Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “predictive modelling”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

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

At least 127 records · Page 7Linked to original sources
PubMed1989

PPS-87: a new event oriented solar proton prediction model.

A new event-oriented solar proton prediction model has been developed and implemented at the USAF Space Environment forecast facility. This new model generates predicted solar proton time-intensity profiles for a number of user adjustable energy ranges and is also capable of making predictions for the heavy ion flux. The computer program is designed so a forecaster can select inputs based on the data available in near real-time at the forecast center as the solar flare is occurring. The predicted event amplitude is based on the electromagnetic emission parameters of the solar flare (either microwave or soft X-ray emission) and the solar flare position on the sun. The model also has an update capability where the forecaster can normalize the prediction to actual spacecraft observations of spectral slope and particle flux as the event is occurring in order to more accurately predict the future time-intensity profile of the solar particle flux. Besides containing improvements in the accuracy of the predicted energetic particle event onset time and magnitude, the new model converts the predicted solar particle flux into an expected radiation dose that might be experienced by an astronaut during EVA activities or inside the space shuttle.

Computer Simulation↗
PubMed2001

Factors influencing predictive models for toxicology.

Comparisons of different models to predict toxicity and evaluation of the predictive power of a model are affected by the variability of the data. We assessed this problem by considering experimental toxicity data and chemical descriptors. We evaluated several toxicological end-points (Oncorhynchus mykiss, Daphnia magna, Acceptable Daily Intake, Anas Platyrhynchos, Colinus virginianus and Muridae) in the case of pesticides and also considered the availability of toxicological data. We calculated hundreds of molecular descriptors (divided into constitutional, electrostatic, geometrical, quantum-chemical and topological ones) for the selected compounds using CODESSA, HyperChem and Pallas. Molecular descriptors may vary depending on the conformation of the molecules and on the software used. We evaluated the extent of this variability, and compared it with the variability of the experimental toxicological values.

Animals↗
PubMed2003

A novel MHCp binding prediction model.

Many statistical and molecular mechanics models have been developed and tested for major histocompatibility complex peptide (MHCp) binding predictions during the last decade. The statistical model prediction using pooled peptide sequence data and three-dimensional modeling prediction by molecular mechanics calculations have been assessed for efficiency and human leukocyte antigen diversity coverage. We describe a novel predictive model using information gleaned from 29 human MHCp crystal structures. The validation for the new model is performed using four different sets of data: (1) MHCp crystal structures, (2) peptides with known IC(50) binding values, (3) peptides tested positive by tetramer staining, (4) peptides with known binding information at the MHCBN database. The model produces high prediction efficiencies (average 60 %) with good sensitivity (approximately 50%-73%) and specificity (52%-58%) values. The average positive predictive value of the model is 89%, while the average negative predictive value is only 18%. The efficiency is very high in predicting binders and very low in predicting nonbinders. This model is superior to many existing methods because of its potential application to any given MHC allele whose sequence is clearly defined.

Amino Acid Sequence↗
PubMed2004

A predictive model of moment-angle characteristics in human skeletal muscle: application and validation in muscles across the ankle joint.

In the present work, a generic model for the prediction of moment-angle characteristics in individual human skeletal muscles is presented. The model's prediction is based on the equation M = V x Lo(-1)sigma c cos phi x d, where M, V, and Lo are the moment-generating potential of the muscle, the muscle volume and the optimal muscle fibre length, respectively, and sigma, phi and d are the stress-generating potential of the muscle fibres, their pennation angle and the tendon moment arm length, respectively, at any given joint angle. The input parameters V, Lo, sigma, phi and d can be measured or derived mechanistically. This eliminates the common problem of the necessity to estimate one or more of the input parameters in the model by fitting its outcome to experimental results often inappropriate for the function modelled. The model's output was validated by comparisons with the moment-angle characteristics of the gastrocnemius (GS) and tibialis anterior (TA) muscles in six men, determined experimentally using voluntary contractions at several combinations of ankle and knee joint angles for the GS muscle and electrical stimulation for the TA muscle. Although the model predicted realistically the pattern of moment-angle relationship in both muscles, it consistently overestimated the GS muscle M and consistently underestimated the TA muscle M, with the difference gradually increasing from dorsiflexion to plantarflexion in both cases. The average difference between predicted and measured M was 14% for the GS muscle and 10% for the TA muscle. Approximating the muscle fibres as a single sarcomere in both muscles and failing to achieve complete TA muscle activation by electrical stimulation may largely explain the differences between theory and experiment.

