ARS | Publication request: SAFE PRODUCTION ZONE, A NEW METHOD FOR VALIDATION OF PREDICTIVE MODELS

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: March 15, 2003
Publication Date: August 10, 2003
Citation: OSCAR, T.P. SAFE PRODUCTION ZONE, A NEW METHOD FOR VALIDATION OF PREDICTIVE MODELS. MEETING ABSTRACT. 2003. pp. 119.

Technical Abstract: Predictions of broth models for lag time (LT) and specific growth rate (SGR) of Salmonella Typhimurium were compared with observed LT and SGR on sterile cooked chicken breast and thigh burgers. The experimental conditions and modeling methods were the same for the broth models and sterile cooked chicken burgers so as not to confound the comparisons. Relative error (RE) plots of broth model predictions were evaluated for prediction bias and accuracy using a safe prediction zone (SPZ) that ranged from RE of -0.3 to 0.15 where positive RE represented fail-dangerous predictions and negative RE represented fail-safe predictions. Relative errors outside the SPZ were considered overly fail-dangerous or overly fail-safe. The fraction of RE that were within the SPZ (RESPZ) was calculated and a minimum RESPZ of 0.85 was established and used as the sole criteria for model validation. For LT, 16 of 24 RE were within the SPZ for a RESPZ of 0.64, whereas for SGR, 21 of 24 RE were within the SPZ for a RESPZ of 0.88. Thus, the broth model for LT was not validated (i.e., RESPZ < 0.85), whereas the broth model for SGR was validated. Of note, the broth model for LT provided overly fail-dangerous predictions of S. Typhimurium growth on sterile cooked chicken burgers at short LT (< 4 h).