The Analysis button becomes available when a sampling plan has been chosen or newly-defined, and at least one contamination profile has been selected to be subjected to the plan. The Analysis Tools page displays the selected sampling plan and contamination profile(s), and offers 4 analysis options.
“Analyze Plan Performance” prompts the tool to produce an Operating Characteristic curve (“OC curve”), which relates each concentration along the x-axis of the between-lot distribution to a probability of acceptance (or rejection). The concentrations of the between-lot distribution represent arithmetic mean hazard concentrations (on the log10 scale) of the individual lots that make up the distribution. Each of these mean concentrations can be thought of as a “Detectable Microbial Load” (DML) that is associated with a particular probability of rejection. Using either the OC curve, or the look-up table provided (for common percentiles), or the compute button provided, the user can easily find the DML associated with any given probability of rejection, and vice versa.
Note that the “Target Probability of Rejection” defined for a sampling plan is used only to produce the value for “Detectable Microbial Load” shown on the PDF Report and on Sensitivity Analysis results.
No inputs are required.
“Design a Plan to Meet a Target” permits the user to adjust the value of one parameter of the sampling plan to attain a desired probability of rejection for a given DML (mean concentration). For example, if a desirable rejection rate for a lot with a mean concentration of -1 log10 cfu/g is 95%, but the current plan produces an OC curve that predicts a 75% probability of rejection for that concentration, then it is necessary to alter the sampling plan so that the resulting OC curve will provide the desired rejection rate. Changing the sampling plan, for example the number of samples or the value of “c”, will change the OC curve.
REQUIRED INPUTS
Target Detectable Microbial Load, in log10 cfu/g
Desired Rejection Rate, in percent
Parameter to Vary (from a drop-down menu of available parameters)
“Study Impact on Microbial Load” requires that a between-lot distribution has been defined. The model applies the probabilities of rejection associated with the OC curve to the concentrations along the between-lot distribution, and predicts the proportion of sampled lots that will be rejected and at which concentrations. From the consideration of the microbial load of the rejected lots as compared to the microbial load in the original distribution, the proportion of microbial load remaining is calculated. This is a measure of risk reduction.
The complete list of results generated in this option is:
- Mean Concentration Pre-sampling
- Mean Concentration in (tested and) Accepted Product
- Mean Concentration in Final Product
- Percentage of Microbial Load Remaining
- Percentage of All Lots Rejected
- Percentage of Tested Lots Rejected
No inputs are required, as the percentage of lots sampled was defined in the sampling plan.
“Perform Sensitivity Analysis” permits the user to see the impact of varying an individual parameter, on both the the OC curve and the risk reduction measures (if a between-lot distribution has been defined).
The parameters that can be varied are Number of Samples, Acceptable Positives, or Analytical Sample Size in the case of presence/absence plans, and Number of Samples, Analytical Sample Size, Acceptable Number of Concentrations within Limits, Small-m, or Big-M in the case of concentration-based plans.
The user defines the range over which to vary the parameter value, and the number of values (“steps”) to be considered within that range. Results are calculated for each step. The OC curves can be shown separately or superimposed on a single graph.
REQUIRED INPUTS
Parameter to Vary (from a drop-down menu of available parameters)
Minimum value of the parameter
Maximum value of the parameter
Number of Steps
Include OC Charts (check-box)
Separate OC Charts (check-box)
Reporting