The eleven-member task force included numbers from Australia, Britain, Canada, Ireland and the U.S.A. The primary interests of each of the members differed, but all had a wide interest in and familiarity with the processing of raw meat, from the slaughter of animals and dressing of carcasses to the preparation of retail-ready products or meals.

The task force was asked to consider six questions. When addressing each question the task force was asked to first decide if an answer agreeable to all could be formed. If not, then the task force would be asked to identify the natures of any disagreements and to indicate the type of information that would be required to resolve each point of disagreement. That procedure was proposed to obtain some indication of likely fruitful lines of research in any area where agreement would at present not be possible.

In traditional meat inspection it was assumed that microbiological and visible contamination are related, and that the microbiological condition of meat could therefore be controlled by preventing or removing visible contamination from meat and meat contacting surfaces. That view has been continued in recommended procedures for implementing HACCP systems in beef packing plants, which are mostly based on subjective assessments of the cleanness of facilities, fixtures, equipment and product.

Obviously, if the microbiological condition of product can indeed be adequately assessed, for practical purposes, by reference to visible contamination, then microbiological testing is not necessary for assuring the safety of beef.

The task force agreed that actions for controlling the microbiological contamination of meat could be usefully guided by consideration of visible contamination on product and equipment. However, they also agreed that apparently clean product could still be heavily contaminated with bacteria. Thus, appropriate actions to control microbiological contamination can be proposed by reference to visible contamination and, during routine production, appropriate actions can be maintained by reference to visible contamination. However, whether or not a proposed controlling action is appropriate can be decided only by reference to microbiological data. Thus, some microbiological testing is necessary if the safety of beef is to be assured.

Microbiological testing can be carried out for the purpose of surveillance, acceptance or rejection of batches of product, or implementation of HACCP systems.

Surveillance sampling is seen as a means of obtaining information which can be refered to for setting policies and standards, comparing regional and seasonal performance, detecting changes in performances over time, etc. The task force agreed that such sampling could be useful for answering questions of the indicated types, but that surveillance studies were by no means essential for assuring the safety of beef.

Acceptance sampling is the form of sampling that most assume should be used for assuring the safety of raw meat. In such sampling, batches of meat are tested for the presence of specific organisms. If those are found in unacceptable numbers, or found at all in some cases, the batch is rejected. The task force agreed that such testing was of little use for raw meats, because failure to detect a rare organism, such as Escherichia coli 0157:H7, gives no real surety that the organism is not present in the batch. Therefore acceptance testing cannot be used to reliably identify and reject contaminated product.

It has long been recognized that the safety as well as other qualities of product cannot be effectively assured by batch testing of the end product. Instead, production processes must be controlled to assure that only product of an acceptable quality is produced. A HACCP system is a special case of process control where the quality to be assured is safety. For the implementation of an effective HACCP system, the Critical Control Points in the process must be validated, and the performance of the controlled process must be periodically verified. Therefore, for raw beef production processes, the points in each process where microbiological contamination must be minimized or is removed must the identified from microbiological data, and the microbiological quality of the product leaving the process must be determined, from time to time. The task force agreed that the microbiological sampling of beef for those HACCP implementation purposes is essential if the safety of beef is to be assured.

The bacteria that may be sought in microbiological sampling are of two broad type; pathogens, which are the organisms that cause disease; and indicator organisms, which are indicative for the possible presence of specific pathogens. Pathogens are usually recovered from beef only sporadically and in small numbers. Thus, any sampling for pathogens usually results in long runs of negative results, which provide little information about process performance. In contrast, indicator organisms can usually be recovered in countable numbers, and changes in the numbers recovered can identify the microbiological effects of procedures for controlling processes. Therefore, the task force agreed that microbiological testing for HACCP implementation should involve the enumeration of indicator organisms, not the detection of pathogens. The indicator organisms that might be used include total aerobes, coliforms, generic E. coli, aeromonads and listerias. Understanding and control of the microbiological effects of meat plant processes requires the enumeration of more than one, preferably several types of indicator organism.

The task force agreed that when seeking sources of microbiological contamination it is necessary to obtain samples from equipment and the environment as well as from product. Samples from equipment and the environment should include obscured areas, where detritus may accumulate and not be reliably removed during cleaning, as well as meat contacting surfaces.

Because of the uneven distribution of bacteria, large samples are preferable to small. The largest samples that it is practicable to collect and process should therefore be obtained . The task force agreed that sampling methods have to be varied to take account of the nature of the material that is being sampled and the various other practical constraints on obtaining samples from product which is being processed, or from equipment which is being used or cleaned. However, for carcasses, cuts and equipment, swabbing of large areas with gauze or sponge, and without delimitation of an exact area, would often be the optimum method.

Question 5. How should samples be processed?

The task force agreed that counting rather than just detection of indicator organisms is necessary. When indicator organisms are numerous, usual plating procedures may be employed. However, when indicator organisms are few it would be desirable to adopt most probable number procedures, which allow enumeration of indicator organisms at numbers as low as one bacterium per sample. One such procedure involves the filtering of samples through a membrane filter which is divided into numerous squares by a hydrophobic grid. All the fluid from swab samples of areas of 1000 cm² or more can be filtered to obtain counts at the level of one organism per 1000 cm². Thus, suitable procedures for counting bacteria at very low numbers are available, although as yet they are little used in systems for assuring meat safety.

Surveillance sampling plans must be devised with some knowledge of the microbiological condition of product, and with a view to answering specific questions about the occurrences of the bacteria of interest. Unfortunately, surveillance sampling to date appears to have been carried out without proper consideration of how the data might be used. Consequently, the task force agreed that comparisons of processing performances as, for examples, between countries or before and after the mandating of HACCP implementation at U.S. packing plans, on the basis of currently available surveillance data are not scientifically legitimate.

Sampling plans for validating and verifying HACCP systems could involve the consideration of attributes or variables. In attributes sampling plans, bacterial numbers are not considered as such. Instead, each number is regarded as the attribute of being above or below one or two defined numbers. Thus, each sample is assigned to one of only two or three classes, and the variability of the numbers recovered from samples is ignored.

With variables sampling plans, numbers are considered as such, and the variation between samples is taken into account. Variable sampling plans can be based on the fact that the distribution of bacteria in or on meat is approximately log normal. That is, the logarithmic values of bacterial counts are normally distributed and so can be analyzed by usual statistical procedures.

The task force agreed that variability must be taken into account when analyzing microbiological data for process control purposes. Therefore, variables sampling plans should be used in the validation and verification of HACCP systems for beef production processes. Conversely, attributes sampling plans are not compatible with HACCP implementation. It is therefore regrettable the current microbiological performance criteria for meat production processes are based on attributes sampling plans.

Despite some differences over detail and emphasis, the members of the task force were in broad agreement about all aspects of microbiological testing for assuring the safety of beef. In summary, they agreed that the proper purpose of microbiological testing is the implementation of demonstrably effective HACCP systems; that HACCP implementation requires testing for indicator organisms rather than pathogens; and that, for HACCP purposes, data must be assessed by reference to variables not attributes sampling plans. Those conclusions are largely agreeable with and extend the findings of a recent American Meat Science Association symposium on the same subject. It therefore seems that scientific opinion is rapidly approaching a consensus on the proper practice of microbiological testing for assuring the safety of beef, and that the scientifically proper practices are greatly at variance with the microbiological testing that is currently required by regulatory authorities.

A detailed account of the task force’s discussion is being prepared. That account will be submitted for publication in an appropriate, international, scientific journal.