PARAMETER: Physiological status.
PARAMETER DESCRIPTION: It represents the proportion of the surviving population after chilled storage that may be irreversibly injured and, thus, is not culturable. This injury is additional to the one accounted for by the predicted model and could be attributed to the combined outcome of cold and osmotic stress (sudden aw downshifts), experienced by cells while residing on the carcass micro‐environments (niches)
Method: 72 h chilling is obtained extending the cold storage beyond 24 h, therefore uncertainty distributions of the physiological status parameter after 72 h of chilling are correlated to its respective distribution elicited under 24 h of cold storage condition Starting from consensus value for the 5 quantiles elicited under 24 h cold storage condition, update of estimates of the quantile of the uncertainty distribution was conducted for the parameter physiological status under the 72 h of cold storage condition Experts were asked to provide their individual estimates for 72 h cold storage condition via email
EVIDENCE Graphical representation of individual experts’ judgement on quantiles of the uncertainty distribution for 24 h of cold storage was provided together with representation of consensus value (GROUP). Best‐fitting distributions for 24 h of cold storage based on each expert's input separately and for the consensus values were additionally provided
CONSENSUS QUANTILE	Elicitation for 24 h of cold storage
Lower bound	25
1st quartile	50
Median	70
3rd quartile	80
Upper bound	90
QUESTION: Consider 100 cells, how many will be injured due to cold storage for 72 h at a temperature up to 7°C?
RANGE: 1st round experts provided their elicitation via email. Range, median and quantiles estimates were projected and discussed with experts. A discussion was run during which experts modified their elicitation of range. A final consensus was reached: lower bound = 30, upper bound = 99
MEDIAN AND QUANTILE: Median and quantiles estimates were projected and were discussed with experts. Visualisation of the individual parameter distribution supports the discussion. There was among expert agreements on using the linear pooling distribution. Several distributions were fitted to the linear pooling distribution, in order to obtain a distribution to enter into the model. Since the fitting was poor, experts finally agree on consensus value for quantiles and median values Median = 75; 1st quantile = 60; 3rd quantile = 85
ELICITED DISTRIBUTION Beta(2.104995, 1.3199485) Experts’ and consensus quantiles and fitted distributions:
PARAMETER: Sponging efficacy (SpEff: 0 to 1)
PARAMETER DESCRIPTION: it determines the proportion of surviving and non‐irreversibly injured cells that can be detached by swabbing. It is scored from 0 to 1 and the value reflects the reduced swabbing efficacy compared to time 0. For instance, a value of 0.6 suggests that the swabbing efficacy after 72 h is reduced by 40% compared to time 0
Method: 72 h chilling is obtained extending the cold storage beyond 24 h, therefore uncertainty distributions of the sponging efficacy parameter after 72 h of chilling are related to its respective distribution elicited under 24 h of cold storage condition Starting from values for min, max and most probable value elicited under 24 h cold storage condition, estimates of these 3 points of the uncertainty distribution for sponging efficacy were updated under the 72 h of cold storage condition
EVIDENCE Pert distribution parameters elicited under 24 h of cold storage condition.
QUANTILE	Elicitation for 24 h of cold storage
Min (1% percentile)	60
Max (99% percentile)	100
Most probable value	90
QUESTION: Consider 100 surviving and non‐irreversibly injured cells, how many will not be detached by swabbing due to cold storage for 72 h at a temperature up to 7°C?
RESULTS Individuals's judgements and consensus values:
Consensus values and distribution:
QUANTILE	Elicitation for 72 h of cold storage
Min (1% percentile)	50
Max (99% percentile)	95
Most probable value	80
PARAMETER: Fitness of Salmonella in the enrichment (Compt: 0 to 1)
PARAMETER DESCRIPTION: it refers to the probability of the surviving population, that is not irreversibly injured and is detached by swabbing, to be outcompeted by indigenous meat microbiota. Such population is not capable of growing in the enrichment medium at levels sufficient to be isolated on a selective plate
Methods: 72 h chilling is obtained extending the cold storage beyond 24 h, therefore uncertainty distributions of the outcompeted proportion parameter after 72 h of chilling are correlated to its respective distribution elicited under 24 h of cold storage condition Starting from values for min, max and most probable value elicited under 24 h cold storage condition, estimates of these 3 points of the uncertainty distribution for the probability to be outcompeted were updated under the 72 h of cold storage condition
EVIDENCE Pert distribution parameters elicited under 24 h of cold storage condition
QUANTILE	Elicitation for 24 h of cold storage
Min (1% percentile)	1
Max (99% percentile)	10
Most probable value	3
QUESTION: Consider 100 cells, that are not irreversibly injured and are detached by swabbing, how many will be outcompeted by indigenous meat microbiota due to cold storage for 72 h at a temperature up to 7°C?
RESULTS Individuals’ judgements
Consensus values and distribution:
QUANTILE	Elicitation for 72 h of cold storage
Min (1% percentile)	1
Max (99% percentile)	15
Most probable value	6