Thursday, June 18, 2009

Drug manufacturing – relationship with PAT and MVA

FDA in the USA is a regulatory body which determines which drugs are favorable for human consumption and which not, but have you ever thought how pharmaceutical companies manufacture drugs adhering to all these standards by bringing the rejection rate to a bare minimum. Well we all say ‘to err is human’ but has anyone thought about how does pharma companies go about doing their work .No!!!…then just read on

Process Analytical Technology (PAT) is a very common term used in the Pharmaceutical sector, now the main advantage here is that it helps in detecting faults very early and also helps in determining the quality of a batch even before it’s completely produced. Now don’t stop here by screaming….Eureka!!!! Because this is just the tip of the iceberg and you still have a lot to understand.

Generally while manufacturing a drug the main feature that should be kept in mind is the conditions in which the process takes place and the stability of the final product. In the past what used to happen was that a drug was developed and then it went for a laboratory analysis to verify the quality. But what is really needed here is a good process that should be in place and that too without flaws so that the moment the final drug comes out we are sure that this product is our desired result and that we will not need to discard it since it matches our quality parameters.

The most important department is the Quality control as they must be present right from raw material collection to the final moment where the drug is packed. They look into the safety of every thing; random samples can be tested to check hardness, colour and other important features.

Now the main problem with the primitive form of testing is that we check to see if the final drug is good only after a batch is manufactured, so there is always a chance of a complete batch rejection in case a randomly tested sample does not pass QC .So it is this factor that has made the FDA encourage companies to use Process Analytical Technology (PAT) in pharmaceutical production and quality control.

Now PAT is a method which is based on timely measurement of critical quality parameters and other performance attributes so that the end result is having the desired quality and is acceptable. They involve optimal applications of process analytical chemistry tools, feedback process control strategies etc

Quality what I feel doesn’t come to a product from space but it should be incorporated into it i.e. it should be built in or should be by design. PAT actually helps to shift focus from the conventional off-line testing i.e. having a separate lab for testing to scientific and in-line testing.

Now in-line testing basically means infrared (NIR), Raman, Mid-IR spectroscopy, acoustic emission signals, and other imaging techniques, or other physiochemical techniques as a primary means of process monitoring. NIR as a method is very useful because while testing using other techniques there is a high chance of a reaction with the chemicals in the existing drug while NIR on the other hand has high penetration power and does not destroy or react with samples. Sampling error would be minimized with in-line probes placed strategically through out the production process.

So the advantage here is that we not only can have a continuous monitoring of the quality but can do timely adjustments as well to make sure the final product is of high quality and is of desired specification.

Chemometrics is the application of mathematical, statistical, graphical or symbolic methods to maximise the chemical information, which can be extracted from data. An important concept in chemometrics is Multivariate Data Analysis (MVA), which can be defined as considering multiple variables simultaneously. Multivariate Data Analysis includes the use of Principal Component Analysis (PCA), Multiplicative Scatter Correction (MSC), Partial Least Squares Regression (PLS), and Standard Normal Variate Transformation (SNV) and once the data from in-line testing is obtained, MVA finds relations between columns in data tables

The objective is often to use one set of variables (columns) to predict another for the purpose of optimisation and to find out which columns are important in the relation. There can be missing data in the data table and if they exist it would mean less information is available in the table to analyse.There are a lot of softwares in the market right now that helps us analyse these data like The Unscrambler from CAMO Software and other companies like Umetrics, Matlab etc also have similiar products.

So we can finally say that PAT as a technique is very useful since it helps us in saving a lot of time and money by reducing the number of rejected batches and prediction of quality of batch even before completion. Where as batch process on the other hand will work only after a batch is completed and if the red light blinks i.e. the rejection sign pops up then well it would mean a huge loss for the company

Wednesday, June 3, 2009

Is this something related to Rocket Science ?



Well the Answer is a BIG NO!!!!! . This is really an interesting field because…….


Data plays a critical role no matter in which ever fields you are working .But how many of you analyze this data or how many of you really find meaning out of this data? Well I would say only very few people really do that. Now we need to find a meaning out of our collected data because that actually reflects our business .A well collected and analyzed data would always push a business to its upswing while the others would who neglect or don’t try to interpret them finally end up as fools


Properly analyzed data is called as information and for any business, information is wealth and it’s a very critical success factor. So how to convert data into wealth is the next big question and it’s all by doing good analysis and this is where multivariate data analysis comes into picture. Now many people reading this would feel it’s something like rocket science to decode your data to know what it means. Let me tell you that even my 15 year old son can do it; all he needs is proper guidance.


So how to analyze data is left to you, either you go manually or use some software to help you out but finally the most important thing is proper interpretation of your result because this interpretation is something that stands between success and failure. I have seen numerous advertisements from different people or organizations claiming to give a 1 week or 5 day course on Multivariate Data Analysis, just make sure that you don’t miss it because it can help you in many ways .


This field of Multivariate Data Analysis is not just restricted to Chemicals, Food and Beverages, Energy or Pharmaceuticals but the scope is really vast and I would say the more deeper you go, you can see with each further step the depth is only increasing I mean the knowledge that you gain .


We have always heard about children who went missing get reunited with their family after 20yrs or so and the question is how the parents recognized their child after such a long time. Well everyone can say DNA sampling did the trick but not many are aware that a Multivariate Analysis can be used for facial recognition, all we need is to mark anthropometric landmark points on the face and Wallah: your trick is done.


We all would have heard the story of Sharbat Gula , her picture was taken by National Geography in Afghanistan during the war when she was only 13yrs old but later on the same photographer went in search for her and was able to capture her frame again when she was 30yrs old, all tests like retina scanning and others were done to prove that this is the same girl that they are looking out for .Even by multivariate analysis this could have been done so just think about what this magical subject can do for you.





Even in the medical field Multivariate data analysis can help you in a big way .So all I would say is that possibilities are unlimited here.

Well I will be back to update you more about the Multivariate data analysis later ………. so don’t forget to come back