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 433 records · Page 24Linked to original sources

PubMed2002

Refining predictive models in critically ill patients with acute renal failure.

Mortality rates in acute renal failure remain extremely high, and risk-adjustment tools are needed for quality improvement initiatives and design (stratification) and analysis of clinical trials. A total of 605 patients with acute renal failure in the intensive care unit during 1989-1995 were evaluated, and demographic, historical, laboratory, and physiologic variables were linked with in-hospital death rates using multivariable logistic regression. Three hundred and fourteen (51.9%) patients died in-hospital. The following variables were significantly associated with in-hospital death: age (odds ratio [OR], 1.02 per yr), male gender (OR, 2.36), respiratory (OR, 2.62), liver (OR, 3.06), and hematologic failure (OR, 3.40), creatinine (OR, 0.71 per mg/dl), blood urea nitrogen (OR, 1.02 per mg/dl), log urine output (OR, 0.64 per log ml/d), and heart rate (OR, 1.01 per beat/min). The area under the receiver operating characteristic curve was 0.83, indicating good model discrimination. The model was superior in all performance metrics to six generic and four acute renal failure-specific predictive models. A disease-specific severity of illness equation was developed using routinely available and specific clinical variables. Cross-validation of the model and additional bedside experience will be needed before it can be effectively applied across centers, particularly in the context of clinical trials.

Acute Kidney Injury↗

PubMed1999

[A predictive model for affect of atopic dermatitis in infancy by neural network and multiple logistic regression].

OBJECTS: To analyze the predictive accuracy of the predictive model for affect of atopic dermatitis in infancy, from the data of the epidemiological survey, which were conducted for 10,000 of mothers of infants and children in 1993. SUBJECTS AND METHODS: A total of 4610 replies were received: 2714 from mothers of infants (12 month old) and 1,896 from mothers of children (2 years old). The sensitivity, specificity and predictive accuracy were calculated from probabilistic model by neural network analysis (NNA) and multiple logistic regression analysis (MLA). RESULTS: Risk factors for probabilistic model by NNA were family history (father, mother, siblings, grand father, grand mother), food restriction, food allergy, age, food restriction of mother, egg introduced time, cow's milk introduced time. The sensitivity, specificity and predictive accuracy of NNA model was 88.6%, 99.5% and 96.4%, respectively and MLA model was 75.1%, 82.6% and 82.3%, respectively. CONCLUSION: These results suggest that the NNA is a good and useful method for prediction of onset of AD than MLA. Furthermore, It is necessary to investigate the artificial neural networks for diagnosis and/or treatment by physician.

Child, Preschool↗

PubMedDate not supplied

Ligand-based molecular modeling study on a chemically diverse series of cholecystokinin-B/gastrin receptor antagonists: generation of predictive model.

Pharmacophore hypotheses were developed for six structurally diverse series of cholecystokinin-B/gastrin receptor (CCK-BR) antagonists. A training set consisting of 33 compounds was carefully selected. The activity spread of the training set molecules was from 0.1 to 2100 nM. The most predictive pharmacophore model (hypothesis 1), consisting of four features, namely, two hydrogen bond donors, one hydrophobic aliphatic, and one hydrophobic aromatic feature, had a correlation (r) of 0.884 and a root-mean-square deviation of 1.1526, and the cost difference between null cost and fixed cost was 81.5 bits. The model was validated on a test set consisting of six different series of 27 structurally diverse compounds and performed well in classifying active and inactive molecules correctly. This validation approach provides confidence in the utility of the predictive pharmacophore model developed in this work as a 3D query tool in the virtual screening of drug-like molecules to retrieve new chemical entities as potent CCK-BR antagonists. The model can also be used to predict the biological activities of compounds prior to their costly and time-consuming synthesis.

Algorithms↗

PubMed2006

External validation of a percutaneous coronary intervention mortality prediction model in patients with acute coronary syndromes.

