Thermal inactivation of pathogens has been studied extensively, which has resulted in a wide range of D- and z-values. Estimating the inactivation rate for a specific condition based on these reported values is difficult, since one has to select representative conditions, and data obtained exactly at the required representative conditions are generally not available. Therefore, a first step could be to globally assess a heat treatment taking into account largest effects only. Once the most important parameters are known, a more precise study of inactivation can be performed. Therefore, in this study a large quantity of D-values (n=4066) was collected from literature for various pathogens and linear regression was applied to obtain average D-values (together with the 95% upper prediction level) and z-values. When comparing these overall data, it can be seen that most factors reported to have an effect on the D-value are smaller than the variability of all published D-values. Even effects of shoulders disappear in the overall analysis. Only a limited number of factors that did have a significant effect (p<0.05) on the D-value were identified: for Salmonella spp., the presence of chocolate ingredients gave protection to the cells, for Listeria monocytogenes the presence of 10% salt (or a(w)<0.92) resulted in a higher heat resistance, for Bacillus cereus there were significant differences for various strains and in oily products and for Clostridium botulinum there were significant differences in heat resistance between different types of C. botulinum. This does not mean that other effects do not occur, but it shows the main effects that have to be included for a first impression on the performance of a heating process. The obtained 95% upper prediction levels of the D-values can be used as a (conservative) estimate of inactivation and can be used to give order of magnitude values in overall process evaluations.