SEARCH · PubMed Health

Results for “Predictive modeling”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 685 records · Page 38Linked to original sources

PubMed2006

Accurate model predicting sustained response at week 4 of therapy with pegylated interferon with ribavirin in patients with chronic hepatitis C.

Current models used to predict response to peginterferon plus ribavirin treatment, based on viral decline during the first 12 weeks of therapy, have focused on creating an early stopping rule to avoid unnecessary prolongation of therapy. We developed a multivariate model that predicted sustained virological response and nonresponse at baseline and during the first 12 weeks of therapy using collected data from 186 unselected patients with chronic hepatitis C treated with peginterferon plus ribavirin. This model employed ordinal regression with similarity least squares technology to assign the probability of a given outcome. Model variables include sex, age, prior treatment status, genotype, baseline serum alanine aminotransferase levels, histologic necroinflammation and fibrosis scores and serum hepatitis C virus RNA concentration at baseline and weeks, 4, 8, and 12. A multivariate model demonstrated high performance values at all time points. At baseline, the model demonstrated a negative predictive value (NPV) and a positive predictive value (PPV) of 91% and 95%, respectively. At week 4, these values improved to 97% and 100%, respectively, with 95% sensitivity, 89% specificity and 93% accuracy. At week 4, the model was equally efficient for naïve or previously treated patients. Internal validation demonstrated 90% PPV, 94% NPV, 95% sensitivity, 88% specificity and 92% accuracy. A week 4 stopping rule for patients with chronic hepatitis C treated with peginterferon with ribavirin might be proposed by using the model developed in our study.

Adult↗

PubMed2004

Predictive models of copper runoff from external structures.

A general model for annual runoff rate predictions of total copper from naturally patinated copper on buildings at specific urban or rural sites of low chloride influence has been deduced from laboratory and field data. All parameters within the model have a physical meaning and include the average annual rain acidity (pH), the annual rain quantity and the geometry of a building in terms of surface inclination. In 70% of all reported annual runoff rates, the predicted values are within 30% from the observed values. The individual and interactive effect of rain composition in terms of pH, sulfate, chloride and nitrate concentration was investigated in immersion experiments in artificial rain water representative of urban and rural sites of Europe. The results show pH to have a dominating effect on patina dissolution, nitrate to have a small inhibiting effect, whereas no significant effect was seen for chloride and sulfate. In case pH data are not available, a model has been statistically deduced from field data by considering SO2 as influencing parameter, rather than pH. The predictability with the SO2 model is not as good as with the pH model i.e. the pH model should preferentially be used since it is a better predictor and all parameters within the model can be physically explained.

Chlorides↗

PubMed2001

A mechanistic sub-model predicting the influence of potassium on radiocesium uptake in aquatic biota.

It is often argued that the quality of science is related to the possibilities of making accurate predictions. It has also long been argued that due to the complex nature of ecosystems, it will never be possible to predict important target variables, especially with more comprehensive dynamic models. New results in radioecology have, however, demonstrated that those arguments are no longer valid. The key to the predictive success lies in the structuring of the model. The accident at Chernobyl has, in fact, provided science with an intriguing opportunity to study how the pulse of 137Cs is transported through ecosystem pathways, thus revealing the basic structure of these ecosystems, i.e. which are the key and the less-important pathways. It is paradoxical to conclude that the Chernobyl accident is, perhaps, the most important factor behind the revolution in predictive ecosystem modelling which lies behind the decrease in the uncertainty factor from 10 to 0.25-0.5. The sub-model for the potassium moderator presented in this paper is an example of a mechanistically based sub-model used within the framework of a more comprehensive lake model for 137Cs. The K-moderator presented is derived from the existing knowledge on ion transport in biological membranes and takes into account ion equilibria modelled by the Nernst equation and the uptake kinetics quantified by the Michaelis-Menten model. It provides the type of structure to this overall lake model that helps to explain the excellent predictive power of this model.

Animals↗

PubMed1995

Validation and comparison of models predicting survival following intracerebral hemorrhage.

