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 217 records · Page 12Linked to original sources

PubMed2001

Predictive modeling techniques in prostate cancer.

A number of new predictive modeling techniques have emerged in the past several years. These methods can be used independently or in combination with traditional modeling techniques to produce useful tools for the management of prostate cancer. Investigators should be aware of these techniques and avail themselves of their potentially useful properties. This review outlines selected predictive methods that can be used to develop models that may be useful to patients and clinicians for prostate cancer management.

Humans↗

PubMed2001

An efficient formulation of Krylov's prediction model for train induced vibrations based on the dynamic reciprocity theorem.

In Krylov's analytical prediction model, the free field vibration response during the passage of a train is written as the superposition of the effect of all sleeper forces, using Lamb's approximate solution for the Green's function of a halfspace. When this formulation is extended with the Green's functions of a layered soil, considerable computational effort is required if these Green's functions are needed in a wide range of source-receiver distances and frequencies. It is demonstrated in this paper how the free field response can alternatively be computed, using the dynamic reciprocity theorem, applied to moving loads. The formulation is based on the response of the soil due to the moving load distribution for a single axle load. The equations are written in the wave-number-frequency domain, accounting for the invariance of the geometry in the direction of the track. The approach allows for a very efficient calculation of the free field vibration response, distinguishing the quasistatic contribution from the effect of the sleeper passage frequency and its higher harmonics. The methodology is validated by means of in situ vibration measurements during the passage of a Thalys high-speed train on the track between Brussels and Paris. It is shown that the model has good predictive capabilities in the near field at low and high frequencies, but underestimates the response in the midfrequency band.

Journal Article↗

PubMed1999

How generalizable are coronary risk prediction models? Comparison of Framingham and two national cohorts.

BACKGROUND: Previous models used to predict individual risk of death from coronary heart disease (CHD) were developed from data of 3 decades ago from the Framingham Heart Study. CHD mortality rates have declined markedly since that period as a result of improvement in both risk factor status and medical interventions. Generalization of the results from this one study to the population at large remains a matter of concern. We compared predictive functions derived from the major risk factors for CHD from Framingham and 2 more recent national cohorts, the First and Second National Health and Nutrition Examination Survey (NHANES I and NHANES II). METHODS AND RESULTS: The participants included 1846 men and 2323 women 35 to 69 years of age and free of CHD at the fourth examination (1954 to 1958) from the Framingham Study; 2753 men and 3858 women from the NHANES I (1971 to 1975); and 2655 men and 3050 women from NHANES II (1976 to 1980). The 3 cohorts were monitored for 24, 20, and 15 years, respectively. Significant heterogeneity existed among studies in the magnitude of the Cox coefficients for the individual factors (ie, age, systolic blood pressure, serum total cholesterol, and smoking status), especially among men. When risk factors were considered collectively, however, functions derived from and applied to different cohorts had a similar ability to rank individual risk. The areas under the receiver operating characteristic curves were 0. 71 to 0.76 in men and 0.76 to 0.81 in women when different risk functions were applied to their own population or to a second population. The cumulative CHD survival observed in women in the 2 national cohorts was close to what was predicted from the Framingham equation. However, Framingham overestimated the cumulative CHD mortality rates in men in NHANES I and NHANES II. CONCLUSIONS: The Framingham risk model for the prediction of CHD mortality rates provides a reasonable rank ordering of risk for individuals in the US white population for the period 1975 to 1990. However, prediction of absolute risk is less accurate.

Adolescent↗

PubMed2001

Outcome prediction model for very elderly critically ill patients.

CONTEXT: Very elderly critically ill patients have three possible hospital outcomes: discharge to home, discharge to a skilled nursing or rehabilitation facility, or death. The factors associated with these outcomes are unknown. OBJECTIVE: To develop a three-outcome prediction model for very elderly critically ill patients. DESIGN: Retrospective chart abstraction with ordered logistic regression analysis. SETTING: Academic medical center. PATIENTS: Four hundred and fifty-five patients 85 yrs or older admitted to intensive care units (ICU) during 1996 and 1997. MEASUREMENTS AND MAIN RESULTS: A fitted ordinal logistic regression predictive model was developed using data from 243 patients hospitalized in 1996, and validated on data from 212 patients hospitalized in 1997. Model variables include age, gender, baseline support level, type of ICU, heart rate at ICU admission, use of mechanical ventilation, vasopressors or a pulmonary artery catheter during the ICU stay, and the development of respiratory, neurologic or hematologic failure or sepsis while in the ICU. When tested on the 1997 data, the model was well calibrated and had a high discriminant index. CONCLUSIONS: This mathematical model can be used to predict the risks of these three hospital outcomes for this population of patients. These predictions can provide a context when discussing goals and expectations with patients, families, and other healthcare providers and to aid in hospital discharge planning.

