Validation

Validation of control measures means obtaining evidence that a control measure, or a combination of control measures, if properly implemented, is capable of controlling the hazard to a specified outcome. It answers one question: can this work?

Validation happens at two levels. Control measures and their critical limits are validated during the development of the plan. The plan as a whole is validated before it is implemented. Codex says this means making sure that the following elements together are capable of ensuring control of the significant hazards: identifying the hazards, CCPs, critical limits, control measures, the frequency and type of monitoring, corrective actions, the frequency and type of verification, and the information to be recorded.

Validation could include a review of scientific literature, mathematical models, validation studies, or guidance developed by authoritative sources. Businesses may not always need to conduct or commission studies themselves. Limits could be based on existing literature, regulations, guidance from competent authorities, or third-party studies, such as those by an equipment manufacturer. Further detail is in the Codex Guidelines for the Validation of Food Safety Control Measures, CXG 69-2008.

UK guidance gives good examples. The FSA's cooking equivalents, such as 70°C for 2 minutes, are authoritative sources for a time and temperature. FSA guidance on less than thoroughly cooked beef burgers, which covers England, Wales and Northern Ireland, expects the cooking method to be validated to achieve at least a four-log reduction, which is 99.99%, under worst-case conditions such as thickness, fat content and cold spots on the grill.

If you use outside guidance, make sure it fits. Codex says care should be taken to ensure that limits fully apply to the specific operation, product or group of products. Guidance on a time and temperature does not tell you how long your thickest pie takes to reach it in your own oven. Cooling data from a shallow tray tells you nothing about a deep one.

Next, gather evidence in operation. During the initial implementation of the system, evidence should be obtained to show that control can be achieved consistently under production conditions. In our example, the team places a data logger at the centre of the deepest tray, in a full chiller, on a busy day, and repeats this over several batches. If the worst case meets the limit, the team has good evidence.

Do not confuse validation with records. A probe record showing 75°C proves that the limit was met. It does not prove that 75°C is the right limit for that hazard. The first is monitoring or verification. The second is validation.

Revalidate when things change. Any change with a potential impact on food safety should lead to a review of the system and, when necessary, a revalidation of the plan. Reducing salt in a recipe, deepening a tray, changing an oven or extending a shelf life may each mean that the old evidence no longer applies.

GHPs are treated differently. Codex says validation of GHPs is generally not carried out by the business. Validation data from authorities, published literature or manufacturers is adequate. A disinfectant maker validates its product, and it is generally sufficient to use it according to the instructions. But the business should be able to show it can follow those instructions.

Keep a validation file that holds the evidence behind each critical limit. When an officer or auditor asks why you chose a limit, you can answer in a minute. The answer 'we have always done it this way' is not validation.