OBJECTIVE: To compare the performance of two previously reported logistic regression models using data independent from those data used to derive the models. DESIGN: Prospective. SETTING: Acute stroke unit of a tertiary care hospital. PATIENTS: One hundred twenty-nine patients with supratentorial intracerebral hemorrhage. MEASUREMENTS AND MAIN RESULTS: Model 1 contains the initial Glasgow Coma Scale score, hemorrhage size, and pulse pressure. The more complex model 2 includes, in addition to those three variables, the presence or absence of intraventricular hemorrhage and a term representing the interaction of intraventricular hemorrhage and Glasgow Coma Scale score. The areas under the receiver operating characteristic curves generated for each model were statistically indistinguishable. CONCLUSIONS: Model 1 predicts 30-day patient status as well as the more complex model 2. Model 1 provides a valid, easy-to-use means of categorizing supratentorial intracerebral hemorrhage patients in terms of their probability of survival.

Cerebral Hemorrhage↗

PubMed2006

Seascape genetics: a coupled oceanographic-genetic model predicts population structure of Caribbean corals.

Population genetics is a powerful tool for measuring important larval connections between marine populations [1-4]. Similarly, oceanographic models based on environmental data can simulate particle movements in ocean currents and make quantitative estimates of larval connections between populations possible [5-9]. However, these two powerful approaches have remained disconnected because no general models currently provide a means of directly comparing dispersal predictions with empirical genetic data (except, see [10]). In addition, previous genetic models have considered relatively simple dispersal scenarios that are often unrealistic for marine larvae [11-15], and recent landscape genetic models have yet to be applied in a marine context [16-20]. We have developed a genetic model that uses connectivity estimates from oceanographic models to predict genetic patterns resulting from larval dispersal in a Caribbean coral. We then compare the predictions to empirical data for threatened staghorn corals. Our coupled oceanographic-genetic model predicts many of the patterns observed in this and other empirical datasets; such patterns include the isolation of the Bahamas and an east-west divergence near Puerto Rico [3, 21-23]. This new approach provides both a valuable tool for predicting genetic structure in marine populations and a means of explicitly testing these predictions with empirical data.

Animals↗

PubMed2000

Validated multivariate models predicting the growth response to GH treatment in individual short children with a broad range in GH secretion capacities.

The aim of the study was to develop and validate models that could predict the growth responses to GH therapy of individual children. Models for prediction of the initial one and 2-y growth response were constructed from a cohort of 269 prepubertal children (Model group) with isolated GH deficiency or idiopathic short stature, using a nonlinear multivariate data fitting technique. Five sets of clinical information were used. The "Basic model" was created using auxological data from the year before the start of GH treatment and parental heights. In addition to Basic model data, the other four models included growth data from the first 2 y of life, or IGF-I, or GH secretion estimated during a provocation test (AITT) or a spontaneous GH secretion profile. The performance of the models was validated by calculating the differences between predicted and observed growth responses in 149 new GH treated children (Validation group) who fulfilled the inclusion criteria used in the original cohort. The SD of these differences (SD(res)) in the validation group was compared with the SD(res) for the model group. For the 1st y, the SD(res) for the Basic model was 0.28 SDscores. The lowest SD(res) (0.19 SDscores), giving the most narrow prediction interval, was achieved adding the 24h GH profile and data on growth from the first 2 y of life to the Basic model. The models presented permit estimation of GH responsiveness in children over a broad range in GH secretion, and with an accuracy of the models substantially better than when using maximal GH response during an provocation test. The predicted individual growth response, calculated using a computer program, can serve as a guide for evidence-based decisions when selecting children to GH treatment.

Adolescent↗

PubMed2001

Phenanthrene desorption from soil in the presence of bacterial extracellular polymer: observations and model predictions of dynamic behavior.

The extracellular polymer produced by a bacterium isolated from soil was employed in laboratory studies of desorption of a model polynuelear aromatic hydrocarbon (PAH), phenanthrene. The experimental results show that the selected extracellular polymer enhances the extent of release of soil-bound phenanthrene. A kinetic model was developed as an aid in interpreting the alterations in phenanthrene desorption resulting from polymer addition. The model employs a statistical gamma (gamma) distribution to describe spectrum of rate constants for transfer of phenanthrene from soil to water, and assumes instantaneous binding of phenanthrene to polymer and of polymer to the test soil. The relevant distribution coefficients and statistical parameters of the gamma distribution needed for the model were evaluated in independent experiments. Using these measured parameters, the model provides a satisfactory independent prediction of phenanthrene release from soil to aqueous phase at two test polymer concentrations, 50 mg TOC/L and 100 mg TOC/L. The success of the independent model predictions suggests a mechanism for the influence of extracellular polymer on phenanthrene desorption. The intrinsic, soil-specific, rate constants for solid to solution transfer of phenanthrene do not appear to be changed by bacterial polymer. Instead, polymer binding of phenanthrene in solution results in an increase in driving force for desorption by decreasing the solution concentration of the free, unbound, PAH molecule.

