 # Sampling in Pharmaceutical

Plz define what is sampling and there types, sampling during PV batches and what is stratified sampling, can we use it in PV batches??

Sampling is the process of selecting individuals from a defined population. There are two groups of sampling technics, probability and non-probability (or random and non-random samplings).

• Probability sampling technics are random samplings, meaning that each individual of the population has the same probability of being selected.
• Non-probability sampling technics are non-random samplings, meaning that indivuals are selected based on specific criteria, intentionally or deliberately.

There are 4 main random samplings:

• Simple random: all individuals are selected utsing a randomized selection method.
• Systematic: firs individual is randomly selected, then you select one individual every n’th individual, like on sample every 50 individuals.
• Stratified: population is divided into defined subsets, and then individuals are randomly selected from each subset.
• Cluster: population is divided into defined subsets, and then entire subsets are randomly selected.

For Validation propuses, we prefer to use probability samplings because they allow us to get unbiased subsets that are representative of the population (process); but we also use non-probability sampling, for example: sampling process using worst case products, materials or parameters; here we use a purposive sampling (non-probability sampling) to deliberately sample the process on its worst case conditions.

From random samplings, the systematic sampling is specially useful for validation since it allow us to select samples in a chronological order, to be able to assess process statistical control and inferential statistics like process capability, control limits, trending, etc.

That being said, you can definitely use stratified sampling when you can divide the population into specific subsets, like processes with several nozzles, powder dispensers, tablet machines, etc. In these examples you would prefer using a combination of stratified and systematic random samplings to be able to sample all different dosing nozzles or dispensers in a chrinologcal order; for example:

“Take 1 sample from each filling nozzle, every 20 items filled”