BACKGROUND: The recently published Michigan outcome prediction model (MM) for inhospital mortality was developed and validated on a series of consecutive patients undergoing percutaneous coronary intervention (PCI). Our purpose was to externally validate the performance of the MM in 2 separate cohorts of patients with acute coronary syndrome (ACS) undergoing PCI in Canada. METHODS: A validation of the MM and development of an extended MM were performed on data describing 10,050 patients from the APPROACH prospective cohort study between January 1995 and December 2000. Performance of both models was assessed on an external data set of 3259 PCI cases from the British Columbia Cardiac Registries. Only patients with a diagnosis of ACS were included in the study. RESULTS: The original MM predicted death rates ranging from 0.1% to 60.6%, but lacked accuracy to predict inhospital mortality as severity increased. The extended MM predicted death rates more widely from 0.0% to a high of 91.0% with better accuracy to predict inhospital death in patients with ACS undergoing PCI. The areas under the receiver operating characteristic curve for the MM and the extended MM on the external validation data set were 0.93 and 0.95, respectively. CONCLUSION: The MM predicts death after PCI in patients with ACS and identifies a clear gradient of risk. However, the enhanced MM developed specifically for the subset of patients with ACS demonstrated better prediction and cross-validated performance. These prediction rules can be useful for risk-adjustment analyses and for prognostication for individual patients.

Aged↗

PubMed1991

Mathematical analysis of perifusion data: models predicting elution concentration.

System models are constructed and analyzed for combined convective flow and for dispersion in distorting concentrations of a chemical signal as it passes through a packed column. We derive general analytical solutions for these models. The results have applications to analyses such as in biological experiments involving hormonal stimulation of perifused cells, elution chromatography, adsorption columns, and studies of groundwater flow. The models reveal that the column distorts an incoming signal (such as a change in solute concentration in the flowing liquid) at the inlet. This distortion is greatest at low values of the Peclet number of the flow and is small at larger values. We explore the effects of the approximations inherent in the mathematical models of the system. Specification of the boundary conditions of the problem are shown to be particularly important. With the use of incorrect models, it is possible to obtain accurate interpolations to data obtained from perfusion experiments. However, the parameters derived (in particular the dispersion constant and the peak concentration of a solute concentration pulse) may be considerably in error. This may lead to errors when these parameter estimates are used to predict results in other experimental situations.

Animals↗

PubMed2005

On-axis and off-axis viewing of images on CRT displays and LCDs: observer performance and vision model predictions.

RATIONALE AND OBJECTIVES: Although cathode-ray tube (CRT) displays typically are used for softcopy display of radiographs in the digital reading environment, liquid crystal displays (LCDs) currently are being used as an alternative. LCDs have many desirable viewing properties compared with a CRT, but significant image degradation can occur with off-axis viewing. This study compares observer performance and predictions from a human visual system model for on-axis and off-axis viewing for CRT displays versus LCDs. MATERIALS AND METHODS: A set of mammograms with different lesion contrasts (n = 400) were shown to six radiologists on CRT and LCD monitors, once on-axis and once off-axis. Observer performance was measured by using receiver operating characteristic techniques. Performance was correlated with results of a human vision model designed to predict observer performance (just noticeable difference [JND]metrix model). Two approaches were used to generate model metrics; a paired discrimination and channelized model observer approach. RESULTS: The performance of human observers indicated that on-axis viewing with the LCD was better than with the CRT, but off-axis viewing was significantly better with the CRT than LCD (F = 8.8175; P < .0001). The paired discrimination model correctly predicted on-axis, but not off-axis, results. The channelized model observer correctly predicted both on- and off-axis results. CONCLUSION: Off-axis viewing of radiographic images on an LCD monitor degrades human observer performance significantly compared with a CRT display. Care should be taken in the clinic to avoid off-axis viewing during diagnostic interpretation.

Breast Neoplasms↗

PubMed2004

A predictive model of human myelotoxicity using five camptothecin derivatives and the in vitro colony-forming unit granulocyte/macrophage assay.