Bacteria↗

PubMed2004

Certification of herds as free of Mycobacterium paratuberculosis infection: actual pooled faecal results versus certification model predictions.

Dutch dairy herds closed for at least 3 years with no history of paratuberculosis were recruited for a study on herd-certification. One hundred dairy herds were tested for Mycobacterium paratuberculosis at 6-month intervals by pooled faecal culture (five individual animal samples per pool) with solid media. Ninety of the herds completed 9 herd tests and 10 herds dropped out of the study for reasons other than a paratuberculosis diagnosis. Of the 90 herds completing the full study, 61% eventually were found to be M. paratuberculosis-infected. The number of infected herds detected decreased with each round of testing. Assuming that all infected herds had been detected by the ninth herd test, the observed percentage of herds that were truly noninfected (P-free) after each round of testing was calculated. The observed P-free was compared to the predicted P-free based on a previously reported herd-certification model. The P-free predicted by the model was significantly different from the observed P-free. When a single assumption in the model was changed and a diagnostic sensitivity of 40-50% was selected, the predicted P-free closely approximated the observed P-free for the 90 Dutch dairy herds studied. The critical assumption that was changed for Version 2.0 of the model was within-herd infection prevalence for infected but test-negative herds after each round of serial testing. Model Version 1.0 had assumed a 50% decrease in within-herd prevalence but Version 2.0 assumed a stable within-herd prevalence. Culture of pooled faecal samples provides a high-sensitive, high-specific, low-cost test for herd-certification programs.

Animals↗

PubMed2005

Optimization model predictions for postural coordination modes.

This paper examines the ability of the dynamic optimization model to predict changes between in-phase and anti-phase postural modes of coordination and to evaluate influence of two particular environmental and intentional constraints on postural strategy. The task studied was based on an experimental paradigm that consisted in tracking a target motion with the head. An original optimal procedure was developed for cyclic problems to calculate hip and ankle angular trajectories during postural sway with a minimum torque change criterion. Optimization results give a good description of the sudden bifurcation phase between in-phase and anti-phase postural coordination modes in visual target tracking. Transition frequency and predicted effects of environmental and intentional constraints are also in line with experimental observations described in existing literature. In particular, these investigations pointed out that postural planning process can be related to the minimization of a dynamic cost criterion with an equilibrium constraint. In conclusion, the optimization technique is well suited for the prediction of postural modes of coordination and seems to offer many opportunities for better comprehension of neuromuscular movement control.

Ankle↗

PubMed2005

Discrimination and calibration of mortality risk prediction models in interventional cardiology.

OBJECTIVES: Using a local percutaneous coronary intervention (PCI) data repository, we sought to compare the performance of a number of local and well-known mortality models with respect to discrimination and calibration. BACKGROUND: Accurate risk prediction is important for a number of reasons including physician decision support, quality of care assessment, and patient education. Current evidence on the value of applying PCI risk models to individual cases drawn from a different population is controversial. METHODS: Data were collected from January 01, 2002 to September 30, 2004 on 5216 consecutive percutaneous coronary interventions at Brigham and Women's Hospital (Boston, MA). Logistic regression was used to create a local risk model for in-hospital mortality in these procedures, and a number of statistical methods were used to compare the discrimination and calibration of this new and old local risk models, as well as the Northern New England Cooperative Group, New York State (1992 and 1997), University of Michigan consortium, American College of Cardiology-National Cardiovascular Data Registry, and The Cleveland Clinic Foundation risk prediction models. Areas under the ROC (AUC) curves were used to evaluate discrimination, and the Hosmer-Lemeshow (HL) goodness-of-fit test and calibration curves assessed applicability of the models to individual cases. RESULTS: Multivariate risk factors included in the newly constructed local model were: age, prior intervention, diabetes, unstable angina, salvage versus elective procedure, cardiogenic shock, acute myocardial infarction (AMI), and left anterior descending artery intervention. The area under the ROC curve (AUC) was 0.929 (SE=0.017), and the p value for the Hosmer-Lemeshow (HL) goodness-of-fit was 0.473. This indicates good discrimination and calibration. Bootstrap re-sampling indicated AUC stability. Evaluation of the external models showed an AUC range from 0.82 to 0.90 indicating good discrimination across all models, but poor calibration (HL p value < or = 0.0001). CONCLUSIONS: Validation of AUC values across all models suggests that certain risk factors have remained important over the last decade. However, the lack of calibration suggests that small changes in patient populations and data collection methods quickly reduce the accuracy of patient level estimations over time. Possible solutions to this problem involve either recalibration of models using local data or development of new local models.

