Is a sampling plan based on the (√n+1) rule acceptable for determining the correct number of samples? The Canadian health authority (Health Canada) listed this question in the section "Good Manufacturing Practices (GMP) Questions and Answers" of their website and answered it as follows:
Sampling plans and procedures must be statistically valid and based on scientifically sound sampling practices.
Among the important issues to be considered regarding the risks are:
A sampling plan based on (√n+1) may generally be acceptable, e.g. for a large number of containers. However, such a sampling plan may pose a considerable risk of accepting defect goods when sampling a small number of containers, for example.
A documented justification is expected for all sampling plans.
Strictly speaking, these requirements only apply for der Canadian market, they can, however, be viewed as state of the art worldwide.
The question with its complete answer can be found in section "Raw Material Testing - C.02.009 & C.02.010" under Q7 - see
"Good Manufacturing Practices (GMP) Questions and Answers / Is the sampling plan based on the (√n+1) acceptable for identifying the number of containers of raw material to be sampled?"