Academic Medical Centers↗

PubMed1993

A predictive model for the behavior of radionuclides in lake systems.

This paper describes a predictive model for the behavior of 137Cs in lacustrine systems. The model was tested by comparing its predictions to contamination data collected in various lakes in Europe and North America. The migration of 137Cs from catchment basin and from bottom sediments to lake water was discussed in detail; these two factors influence the time behavior of contamination in lake water. The contributions to the levels of radionuclide concentrations in water, due to the above factors, generally increase in the long run. The uncertainty of the model, used as a generic tool for prediction of the levels of contamination in lake water, was evaluated. Data sets of water contamination analyzed in the present work suggest that the model uncertainty, at a 68% confidence level, is a factor 1.9.

Cesium Radioisotopes↗

PubMed2003

Comparing statistical and machine learning classifiers: alternatives for predictive modeling in human factors research.

Multivariate classification models play an increasingly important role in human factors research. In the past, these models have been based primarily on discriminant analysis and logistic regression. Models developed from machine learning research offer the human factors professional a viable alternative to these traditional statistical classification methods. To illustrate this point, two machine learning approaches--genetic programming and decision tree induction--were used to construct classification models designed to predict whether or not a student truck driver would pass his or her commercial driver license (CDL) examination. The models were developed and validated using the curriculum scores and CDL exam performances of 37 student truck drivers who had completed a 320-hr driver training course. Results indicated that the machine learning classification models were superior to discriminant analysis and logistic regression in terms of predictive accuracy. Actual or potential applications of this research include the creation of models that more accurately predict human performance outcomes.

Adolescent↗

PubMed2026

Genomic selection in timothy (Phleum pratense L.): a comprehensive evaluation of prediction models, multi-trait strategies, and forward validation across Norwegian environments.

This study presents a comprehensive evaluation of genomic selection (GS) in timothy (Phleum pratense L.), comparing nine prediction models across yield and quality traits at two Norwegian locations. Forward validation with independent full-sib (FS2) families revealed a substantial generalization gap, highlighting the need for realistic accuracy assessment in polyploid forage breeding. Timothy (Phleum pratense L.) is the most important forage grass in Northern Europe, yet genomic selection has not been systematically evaluated in this hexaploid species. We assessed 889 FS2-families originating from biparental crosses among 49 cultivars/populations. The FS2-families were genotyped with 30,698 SNP markers derived from genotyping-by-sequencing (GBS) and field tested for three harvest years at a highland and a lowland continental location in Southern Norway. Nine genomic prediction models were compared for six yield traits (dry matter yield per cut and total) and six quality traits (protein, digestibility, and fiber fractions) across three cuts/year. Within-training cross-validation accuracies were moderate to high (mean r = 0.62), with Random Forest and SVR consistently outperforming GBLUP. However, forward validation using 213 independent FS2-families revealed dramatically lower accuracies (mean r = 0.16), with only 16 of 30 trait-dataset combinations reaching statistical significance (p < 0.05). Genomic heritabilities (GREML), estimated across environments, ranged from near zero for the quality traits to 0.55 for the yield traits. Multi-trait models improved accuracy by 3-5% over single-trait approaches, while FS2 families-by-environment interaction models with Random Forest achieved the highest within-training accuracy (mean r = 0.71). Marker density analysis showed accuracy plateauing at approximately 15000 SNPs. Genetic correlations among the yield component traits were estimated by multi-trait REML; correlations among the quality traits could not be estimated reliably because their genomic heritabilities were low. A multi-trait selection index identified top-performing FS2-families for further crossing recommendations. These results provide a benchmark for GS implementation in hexaploid timothy and emphasize that cross-validation substantially overestimates prediction accuracy for truly independent material.

Norway↗

PubMed2001

Internal validation of predictive models: efficiency of some procedures for logistic regression analysis.

The performance of a predictive model is overestimated when simply determined on the sample of subjects that was used to construct the model. Several internal validation methods are available that aim to provide a more accurate estimate of model performance in new subjects. We evaluated several variants of split-sample, cross-validation and bootstrapping methods with a logistic regression model that included eight predictors for 30-day mortality after an acute myocardial infarction. Random samples with a size between n = 572 and n = 9165 were drawn from a large data set (GUSTO-I; n = 40,830; 2851 deaths) to reflect modeling in data sets with between 5 and 80 events per variable. Independent performance was determined on the remaining subjects. Performance measures included discriminative ability, calibration and overall accuracy. We found that split-sample analyses gave overly pessimistic estimates of performance, with large variability. Cross-validation on 10% of the sample had low bias and low variability, but was not suitable for all performance measures. Internal validity could best be estimated with bootstrapping, which provided stable estimates with low bias. We conclude that split-sample validation is inefficient, and recommend bootstrapping for estimation of internal validity of a predictive logistic regression model.

