Underestimating the complexity of pharmaceutical systems and their underlying raw materials is a good way to generate “Black Swans.” Black Swans are rare, unpredictable events that have an impact on other events. Their explanations seem obvious in hindsight. The term was coined by philosopher Nassim Nicholas Taleb, in his book The Black Swan: The Impact of the Highly Improbable, which was first published in 2007.
Although rare, Black Swans occur with greater than expected frequency, in what Taleb called the “fat-tail” distribution. We continue to see these events in the pharmaceutical industry. Even when we comply with current Good Manufacturing Practices (cGMPs), we still experience unanticipated quality problems in apparently robust pharmaceutical products, leading to out-of-specification (OOS) excursions, rejections, or recalls. What can we do about this?
If robustness is defined as the ability of a manufacturing process to tolerate the expected variability of raw materials, operating conditions, process equipment, environmental conditions, and human factors, perhaps the expectations were not appropriate. As documented in a PQRI white paper, "well understood, robust processes suggest greater process certainty in terms of yields, cycle-times, and level of discards." What uncertainties or unknowns might affect diligently controlled products and their underlying design assumptions?
Never underestimate complexity. Even the simple tablet assumes a homogeneous force transmission and resultant microstructure, which is the exception rather than the rule. The collective particulate properties of the compression mix are difficult to predict, and the composition and performance of the individual excipients are not always understood. Reliance on univariate experiments during development may miss interactions.
Unknowns add to system complexity, undermine risk assessment, and make it difficult to develop meaningful models. Excipient unknowns fall into several categories and may be further categorized as being unknown to the user, unknown to the supplier, and perhaps even unknown to both. Risk assessment requires that unknowns (not unknowable) be addressed with all stakeholders, including the excipient suppliers.
Excipient risk assessment starts with risk identification, defined in ICH Q9 as a “systematic use of information to identify hazards referring to the risk question or problem description. Information can include historical data, theoretical analysis, informed opinions, and the concerns of stakeholders.” Are excipient suppliers not stakeholders with informed opinions? They have the knowledge of variability and unspecified attributes, without which excipient risk assessment is flawed. Their application knowledge can identify potential excipient-related modes of failure.
Identifying as many potential modes of failure as possible and developing appropriate responses (elimination, mitigation, contingency, or no action required) is a good antidote to Black Swans. However, this adds to the complexity of the regulatory submission, and unless the resulting increased level of regulatory scrutiny is offset by the long-promised flexibility of oversight, the incentive is to minimize discussion of potential failure modes.
Perversely, the resulting focus on only a minimum number of critical material attributes or critical process parameters greatly increases the risk of failure.
Rearranging the deck chairs on the Titanic will not impart resistance to icebergs. The challenge for quality by design is to provide regulatory encouragement to the Black Swan hunters, exercising joint due diligence with all stakeholders. The burden of regulatory scrutiny should then fall on less informative applications, which in practice may be too good to be true. Absence of evidence of a problem is not evidence of absence of that problem.