Angioplasty, Balloon, Coronary↗

PubMed2000

Chromosomal regions involved in hybrid performance and heterosis: their AFLP(R)-based identification and practical use in prediction models.

In this paper, a novel approach towards the prediction of hybrid performance and heterosis is presented. Here, we describe an approach based on: (i) the assessment of associations between AFLP(R) markers and hybrid performance and specific combining ability (SCA) across a set of hybrids; and (ii) the assumption that the joint effect of genetic factors (loci) determined this way can be obtained by addition. Estimated gene effects for grain yield varied from additive, partial dominance to overdominance. This procedure was applied to 53 interheterotic hybrids out of a 13 by 13 half-diallel among maize inbreds, evaluated for grain yield. The hybrid value, representing the joint effect of the genetic factors, accounted for up to 62.4% of the variation in the hybrid performance observed, whereas the corresponding efficiency of the SCA model was 36.8%. Efficiency of the prediction for hybrid performance was evaluated by means of a cross-validation procedure for grain yield of (i) the 53 interheterotic hybrids and (ii) 16 hybrids partly related to the 13 by 13 half-diallel. Comparisons in prediction efficiency with the 'distance' model were made. Because the map position of the selected markers is known, putative quantitative trait loci (QTL) affecting grain yield, in terms of hybrid performance or heterosis, are identified. Some QTL of grain yield detected in the present study were located in the vicinity of loci reported earlier as having quantitative effects on grain yield.

Alleles↗

PubMed2004

Predictive models of CYP3A4 Heteroactivation: in vitro-in vivo scaling and pharmacophore modeling.

Although activation of CYP3A4 is frequently observed in vitro, predictive computational-based models and methods for in vitro-in vivo scaling are scarce. It has been previously shown that in vitro CYP3A4 heteroactivation of carbamazepine (CBZ)-epoxide (ep) formation can be associated with the clinical drug interaction between felbatame and CBZ. The previously reported prediction methodology is applied here to an additional set of in vitro CYP3A4 heteroactivators, some exerting this effect at concentrations relevant in vivo. The antimalarial artemisinin potently increases CBZ-ep formation by a maximum of 500% at 300 microM. Testosterone and progesterone activates by a maximum of 1680 and 920%, respectively, at 150 microM, and quinidine causes a 130% increase at 300 microM. The predicted maximum in vivo decrease in steady-state concentration of carbamazepine (Css(CBZ)) at saturating effector concentrations is 85 to 90% for testosterone and progesterone, 75% for artemisinin, and 45% for quinidine. The corresponding predicted in vivo increase in Css(CBZ-ep) is 50, 60, 55, and 30% for artemisinin, testosterone, progesterone, and quinidine, respectively. At effector concentrations relevant in vivo, the Css(CBZ) change is predicted to </=20% for testosterone, artemisinin, and quinidine and </=10% for progesterone, with a concomitant Css(CBZ-ep) increase of 12% for testosterone and </=10% for progesterone, artemisinin, and quinidine. Structure-heteroactivation relationships were evaluated by generating a pharmacophore. The model includes two hydrogen bond acceptor features separated by hydrophobic features. Internal predictivity is high, and heteroactivation of an external test set correlate to observed in vitro heteroactivation.

Artemisinins↗

PubMed2005

Practical considerations on the use of predictive models for regulatory purposes.