Aged↗

PubMed2004

Genetic markers applied in regression tree prediction models.

Classification and regression tree (CART) modelling was used to determine infectious hypodermal and haematopoietic necrosis virus (IHHNV) resistance and susceptibility in Penaeus stylirostris. In a previous study, eight random amplified polymorphic DNA (RAPD) markers and viral load values using real-time quantitative PCR were obtained and used as the training data set in order to create numerous regression tree models. Specifically, the genetic markers were used as categorical predictor variables and viral load values as the dependent response variable. To determine which model has the highest predictive accuracy for future samples, RAPD fingerprint data was generated from new Penaues stylirostris IHHNV resistant and susceptible individuals and used to test the regression models. The best performing tree was a four terminal node tree with three genetic markers as significant variables. Marker-assisted breeding practices may benefit from the creation of regression tree models that apply genetic markers as predictive factors. To our knowledge this is the first study to use RAPD markers as predictors within a CART prediction model to determine viral susceptibility.

Animals↗

PubMed1997

Predictive model to identify trauma patients with blood alcohol concentrations > or = 50 mg/dl.

OBJECTIVE: To develop a simple model for identification of trauma patients who are likely to have a blood alcohol concentration > or = 50 mg/dL (BAC + 50). METHODS: Demographic, clinical, and BAC data were collected from the clinical trauma registry and toxicology data base at a Level I trauma center. Logistic regression was used to analyze data from 11,206 patients to develop a predictive model, which was validated using a subsequent cohort of 3,523 patients. RESULTS: In the model development cohort, alcohol was detected in the blood of 3,180 BAC-tested patients (28.7%), of whom 91.2% had a BAC + 50 status. Preliminary analysis revealed associations between a BAC + 50 status and sex, age, race, injury type (intentional vs. unintentional), and time of injury (night vs. day and weekend vs. weekday). A predictive model using four attributes (sex and injury type) identified patients at low, medium, and high risk for being BAC + 50. The model was validated using the second group of patients. CONCLUSION: Injured patients with a high probability of being alcohol positive can be identified using a simple scoring system based on readily available demographic and clinical information.

Adolescent↗

PubMed2003

Validation and calibration of the Kabi Pharmacia International Growth Study prediction model for children with idiopathic growth hormone deficiency.

In 1999 a model was published for prediction of growth in children with idiopathic GH deficiency (IGHD) during GH therapy, derived using data from the Kabi Pharmacia International Growth Study (KIGS) database (Pharmacia \|[amp ]\| Upjohn, Inc., International Growth Database). We validated and calibrated this KIGS model for growth in the first year of GH therapy using data from 136 Dutch children with IGHD. Observed vs. predicted outcomes were plotted, and the fitted regression line was significantly different from the line of identity (P = 0.03). It appeared that the predictions were too extreme: relatively low predictions were too low, relatively high predictions were too high. This is a well known phenomenon in the context of prediction models, called overoptimism. For valid application to other data the KIGS predictions should be calibrated. Calibrated predictions are obtained using Y(cal) = Y(orig) + (2.153 - 0.192 x Y(orig)), where Y(cal) is the calibrated prediction, and Y(orig) is the KIGS prediction. The calibrated prediction will be higher than the original KIGS prediction when the original prediction is less than 11.2 cm/yr and will be lower otherwise. The variability of the prediction errors of the calibrated predictions was positively related to the value of the prediction (P < 0.001), described by the equation SD(pred err) = -1.017 + 0.286 x Y(cal). Our calibrated model will give better predictions for children with IGHD fulfilling the same criteria.

Calibration↗

PubMedDate not supplied

In vitro phototoxicity testing: development and validation of a new concentration response analysis software and biostatistical analyses related to the use of various prediction models.

