The data collected from studies to monitor inactivation of Escherichia coli
O157:H7 in uncooked fermented salami were used to develop models to descri
be survival of the organism. Three models were developed that included diff
erent variables to best describe E. coli O157:H7 reduction. Model A include
d the variables water activity (a,), pH, time, and quadratic variables pH a
nd time. Model B separated the processing stages into fermentation and dryi
ng. The fermentation included the variables pH and temperature x time (ttar
ea) and interaction between the two variables. The drying stage was modeled
using the variables time and a, and interaction between the two. Model C l
ooked at variables a, and the time at pH 5.3 to achieve a 2-D log reduction
of E. coli O157:H7 and the interaction between the variables. The variable
s selected for inclusion in all the models were significant at the P < 0.00
01 level. The predicted values for all models correlated well to the observ
ed values with R-2 Of 0.888, 0.828, 0.836, and 0.818 for models A, B-ferm,
B-drying and C, respectively. The models were validated using data from a t
rial not used to develop the model. In general the predicted reduction in E
. coli O157:H7 count in uncooked fermented salami was greater than for the
observed E. coli O157:H7 log reductions for all models, but the relation be
tween the two was linear. The results demonstrate that modeling can be a us
eful tool in assessing manufacturing practices for uncooked fermented salam
i processes.