Biofilm Efficacy = CNT − ACT

where CNT and ACT were pathogen cell numbers (log CFU/cm2) in the control (without probiotic biofilm) and in the active sample (with probiotic biofilm). These differences were analysed through one-way ANOVA and Tukey's test as the post-hoc comparison test (*p* < 0.05). Results about biofilm formation on different materials were expressed as log CFU/cm2, presented as the average of replicates (*n* = 4) and analyzed through one-way ANOVA and Tukey's test as the post-hoc comparison test (*p* < 0.05).

The microbiological data collected during the challenge tests were expressed as the average of two replicates and the obtained mean values (one for experiment) were modelled according to the Gompertz equation modified by Zwietering et al. [30]:

$$y = k + A \ast \exp[-\exp[\left(\mu\_{\text{max}} \ast \varepsilon/A\right) \ast \left(\lambda - t\right) + 1]],\tag{1}$$

where *y* is the concentration of the microorganism (Log CFU/ g), *k* is the initial level of the dependent variable to be modelled, *A* is the difference between the decimal logarithm of the initial value of cell concentration and the decimal logarithm of maximum bacteria growth attained at the stationary phase (Log CFU/g), μ*max* is the maximal growth rate (1/day), λ the lag time (day) and *t* the time.

Following Castillejo Rodriguez et al. [31], the sanitary risk time for the growth of *L. monocytogenes* in our samples was determined as the time (in days) that it took to observe an increase of 2 Log CFU/g of the count of this microorganism in food as follow:

$$\text{Sarnitary risk time [SRT]} = \text{2/\mu},\tag{2}$$

where μ is the maximal growth rate.

For the growth of *Ps. fluorescens*, the maximum acceleration of microbial growth (dy2/dt2 (day)), known as stability time, was also estimated with the Gompertz equation, following Riva et al. [32].

To determine whether significant differences (*p* < 0.05) existed among the parameters calculated by using the Gompertz equation, a one-way analysis of variance (ANOVA), followed by Tukey's test, was conducted.

Modeling was performed through the software Statistica for Windows version 10.0 (Statsoft, Tulsa, OK, USA).