As demonstrated in several validation studies, the dermal phototoxic potential of chemicals in humans can be effectively assessed by in vitro methods. The core of these methods is to monitor dose-response curves of a chemical in the absence and presence of light, to quantify the difference between these two curves by appropriate measures (either the photo-irritancy factor [PIF], or the mean photo effect [MPE]), and to use these measures as predictors of in vivo phototoxicity. We present new concentration-response analysis software for in vitro phototoxicity testing, which runs on current personal computers, and takes into account all the limitations identified when using a former program. We also demonstrate the validity and robustness of this new software by applying it retrospectively to all data available from two phases of the EU/COLIPA validation trial for the 3T3 neutral red update in vitro phototoxicity test. Some frequently raised questions pertaining to the use of prediction models in phototoxicity testing are addressed, including: the necessity of using prediction models based on a cut-off; whether it is justifiable to use sharp prediction cut-off values; whether there is a biostatistical justification for the highest concentration of the test chemical; and whether repeated testing of a chemical is required.

3T3 Cells↗

PubMed2001

Predictive modeling care management program produces dramatic bottom line results.

Predictive modeling program identifies high-risk patients before complications lead to skyrocketing costs. It's a new-generation effort, but it has already proven its worth in a large population, producing savings of $1 million to $2 million per month. The effort begins with the development of a constantly evolving registry of high-risk patients. Specialized case management efforts are then focused on stabilizing individuals within this group. See how traditional interventions have been combined with new technology to deliver results.

Algorithms↗

PubMed2003

Quantitative predictive models for octanol-air partition coefficients of polybrominated diphenyl ethers at different temperatures.

Quantitative predictive models for octanol-air partition coefficients of polybrominated diphenyl ethers at different environmental temperatures (T) were developed. Partial least squares (PLS) regression was used for model development. A list of 18 theoretical molecular structural descriptors was screened by PLS analysis. The optimal model was selected from the one containing nine theoretical molecular descriptors and 1/T as predictor variables. The cross-validated Q(2)(cum) value for the optimal model is 0.975, indicating a good predictive ability and stability of the model. Intermolecular dispersive interactions play a leading role in governing the magnitude of logK(OA). The lower the E(LUMO) (the energy of the lowest unoccupied molecular orbital), the greater the intermolecular interactions between octanol and PCB molecules, and thus the greater the logK(OA) values.

Environmental Pollutants↗

PubMed1996

Evaluation of the white finger risk prediction model in ISO 5349 suggests need for prospective studies.

The risk prediction model for white fingers in Annex A of ISO 5349 is not likely to offer protection from all tools and all work processes. It is also probable that some work place changes it has initiated are either redundant or lack the intended effect. The main reasons for these shortcomings are the following. The often demonstrated disagreement between predicted and observed white fingers occurrence may be related to the fact that the model is based on latency data. This leads to an overestimation, to an unknown extent, of true group risks. A possible healthy worker effect, resulting in underestimation, has not been considered, and uncertainty because of recall bias is connected with using latency as effect variable in a slowly developing disorder like white fingers. The diagnostic criteria for white fingers have varied over the years, causing a possible inclusion of circulatory disturbances other than those induced by vibration. Among insufficiently clarified matters unrelated to vibration are variations in individual susceptibility and other host factors that modify vibration effects, uncertainty concerning daily or total effective exposure, and the fact that variation in work methods and processes as well as ergonomic factors other than vibration tend to make different groups incomparable form the viewpoint of risk of injury. Lack of sufficient data on vibration measurements and employment durations add to the uncertainty, as do variations in tool conditions (grinder wheels, etc) and inherent difficulties in measurement. Finally, the ISO 5349 frequency-weighting curve only relates to acute sensory effects rather than chronic effects on vascular functions like white fingers, and directional difference in sensitivity has not been incorporated in the curve. Data on exposure-response relationships are needed from prospective studies that monitor the dose of exposure to special vibration types and all relevant environmental agents, employ diagnostics with good sensitivity, specificity and predictive value, and pay attention to environmental or individual confounding factors and effect modifiers. Before such data are available, the ISO 5349 model should not be used for risk prediction. It can serve, however, as an incentive for manufacturers to produce tools that vibrate less, and for employers to implement practical measures in order to reduce the total and dose of effective exposure.

Fingers↗

PubMed1997

Analyzing predictive models following definitive radiotherapy for prostate carcinoma.

