I’ve found a few WWW sites which summarise microbiological risk and criteria. 
Database of microbiological specifications  World Health Organization  Microbiological criteria in the EU  Operating characteristic curve   The above graph is the ’Operating characteristic curve for n=5, c= 1 to 3 (Fig 8.5a in the book). It illustrates that although a batch of food may be 30% defective there is a 52, 85 or 98% chance of accepting it using a sampling plan of 5 samples with a reject batch criteria (c) of 1, 2 or 3 respectively.
This clearly illustrates that end-product testing cannot guarentee a safe product ! Hence the need for HACCP and more recently microbiological risk assessment. 
Since the publication of my 2000 book the PHLS (UK) guidelines for ready-to-eat foods (Table 8.20, p333) has been modified already (Such is the pace of food microbiology and the need for web page up-dates!). The ’Aerobic plate count’ is now referred to as ’Aerobic colony count’’ and the Clostridium perfringens Satisfactory level is 20/g. The units for the Table should be per gram unless otherwise stated (ie per 25 g as for Salmonella etc.
The reference for the new PHLS ready-to-eat guidelines is Anon (2000) Guidelines for the microbiological quality of some ready-to-eat foods sampled at the point of sale. Communicable Disease and Public Health 3, 163-167. You can also download a .pdf file version by going to the PHLS (UK) web site and clicking on ’Food Sampling’.
In 2002 HACCP implementation for companies handling meat was released by the HMSO for Northern Ireland and includes swabbing sites and microbiological criteria.