Interest in the use of quantitative structure-activity relationships (QSARs) for regulatory purposes has been growing steadily over the years, and many models have been evaluated under the guidance and acceptability criteria defined at the Setubal workshop held in March 2002. This work explores some of the practical issues related to the use of QSARs for regulatory purposes using results obtained from rat oral lethality and fish acute toxicity estimates generated from computational models (including TOPKAT, MCASE, OASIS, and ECOSAR). Using data submitted under the Environmental Protection Agency's (EPA's) High Production Volume (HPV) Challenge Program, the results on the quality of the estimations are compared using a standard statistical review and an additional classification approach in which the hazard predictions were grouped using well-defined regulatory criteria (those used in EPA's New Chemical Program). Our results indicate that an evaluation of a model's regulatory applicability and predictive power is ultimately dependent on the specific criteria used in the assessment process. This work also discusses the practical difficulties associated with defining the domain of a predictive model using the estimates of four different ready biodegradation models and experimental data submitted under the EPA's New Chemical program. Our results suggest that the method a model employs for its predictions is as important as the training set in determining its domain of applicability. Together, these results highlight the challenges associated with developing reliable and easily applied acceptability criteria for the regulatory use of QSAR models.

Biodegradation, Environmental↗

PubMed1991

Pharmacokinetics of acetaminophen, antipyrine, and salicylic acid in the lactating and nursing rabbit, with model predictions of milk to serum concentration ratios and neonatal dose.

The rabbit was utilized for examining the pharmacokinetics of three compounds (acetaminophen, AC; antipyrine, AN; and salicylic acid, SA) in nursing adults and their suckling offspring and for assessing the ability of a diffusional model to predict milk to serum drug concentration ratios (M/S) from in vitro experiments. AC, AN, and SA serum concentration time profiles declined monoexponentially for both adults and their pups. The mean systemic clearance (Cls) for AC in the adults and pups was 16.1 and 13.7 ml/min/kg, respectively. The mean half-lives of AC (t1/2) were 25.5 and 33.3 min in the adult and pup groups, respectively. AN declined in parallel for adult rabbits and an older group of suckling pups (23-25 days old). In a younger group of pups (18-21 days old) it declined with a longer t1/2 (97.5, 95.1, and 347.6 min in the adults, older pups, and younger pups, respectively). The mean AN Cls in the adults, the older pups, and the younger pups was 5.34, 6.30, and 1.91 ml/min/kg, respectively. The time course of SA was prolonged in the suckling pups (t1/2 of 633 min in the pups vs 78.7 min in the adult). The mean Cls values in the adults and the pups were 1.05 and 0.27 ml/min/kg, respectively. The mean systemic clearance of unbound drug (Clu) for SA was 11.2 ml/min/kg in the adults and 0.92 ml/min/kg in the pups. The serum protein binding of AC and AN was limited, whereas the mean free fraction for SA was 9.7% in adult serum and 32.5% in pup serum. AC and AN in milk paralleled serum drug profiles; a time lag was noted for milk SA. M/S ratios were determined in vivo (M/Sobs; AN = 0.885, AC = 0.580, and SA = 0.125) using area under the milk and serum concentration time profiles. Predicted M/S values (M/Spred; AN = 0.779, AC = 0.578, and SA = 0.085) were calculated from in vitro measurements of the unbound fractions of drug in skim milk and serum, the skim to whole milk drug concentration ratio, milk and serum pH, and the pKa of the model compound. Mean values for M/Sobs were highly correlated with M/Spred values (r2 = 0.976) when the present data were combined with previous data for propranolol, phenobarbital, phenytoin, and diazepam (Fleishaker, J.C., and McNamara, P.J., J. Pharmacol. Exp. Ther. 244, 919, 1988). These results support the usefulness of the diffusional model for predicting M/S in vivo, provided that the distributional process is governed by passive diffusion.(ABSTRACT TRUNCATED AT 400 WORDS)

Acetaminophen↗

PubMedDate not supplied

Comparison of diagnostic usefulness of predictive models in preliminary differentiation of adnexal masses.