BACKGROUND: As we approach the 21st century, clinically useful predictive models for prostate carcinoma are urgently needed to stratify patients reliably for future treatment strategies. Recently, many investigators have developed models that employ prostate specific antigen (PSA)-based constructs or groupings in an attempt to predict outcome accurately following definitive radiotherapy. This investigation was conducted to determine which of these models provides the closest "fit" to independent clinical outcome data measuring biochemical freedom from failure (bNED control), thereby warranting further exploration. METHODS: Six models were analyzed in a definitive radiotherapy series of 421 patients with localized prostate carcinoma treated with a median dose of 74 Gray (Gy) between March 1988 and November 1994. A stepwise Cox proportional hazards multivariate analysis (MVA) was performed to predict for bNED control using the following covariates: PSA, Gleason's score, stage, dose, PSA density, and perineural invasion. Subsequent MVAs were performed for each model incorporating the new construct or prognostic groupings. The adequacy of the models was confirmed using plots of score residuals against time to bNED failure and comparisons were made used Akaike's Information Criteria (AIC) in which a smaller value corresponds to a statistically improved model based on explained variation and the number of predictors. Because PSA was distributed in a log-normal fashion in the current study population, the model-building process was duplicated using a logarithmic transformation analysis. Biochemical failure was defined as 2 consecutive elevations in the PSA > or = 1.5 ng/mL. The median follow-up time was 34 months (range, 2-87 months). RESULTS: Initially, the model developed by Pisansky et al. appeared the most predictive due to the parsimony in their risk estimate, which is the sole predictor of outcome, as well as its associated lowest AIC value. However, after the logarithmic transformation analysis, all the models appeared to be equally predictive of bNED outcome. CONCLUSIONS: A plethora of accurate models for predicting outcome following definitive radiotherapy for prostate carcinoma recently have been engineered, all of which are essentially equally predictive in this data base (via a logarithmic conversion process). This analysis should be corroborated in other large radiotherapy series.

Follow-Up Studies↗

PubMed1994

Development of a predictive model for optimal zona pellucida binding using insemination volume and sperm concentration.

OBJECTIVE: To develop a predictive model under hemizona assay (HZA) conditions for human spermatozoa concentrations and insemination volume for optimum zona pellucida (ZP) binding. DESIGN: Analysis of 20 different insemination volumes for zona binding and sperm morphology under HZA conditions. SETTING: Reproductive biology unit, tertiary medical center. PATIENTS: Four proven fertile sperm donors. MAIN OUTCOME MEASURES: 5-, 20-, 50-, 80-, and 100-microL droplets were analyzed with four different concentrations of 0.5 x 10(6), 1.0 x 10(6), 2.0 x 10(6), and 4.0 x 10(6) cells/mL to determine the number of sperm bound to each hemizona. Fifteen hemizonae were used for each insemination volume or microdroplet. Response surface regression model with volume and concentration as the regressor variables has been used. RESULTS: The response surface of binding for the factors concentration and volume showed nonlinear association. A formula, indicating the optimal sperm insemination volume for maximum sperm binding to the ZP, Vmax = -(b1 + b5c)/2b6c, is described. The transformed data indicated 60 microL containing 4 x 10(6) sperm/mL to be optimal. Although morphology of zona spermatozoa is superior compared with seminal and postswim-up samples, no difference among the percentage of the normal morphology in different microdroplets could be demonstrated. CONCLUSION: Optimal volume for the obtained concentration of spermatozoa from a patient can be calculated and therapeutically used for cases of severe oligozoospermic patients by microvolume inseminations in IVF practice.

Female↗

PubMed1995

Comparison of two predictive models for prognosis in critically ill patients in a Veteran's Affairs Medical Center coronary care unit.

STUDY OBJECTIVE: The acute physiologic and chronic health evaluation score has been developed to assess prognosis in critically ill patients. Knaus et al initially determined and validated diagnosis-specific coefficients for prediction of outcome in a group of multidisciplinary ICUs. Teskey et al found different coefficients for cardiac diagnoses in a retrospective analysis of coronary care unit (CCU) only patients. This study compares the actual mortality in a Veteran's Affairs Medical Center (VAMC) CCU with the mortality predicted by the two equations. DESIGN AND SETTING: Data were prospectively collected for patients admitted to the medical CCU at a university-affiliated, tertiary care VAMC. PATIENTS: Patients (n = 338) admitted to the CCU with the diagnoses of coronary artery disease (CAD), myocardial infarction (MI), congestive heart failure (CHF), arrhythmia (Arr), and other cardiac-related diagnoses. RESULTS: The entire CCU population showed no significant differences from either the predictions by Knaus et al and Teskey et al in 1991. However, when specific disease states were analyzed as a whole, significant differences from both prediction models for CAD and MI actual mortality were found. Teskey et al was a better predictor for the Arr and CHF population, while both were equally reliable for other. CONCLUSIONS: Either prediction model is reliable for a CCU population in general. However, diagnosis-specific coefficients of Teskey et al appear to correlate better with actual mortality for this VAMC. Significant outcome differences compared with those predicted may reflect the patient population specific to a VAMC.

APACHE↗