PURPOSE: Many promising anticancer drugs are limited by myelosuppression. It is difficult to evaluate human myelotoxicity before a Phase I study because of the susceptibility of humans and animals to hematotoxicity. The purpose of this study was to establish a reliable method to predict the human maximum tolerated dose (MTD) of five camptothecin derivatives: SN-38, DX-8951f, topotecan, 9-aminocamptothecin, and camptothecin. EXPERIMENTAL DESIGN: The myelotoxicity of SN-38 and DX-8951f were evaluated on bone marrow from mice, dogs, and humans using a 14-day colony-forming unit, granulocyte-macrophage (CFU-GM) assay to determine the 50%, 75%, and 90% inhibitory concentration values (IC50, IC75, and IC90, respectively). RESULTS: Species differences in myelotoxicity were observed for SN-38 and DX-8951f. Using human and murine IC90s for myelotoxicity of these compounds and other camptothecin compounds (topotecan, 9-aminocamptothecin, and camptothecin), in vivo toxicological data, and pharmacokinetic parameters (data referred to in the literature), human MTDs were predicted retrospectively. The mechanism-based prediction model that is proposed uses the in vitro camptothecin assay and in vivo parameters on the basis of free fraction of area under the concentration-curve at the MTD (r2 = 0.887) and suggests that the human MTDs were well predicted for the five camptothecin derivatives by this model rather than by other models. CONCLUSION: The human MTDs of the camptothecin drugs were successfully predicted using the mechanism-based prediction model. The application of this model for in vitro hematotoxicology could play an important role for the development of new anticancer agents.

Animals↗

PubMed2004

Development of predictive modelling approaches for surface temperature and associated microbiological inactivation during hot dry air decontamination.

This research deals with the development of predictive modelling approaches in the field of heat transfer and microbial inactivation. Upon making some backstage microbiological considerations, surface temperature predictions during hot dry air decontaminations are incorporated in a microbial inactivation model, in order to describe inactivation kinetics under realistic (time-varying) temperature conditions. In the present study, the following parts are presented. (i) First, a one-dimensional heat transfer model is developed taking into account exchanges by convection, radiation and evaporation. The model is subsequently validated on a laboratory setup and on a test rig, assuming no water activity changes. This test rig is developed for studying-at a later stage-surface pasteurisation treatment on food products with the use of hot dry air. (ii) Isothermal inactivation data of Escherichia coli K12 MG1655 have been collected and inactivation parameters are accurately estimated by using a primary and a secondary model in a global modelling approach. (iii) Microbiological considerations such as microbial growth effects during come-up times, initial temperature of inactivation, and heat resistance effects, based on experimental observations and on literature studies, are formulated in order to evaluate possible microbial effects arising under the dynamic temperature conditions modelled in step (i). (iv) Microbial inactivation simulations with the incorporation of surface temperature predictions are presented. (v) Finally, the level of the microbial decontamination in an example based on the design of an industrial installation is presented, outlining the importance of the combination of surface temperature and microbial inactivation modelling approaches.

Escherichia coli↗

PubMed1999

[Treatment dropout as failed utilization--development of a predictive model for inpatient psychosomatic rehabilitation].

Premature discontinuance of rehabilitative measures can be considered a revised decision by the patient concerning participation or a revised decision by the clinic concerning admittance. Such termination of therapy suggests that the individual need for rehabilitation (defined as the fit between the patient's need for rehabilitation and the treatment offered by the rehabilitation facility) is not (or no longer) present, at least at this point in time. Assuming that the individual need for rehabilitation actually existed when treatment was requested or approval for rehabilitative measures was granted, the question arises as to when and how the need changed in such a way that would prevent the treatment from continuing as planned and being concluded in a regular fashion. This question will be taken into focus in the present article by developing a model for the prediction and explanation of prematurely discontinued treatment. This model will give central significance to the intention to co-operate, which is considered a dependent variable by reference to individual symptoms and treatment related expectations. Furthermore, various factors of influence during the stay in the clinic are formulated, which, firstly, have a presumed effect on the intention to cooperate and, secondly, have an influence on whether the intention to prematurely terminate treatment develops from a specific intention to cooperate and whether this is then realized. The model is discussed with regard to its practicability and possibilities for operationalization.

Germany↗

PubMed2005

Predictive model for the survival, death, and growth of Salmonella typhimurium in broiler hatchery.

