly communicates risk assessment details to regulatory agencies. Beyond that, there is
Unit Operation
Parameter Potential Failure Mode
Mixing speed Final mix time Fill level Mix scale Temperature Too slow or fast Too short or long Too low or high
much fl exibility in which risk assessment approaches are selected to communicate that information. What is important is to communicate clearly and completely that which is considered low or high risk and the reasons why.
Statistical Design of Experiment
Let us assume that in the FMECA for the mixing process above, the RPN value for the fi nal mix time had a high risk of causing a failure that would negatively impact the drug product viscosity. In this case, it would be appropriate for the process engineer to design experiments to gain an understanding of the eff ect of mixing time on viscosity. The same principle applies in a situation in which blade speeds aff ect viscosity. Now assume that both time and speed have the potential to impact
Impact of Change
Potential Cause of Failure
Equipment, operator Operator Operator Operator
viscosity, or even multiple CQAs such as dose uniformity and viscosity, as it is in this case. In this situation, where multiple factors have the potential to aff ect one or more responses, it is appropriate to use statistical DOE. Depending on the overall process map and business drivers for a project, it may not be possible to perform a DOE due to a lack of resources, a lack of management support, or other reasons. However, if designed properly, DOEs are the most effi cient way to perform a multivariate analysis on a system with potential main and interaction of factor eff ects on responses. In the context of this discussion, both material attributes and process parameters are considered to be factors of the DOE and drug product quality attributes are referred to as responses.
Before designing an experiment, whether DOE or not, a strategy should be determined. Experimental drivers and primary and secondary goals should be established before executing any experiment. The primary goals for the DOE designed for the container fi lling process in this case study were to identify and understand main and interaction eff ects of process parameters such that CPPs could be established and ranked, and to identify CPP ranges that would result in acceptable product. Although identifying CPPs is a requirement for manufacturing R&D processes, ranking the criticality of key process parameters is not mandatory for every situation, but may be useful depending on the specifi c process or formulation risks.
Table 3. SSL Topical Product: Mixing Process Risk Assessment Example Controls
Dose uniformity, viscosity, degradation
Dose uniformity, viscosity, degradation
Dose uniformity or over spilling
Too small or large Dose uniformity Too low or high
Dose uniformity, degradation
BPR: inspection 5 BPR: inspection 4 BPR: visual BPR: visual Friction, ambient temp Visual
Potential Failure Mode: communicates how the CPP could cause a failure. Impact of Change: communicates how a failure could impact drug product CQA. Potential Cause of Failure: communicates what could cause a failure. Controls: communicates what controls are in place to prevent a failure. S = Severity of the failure, expressed numerically. O = Occurrence, or probability of a failure occurring, expressed numerically. D = Detectability, or how detectable is the failure, expressed numerically. RPN = Product of S × O × D. Risk Rating = Communicates qualitatively how high or low the sponsor considers the numerical RPN value.
5 5 4 4 2 1 1 2 S O D RPN 3 3 3 3 3 60 24 15 15 24 Risk Rating Medium Low Low Low Low
20 |
| January/February 2015
Mixing Process
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76