The purpose of this study was to compare prognostic models evaluating the probability of an ovarian cancer occurrence based on a number of clinical and ultrasonographic data in women with adnexal masses. A total of 686 women with adnexal masses underwent the examinations between 1994 and 2002. The recorded parameters included: age, menopausal status, body mass index, the grayscale and Doppler ultrasonographic examination, and selected markers concentration levels. In order to find the best combination of features, which significantly influences the probability of malignancy, stepwise logistic regression analysis, as well as artificial neural network, was used. The diagnostic efficiency of received models was estimated and compared using receiver-operating characteristics (ROC) curve. The results indicate that 431 and 255 patients had a benign and malignant ovarian tumor, respectively. Application of stepwise logistic regression analysis revealed statistically significant importance of eight features. The sensitivity and specificity for the received model were 65.71% and 77.59%, respectively. Three-layer perceptron network shows 13 features as significant predictors of malignancy. The network gave a sensitivity of 85.7% and specificity of 93.1%. Comparison of area under ROC curve for received models was 0.9679 vs 0.9716. Prognostic values of the analyzed neural model are not optimal but seem to surpass logistic regression model in terms of the predictive possibilities.

Adult↗

PubMed1999

The effects of stochastic neural activity in a model predicting intensity perception with cochlear implants: low-rate stimulation.

Most models of auditory nerve response to electrical stimulation are deterministic, despite significant physiological evidence for stochastic activity. Furthermore, psychophysical models and analyses of physiological data using deterministic descriptions do not accurately predict many psychophysical phenomena. In this paper we investigate whether inclusion of stochastic activity in neural models improves such predictions. To avoid the complication of interpulse interactions and to enable the use of a simpler and faster auditory nerve model we restrict our investigation to single pulses and low-rate (< 200 pulses/s) pulse trains. We apply signal detection theory to produce direct predictions of behavioral threshold, dynamic range and intensity difference limen. Specifically, we investigate threshold versus pulse duration (the strength-duration characteristics), threshold and uncomfortable loudness (and the corresponding dynamic range) versus phase duration, the effects of electrode configuration on dynamic range and on strength-duration, threshold versus number of pulses (the temporal-integration characteristics), intensity difference limen as a function of loudness, and the effects of neural survival on these measures. For all psychophysical measures investigated, the inclusion of stochastic activity in the auditory nerve model was found to produce more accurate predictions.

Cochlear Implants↗

PubMed2002

Artificial neural network predictive model for allergic disease using single nucleotide polymorphisms data.

The purpose of this study was to develop a novel diagnostic prediction method for allergic diseases from the data of single nucleotide polymorphisms (SNPs) using an artificial neural network (ANN). We applied the prediction method to four allergic diseases, such as atopic dermatitis (AD), allergic conjunctivitis (AC), allergic rhinitis (AR) and bronchial asthma (BA), and verified its predictive ability. Almost all the learning data were precisely predicted. Regarding the evaluation data, the learned ANN model could correctly predict a diagnosis with more than 78% accuracy. We also analyzed the SNP data using multiple regression analysis (MRA). Using the MRA model, less than 10% of patients with the above allergic diseases were correctly diagnosed, while this figure was more than 75% for persons without allergic diseases. From these results, it was shown that the ANN model was superior to the MRA model with respect to predictive ability of allergic diseases. Moreover, we used two different methods to convert the genetic polymorphism data into numerical data. Using both methods, diagnostic predictions were quite precise and almost the same predictive abilities were observed. This is the first study showing the application and usefulness of an ANN for the prediction of allergic diseases based on SNP data.

Journal Article↗

PubMed2001

On the need for another type of predictive model in structured foods.

Most of the models discussed up till now in predictive microbiology do not take into account the variability of microbial growth with respect to space. In structured (solid) foods, microbial growth can strongly depend on the position in the food and the assumption of homogeneity can thus not be accepted: space must be considered as an independent variable. Indeed, experimental evidence exists of bacteria competition on agar not showing the same behavior as the competition in a well-mixed liquid culture system. It is conjectured that this is due to the spatially structured habitat. Therefore, in the current paper, a prototype two species competition model proposed in previous work by the authors is extended to take space into account. The extended model describes two phenomena: (i) local evolution of biomass and (ii) transfer of biomass through the medium. The structure of the food product is taken into account by limiting the diffusion through the medium. The smaller mobility of the micro-organisms in solid foods allows spatial segregation which causes pattern formation. Evidence is given for the fact that taking space into account indeed has an influence on the behavior (coexistence/extinction) of the populations. Although the reported simulations are by no means to be interpreted as accurate predictions, the proposed model structure allows one to highlight (i) important characteristics of microbial growth in structured foods and (ii) future research trends in predictive microbiology.

Bacteria↗