Contamination and penetration of salmonellae into hatching eggs may comprise an important link in the transmission of these bacteria to growing birds, processed carcasses, and eventually to the consumer. In this study, a predictive model for Salmonella typhimurium as a function of initial cell number and storage or incubation time at a nearly constant temperature and humidity was developed and evaluated to compute the bacterial load after 1 d (holding), 10 d (candling), 17 d (incubation), and 21 d (chick processing). Experiments were conducted for S. typhimurium with both high initial bacterial load (HIBL) and low initial bacterial load (LIBL) of 6.0 and 3.5 log cfu/egg, respectively. Eggs with HIBL experienced 2.0 log reduction in the bacterial load after holding at 4 degrees C for 24 h and 3.0 log increase in the bacterial load during incubation and hatch at approximately 37 degrees C between 17 d and 21 d. Experimental data showed that bacterial load of S. typhimurium from holding to chick processing changed from 3.7 to 6.6 log cfu/egg and from 3.7 to 2.7 log cfu/egg in HIBL and LIBL eggs, respectively. The developed model was able to predict bacterial load of S. typhimurium from 3.6 to 6.6 log cfu/egg in HIBL eggs and from 3.4 to 2.7 log cfu/egg in LIBL eggs from holding to chick processing. Root mean square errors and plot of predicted compared with observed bacterial load of S. typhimurium in contaminated eggs yielded a good fit and prediction. The predicted and experimental results indicated that incubated broiler eggs have an increase in internal bacterial loads between incubation and hatch. This model can be used as a tool to predict bacterial load of S. typhimurium in contaminated eggs as well as help predict the behavior of S. typhimurium during hatch.

Animal Husbandry↗

PubMed2005

A new model predicting locomotor cost from limb length via force production.

Notably absent from the existing literature is an explicit biomechanical model linking limb design to the energy cost of locomotion, COL. Here, I present a simple model that predicts the rate of force production necessary to support the body and swing the limb during walking and running as a function of speed, limb length, limb proportion, excursion angle and stride frequency. The estimated rate of force production is then used to predict COL via this model following previous studies that have linked COL to force production. To test this model, oxygen consumption and kinematics were measured in nine human subjects while walking and running on a treadmill at range of speeds. Following the model, limb length, speed, excursion angle and stride frequency were used to predict the rate of force production both to support the body's center of mass and to swing the limb. Model-predicted COL was significantly correlated with observed COL, performing as well or better than contact time and Froude number as a predictor of COL for running and walking, respectively. Furthermore, the model presented here predicts relationships between COL, kinematic variables and body size that are supported by published reduced-gravity experiments and scaling studies. Results suggest the model is useful for predicting COL from anatomical and kinematic variables, and may be useful in intra- and inter-specific studies of locomotor anatomy and performance.

Biomechanical Phenomena↗

PubMed2004

A predictive model to estimate total skin tetrodotoxin in the newt Taricha granulosa.

We developed a predictive model to estimate the total amount of tetrodotoxin (TTX) in the skin of individual newts (Taricha granulosa) based on measures of the amount of TTX present in dorsal skin. We found that regions of skin on a newt could be reliably differentiated using granular gland density and that patterns of variation in granular gland density matched intra-individual variation in TTX levels. Tetrodotoxin is uniformly distributed in dorsal skin and TTX levels in dorsal skin are strongly predictive of TTX levels in other regions of skin on a newt. Our model is both accurate and precise and includes the effect of body size through surface area, variation in granular gland density, as well as positional variation in toxicity apparently associated with variation in granular gland density. This technique allows us to detect patterns that would be unclear if only dorsal skin toxicity or total skin toxicity based on extracts from an entire animal were used and demonstrates that the levels of TTX present in some populations of T. granulosa are remarkably high.

Animals↗

PubMed1999

[Risk factors, diseases and health care acceptance in perinatology: a predictive model of hospital use].

