Predictive microbiology is a description of the responses of microorganism's to particular environmental conditions such as temperature, pH and water activity. Predictive microbiology utilises mathematical models (built with data from laboratory testing) and computer software to graphically describe these responses.

Predictive microbiological models do not replace laboratory analysis or the training and judgment of an experienced food microbiologist. Predictive microbiological models must be used with great caution and only used by trained, experienced personnel with an understanding of the limitations of use.

In all predictive microbiology a prediction must only be used as a guide to the response of microorganism(s) to a particular set of environmental conditions. Consultation with a competent body is strongly recommended before their use. Food businesses should never rely solely on any predictive microbiological model to determine the safety of foods and/or processing systems. Determining the growth, survival or inactivation of pathogens in food requires:

1. The determination of the intrinsic and extrinsic properties of the product, taking into account the storage and processing conditions, the possibilities for contamination and the foreseen shelf-life.
2. Consultation of available scientific literature and research data regarding the survival, growth and inactivation of microorganisms of concern.

Where necessary on the basis of these studies food businesses should also conduct additional studies, which may include:

1. Laboratory-based microbiological sampling and analysis
2. Predictive microbiological modelling
3. Challenge tests to investigate the ability of microorganisms of concern to grow or survive in the food product under reasonably foreseeable conditions of distribution and storage.

General Information

Laboratory-based microbiological tests are typically used to make the critical decisions regarding food safety and product shelf-life. However, the growth, survival and inactivation of microorganisms in foods are reproducible responses.

Predictive microbiological models which quantitatively describe the combined effect of specific environmental conditions can be used to predict growth, survival or inactivation of microorganisms. In product development, a predictive microbiological model may allow a food business to evaluate the safety and stability of new formulations and identify those which may give a desired shelf-life.

Predictive microbiological models are also useful when the shelf-life has been determined, but the product is then subject to a minor process or formulation change (either planned or unplanned through loss of process control). A predictive microbiological model can then be used to initially establish if the change might have any effect on the safety and shelf-life of the product.

Predictive microbiological models allow product developers to pinpoint the combinations of hurdles that may achieve a desired shelf life. These specific conditions can then be tested by experiment if necessary thus reducing the cost of challenge studies.

Development and Limitations of Models

Predictive microbiological models are normally developed assuming that microbial responses are consistent. While predictive models can provide a cost effective means to minimise microbiological testing in determining shelf-life, there may be occasions when the model's predictions may not be accurate, due to inconsistent microbial responses and variations in the growth media. Research has indicated that this is often why some predictive microbiological models fail to accurately predict the survival, growth or inactivation of pathogens in food products.

Furthermore predictive microbiological models must undergo validation before they are used to aid in food safety decisions. Validation involves comparing model predictions to experimental observations not used in model development.

Microbiological Modelling Programs

Initiatives to develop microbiological modelling programs have been ongoing in the United States, the United Kingdom, Denmark, France, Australia and other countries for a number of years. These programs have resulted in the development of a wide range of microbiological modelling software packages becoming available on the internet for download. Some of the more commonly used models are listed below:
 
1. ComBase
 
2. Pathogen Modeling Program
 
3. The Food Spoilage & Safety Predictor
 
4. Sym'Previus (an integrated database and predictive software)