Gertrude Stein revisited
Richard Prince
A rose is a rose is a rose. But, Gertrude Stein would have a hard time succinctly describing what passes for pharmaceutical-styled validation. After all, what crystalline phrase can we conjure up to aptly describe, say, that a system is working reproducibly? Is it validated or qualified?
A “validated” manufacturing system or testing system is understood to mean that such systems are working as intended through the generation of predicted (i.e., pre-defined) outcomes. This is accomplished by means of the installation and maintenance of suitable engineering, manufacturing and quality control/quality assurance controls. A sterile manufacturing system will consistently produce sterile product batches if all steps in the manufacturing process are properly executed. How many successful, consecutive trials are therefore needed to ascertain that such a system or process should be deemed validated? The Food and Drug Administration (FDA) in a May 1996 Federal Register notice stated that an analytical release test method, to be considered validated, should be shown to give the predetermined results three consecutive times. At first blush, this seems reasonable and practical. Indeed, demonstrating process or system reproducibility provides initial evidence of scientific robustness as well as satisfying regulatory requirements. But, is it necessarily axiomatic to proffer that 3x reproducibility equals a state of validation? Do we delude ourselves into believing that true validation can be proven so relatively cheaply? Possibly so.
It is fair to state that merely demonstrating that some desired outcome has been shown to be reproducibly obtained does not mean that future outcomes emanating from the same system or process will be consistently and uniformly the same. Firms, for obvious reasons, do not have the luxury of conducting statistically valid protocols for systematically assessing a process`s basic rigor, such as is the case when gearing up for new product launches. Firms instead must show that their “validation” approach is in alignment with regulatory expectations and industry practice, e.g., the “rule of 3.” One nice feature of performing (default) retrospective validation activities is that they can provide more statistically meaningful determinations of the reproducibility of a given system or process. Proving, for example, that the last 30 batches consecutively manufactured passed all release testing conditions gives pretty good proof that a system or process is robust. This comes closer to a more idealized definition of validation, i.e., accumulating substantial evidence of process/ system reproducibility as benchmarked against pre-defined specifications. Thus, that which is reproducible is necessarily predictive. And that which is predictive constitutes proof of validation.
If one wishes to initially demonstrate process reproducibility, what name should be used to describe this activity? Some use the term validation. Others use the term qualification. I suggest the term “process readiness.” This would permit firms to demonstrate the intrinsic validatability of a given process or system. This presupposes that the specifications are rationally set. The imprimatur of a validated system would come way down the line, long after regulatory approval was garnered, e.g., pre-approval inspection. The firm could then issue a certification statement to the regulatory authority formally declaring, after the requisite accumulation and analysis of data, that a process has been proven to be truly validated on the basis of substantial and impressive historical data. This could then translate into an enlightened regulatory authority exempting the validated system from future on-site regulatory scrutiny.
Pharmaceutical scientists establish (commercial) sterility assurance by various mathematical equations. We use terms like sterility assurance level (SAL) and probability of a non-sterile unit (PNSU). The outcome is relative, not absolute. We really can`t absolutely say that a parenteral batch is truly sterile, although we have pretty good confidence of such. Conversely, physicists are able to reduce the mystery of the universe into equations that are fundamentally and absolutely predictive, e.g., e=mc2, law of gravity. Talk about validated systems!
While a rose is a rose is a rose, it can be argued that a validation is a verification is a qualification … in the sense that these latter activities all seek to establish that the work activity at hand will deliver or yield predicted outcomes. But, it is really validation with a lower case `v`, not a capital `V`. Invoking and slightly rewording Gertrude Stein`s equally famous phrase regarding the City of Oakland, California, it can be said of our current notion of validation: there isn`t much there there.
Richard Prince, Ph.D., is president of Richard Prince Associates Inc. (Short Hills, NJ), a compliance and technical based consultancy, and an officer of Microgen, a provider of quality disinfectant-cleaner products. He can be reached at (973) 564-8565, fax at (973) 564-8731 or by e-mail: [email protected].