BACKGROUND: The goal of our study was to develop a predictive model of resource use for pregnancy and perinatal care based on the knowledge of the distribution of risk factors in a given population of pregnant women. METHODS: Data recorded in Outcome of Pregnancy Certificates (CIG) from 11 voluntary maternities of the district of Seine-Saint-Denis allowed us to identify those pathologies that were predictive of premature births and prenatal hospitalization of mothers. We built a classification of disease states and of risk level. A logistic regression using disease states as dependent variables and risk levels as independent variables allowed us to compute expected rates with their confidence intervals. RESULTS: Among singletons, malformations, diabetes, toxemia, intra-uterin growth retardation, premature rupture of membranes covered 25% of all pregnancies but explained 64% of maternal hospitalizations; 90% of all mothers hospitalized and with delivery before 37 weeks gestation had at least one of these disease states. But 85% of the women who did not belong to disease classes had a normal pregnancy and delivery. CONCLUSIONS: In a given population, the distribution of risk levels is predictive of the incidence of disease per class. Then, given the length of stay of mothers per class, the rate of transfer of babies and the length of stay in postnatal care, we can simulate bed occupancy and compute bed capacities. The precision of the model is globally good, despite the relatively modest size of our initial data base: it will improve with the use of the model and the expected more widespread availability of data in France.

Adult↗

PubMed2003

Reduction and lumping of physiologically based pharmacokinetic models: prediction of the disposition of fentanyl and pethidine in humans by successively simplified models.

Physiologically based pharmacokinetic (PBPK) models can be used to predict drug disposition in humans from animal data and the influence of disease or other changes in physiology on the pharmacokinetics of a drug. The potential usefulness of a PBPK model must however be balanced against the considerable effort needed for its development. Proposed methods to simplify PBPK modeling include predicting the necessary tissue:blood partition coefficients (kp) from physicochemical data on the drug instead of determining them in vivo, formal lumping of model compartments, and replacing the various kp values of the organs and tissues by only two values, for "fat" and "lean" tissues, respectively. The aim of this study was to investigate the effects of simplifying complex PBPK models on their ability to predict drug disposition in humans. Arterial plasma concentration curves of fentanyl and pethidine were simulated by means of a number of successively reduced models. Median absolute prediction errors were used to evaluate the performance of each model, in relation to arterial plasma concentration data from clinical studies, and the Wilcoxon matched pairs test was used for comparison of predictions. An originally diffusion-limited model for fentanyl was simplified to perfusion-limitation, and this model was either lumped, reducing 11 organ/tissue compartments to six, or changed to a model based on only two kp values, those of fat (used for fat and lungs) and muscle (used for all other tissues). None of these simplifications appreciably changed the predictions of arterial drug concentrations in the 10 patients. Perfusion-limited models for pethidine were set up using either experimentally determined [Gabrielsson et al. 1986] or theoretically calculated [Davis and Mapleson 1993] kp values, and predictions using the former were found to be significantly better. Lumping of the models did not appreciably change the predictions; however, going from a full set of kp values to only two ("fat" and "lean") had an adverse effect. Using a kp for lungs determined either in rats or indirectly in humans [Persson et al. 1988], i.e., a total of three kp values, improved these predictions. In conclusion, this study strongly suggested that complex PBPK models for lipophilic basic drugs may be considerably reduced with marginal loss of power to predict standard plasma pharmacokinetics in humans. Determination of only two or three kp values instead of a "full" set can mean an important reduction of experimental work to define a basic model. Organs of particular pharmacological or toxicological interest should of course be investigated separately as needed. This study also suggests and applies a simple method for statistical evaluation of the predictions of PBPK models.

Adult↗

PubMed2004

A predictive model for the reactor inorganic suspended solids concentration in activated sludge systems.

A simple predictive model for the activated sludge reactor inorganic suspended solids (ISS) concentration (excluding that from chemical precipitant dosing) is presented. It is based on the accumulation of influent ISS in the reactor and an ordinary heterotrophic organism (OHO) ISS content (fiOHO) of 0.15 mg ISS/mg OHO organic (volatile) suspended solids (VSS) and a variable phosphate accumulating organism (PAO) ISS content (fiPAO) proportional to their P content (fXBGP). Organism ISS content is conceptualized as the uptake of dissolved inorganic solids by active organisms, which when dried in the total suspended solids (TSS) test procedure, precipitate and manifest as ISS. The model is validated with data from 22 investigations conducted over the past 15 years on 30 aerobic and anoxic-aerobic nitrification-denitrification (ND) systems and 18 anaerobic-anoxic-aerobic ND biological excess P removal (BEPR) systems variously fed artificial and real wastewater, and operated from 3 to 20 days sludge age. The predicted reactor VSS/TSS ratio reflects the observed relative sensitivity to sludge age, which is low, and to BEPR, which is high. To use the model for design, two parameters need to be known: (1) the influent ISS concentration, which is not commonly measured in wastewater characterization analyses and (2) the P content of PAOs (fXBGP), which can vary considerably depending on the extent of anoxic P uptake BEPR that takes place in the system. Some guidance on the measurement of influent ISS concentration and selection of the PAO P content to calculate the mixed liquor VSS/TSS ratio for design is given.

