As a part of an ongoing discussion on the LinkedIn Network group (Quality-by-Design), I posted the following response. For the interest of people who do not participate in the LinkedIn Network or the particular group, I am also posting the response on the blog. I hope that you will find the post useful.
Bill:
With all due respect and with all my humbleness, I say that your post in response to my question about a fairy tale itself is a fairy tale. Do this or do that and fire some people, who file papers/data which were created by others, you know who, but keeping the lobbyists should hardly be considered as defining and describing the problem and let alone solving it.
Let me go one step further. In my view, a bigger and juicier fairy tale is QbD itself. It is an adult version of children’s fairy tales. Only it costs a lot more than children’s versions. Let me explain:
In our area (related to drug product quality), the quality of products, in particular oral products such as tablets and capsules, is determined by plasma levels of the drug. However, in most cases, the quality after initial stages of product development is established based on in vitro methods. The leading test in this regard is a drug dissolution test. People, who are not familiar with this subject, can take my words for it (considering my 25+ years of experience in all aspect of such testing) that it is the most simple test one can have in the entire science area perhaps after the procedure for taking body temperature. The idea behind this test is that if the drug dissolves (which we measure by this test), the drug will be absorbed in the body and will provide its intended efficacious effect. It is part of all GMP requirements and all national and international guidances and standards.
Now, is it too much to ask that anyone please show me that if the test, as it is performed, is indeed repeatable/reproducible and relevant to plasma drug levels? The answer is: There is absolutely no evidence available. In fact, there are tens, if not hundreds, of examples that exist to show that the test is not valid and it cannot predict the plasma drug levels with any accuracy. The FDA’s experimental studies confirmed this lack of an in vitro-in vivo link, formally known as in vitro-in vivo correlation (IVIVC). However, guidance has been developed, with the expectation that everyone should follow them to develop such IVIVC and predict plasma levels. Now the question is, how come we have the guidances and requirements of something when experimentally it has been shown that the test is neither valid nor relevant. Is it not a fairy tale? I think it is.
I often refer to our over indulgence with statistics, statistical modeling, or computing because, I believe, people think that if we have enough data and powerful statistical software packages, we may be able to show IVIVC. For at least 10 years, people and guidance have suggested and demanded such practices to show IVIVC. Furthermore, believe it or not, people make claims of successes by using complex and propriety software (black boxes) which I am almost 100% sure that they do not understand. Remember, it is scientifically and mathematically impossible to achieve such IVIVCs and use these (Inna please, note your belief of checking/verifying the underlying science). If you would like to know more about the bizarre practices of drug dissolution testing, IVIVC, or often commonly made claims (fairy tales) please visit the website (www.drug-dissolution-testing.com).
Recently, the FDA has released a QbD-based document for establishing the quality of products based on the very concept, you guessed it, IVIVC. They assume that if the title contains the word QbD in it, then the technique might work. I say; good luck. For my comments on the document, please see the link). I know of at least two upcoming conferences in the next few months on the very subject of developing IVIVC (storytelling), based on the same old concepts, instrumentation and procedures. What should we call these conferences? I will leave it up to you!
So, to conclude, believe it or not, we are telling adult versions of the fairy tales under the name of QbD. The mystery in our story comes from the statistical and computing parts. The story become interesting, mystical and sellable.