Bacteria, Aerobic↗

PubMed2002

Growth prediction models, concept and use.

There is a principal and qualitative difference between using regression results on data from groups of children, and using validated prediction models for individual children. Using accurate models, it is now possible to predict the growth response to growth hormone (GH) treatment in a slowly growing child with GH deficiency (GHD) or in a child with idiopathic short stature (ISS). The growth response to the standard dose of GH can be regarded as a bioassay for GH (i.e. the tissue GH responsiveness) and the information on this growth response can be used for different purposes: to decide about treatment or not, for monitoring, and for adjusting the GH dose in order to reach a defined goal for height. This last concept is now used in an ongoing prospective randomized GH dose-finding trial.

Adolescent↗

PubMed2006

Comparison of mortality prediction models after open abdominal aortic aneurysm repair.

OBJECTIVES: Comparison of the accuracy of prediction of contemporary mortality prediction models after open Abdominal Aortic Aneurysm (AAA) surgery. METHODS: Post-operative data were collected from AAA patients from 2 UK Intensive Care Units (ICU). POSSUM and VBHOM based models were compared to the APACHE-AAA model which was able to adjust for the hospital-related effect on outcome. Model performance was assessed using measures of calibration, discrimination and subgroup analysis. RESULTS: 541 patients were studied. The in-hospital mortality rate for elective AAA repair (325 patients) was: 6.2% (95% confidence interval (c.i.) 3.5 to 8.8) and for emergency repair (216 patients) was: 28.7% (95% c.i. 22.5-34.9). The APACHE-based model had the best overall fit to the whole population of AAA patients, and also separately in elective and emergency patients. The V-POSSUM physiology-only (p<0.001) and VBHOM (p=0.011) models had a poor fit in elective patients. The RAAA-POSSUM physiology-only (p<0.001) and VBHOM models (p=0.010) had a poor fit in emergency patients. CONCLUSIONS: The APACHE-AAA model with its ability to adjust for both the hospital-related "effect" as well as the patient case-mix, was a more accurate risk stratification model than other contemporary models, in the post-operative AAA patient managed in ICU.

APACHE↗

PubMed2004

Biochemical recurrence and survival prediction models for the management of clinically localized prostate cancer.

A number of new predictive modeling techniques have emerged in the past several years. These methods, which have been developed in fields such as artificial intelligence research, engineering, and meteorology, are now being applied to problems in medicine with promising results. This review outlines our recent work with use of selected advanced techniques such as artificial neural networks, genetic algorithms, and propensity scoring to develop useful models for estimating the risk of biochemical recurrence and long-term survival in men with clinically localized prostate cancer. In addition, we include a description of our efforts to develop a comprehensive prostate cancer database that, along with these novel modeling techniques, provides a powerful research tool that allows for the stratification of risk for treatment failure and survival by such factors as age, race, and comorbidities. Clinical and pathologic data from 1400 patients were used to develop the biochemical recurrence model. The area under the receiver operating characteristic curve for this model was 0.83, with a sensitivity of 85% and specificity of 74%. For the survival model, data from 6149 men were used. Our analysis indicated that age, income, and comorbidities had a statistically significant impact on survival. The effect of race did not reach statistical significance in this regard. The C index value for the model was 0.69 for overall survival. We conclude that these methods, along with a comprehensive database, allow for the development of models that provide estimates of treatment failure risk and survival probability that are more meaningful and clinically useful than those previously developed.

Aged↗