Analysis of Variance↗
PubMed1986

Development of a predictive model for symptomatic neuropathy in diabetes.

In evaluating therapeutic interventions aimed at preventing diabetic neuropathy, choosing a suitable measure of neural function is difficult, partly because the relation between most available objective measures and the development of symptomatic neuropathy (SN) is unclear. Using data from 67 diabetic patients, we developed a linear logistic regression model to assess the relationship between SN and a set of neural measures, including dark-adapted pupil size; pupillary latency; heart rate; a measure of respiratory sinus arrhythmia (R); the Valsalva ratio; and conduction velocities for the peroneal, median motor, and median sensory nerves. Models allowed for possible effects related to age, sex, duration and type of diabetes, glycosylated hemoglobin, and adiposity. Thirty-two of the patients reported SN (autonomic and/or sensorimotor). The best-fitting model for predicting the presence of any SN included only heart rate, duration of disease, and R. Exclusion of duration (P less than .01), or heart rate (P less than .05), or R (P less than .001) significantly impaired the fit of the model. To evaluate the temporally predictive power of the model, nine of the asymptomatic patients were reinterviewed 2 yr later by the same interviewer, who was unaware of the results of the modeling. Four of five to whom the model had assigned high probability of symptoms had indeed developed SN during the follow-up period, whereas none of the four assigned low probability had developed SN (P less than .05). Thus it seems that a measure of respiratory sinus arrhythmia provides an index of neural function strongly related to SN, and our follow-up data suggest that diminished R can be used to predict the development of SN in diabetes.

Adult↗
PubMed2026

A machine learning-based predictive model for radiosensitivity in nasopharyngeal carcinoma utilizing serum proteomics.

BACKGROUND: Nasopharyngeal carcinoma (NPC) remains highly sensitive to radiotherapy; however, radioresistance in a subset of patients leads to local recurrence and distant metastasis. Serum proteomics provides a minimally invasive approach to capturing dynamic physiological changes, and machine learning enables efficient construction of predictive models. This study aimed to develop and validate a serum proteomics–based machine-learning model for predicting radiotherapy sensitivity in nasopharyngeal carcinoma (NPC). METHODS: Pretreatment serum samples from newly diagnosed NPC patients were analyzed using SELDI-TOF-MS. Differentially expressed proteins between radiosensitive and radioresistant groups were identified using limma. GO and KEGG analyses were performed to explore functional enrichment. Twelve machine-learning algorithms were used to construct predictive models, and the top-performing models were optimized through feature selection. A Random Forest model with seven features was identified as the optimal model. External validation was performed using an independent cohort with ELISA-quantified protein levels. Model performance was assessed using Receiver operating characteristic curve (ROC), calibration analysis, decision curve analysis (DCA), and 10-fold cross-validation. SHapley Additive exPlanations (SHAP) analysis was applied for model interpretability, and the final model was deployed via a ShinyAPP. RESULTS: A total of 96 differentially expressed proteins were identified, which involved multiple function and signaling pathways. The Random Forest model demonstrated the best predictive performance, achieving an area under the curve (AUC) of 0.963 in the training set and 0.975 in the validation set. Cross-validation yielded an average AUC of 0.965. DCA indicated high clinical utility across a broad threshold range, and calibration curves showed good model agreement. Seven proteins (PLXND1, GSR, PGD, PTPRC, OR2T29, ACTG2, CHAD) were selected as final features. SHAP analysis provided global and individual-level interpretability. A web-based tool was developed to facilitate clinical application. CONCLUSION: This study establishes a robust serum proteomics–based machine-learning model capable of accurately predicting radiotherapy sensitivity in NPC. The model offers clinical interpretability and practical implementation, supporting personalized radiotherapy decision-making.

Humans↗
PubMed2005

A predictive model for cesarean section in low risk pregnancies.

OBJECTIVE: A small number of women with low risk pregnancies undergo cesarean section. A model that can predict this risk and therefore identify these women will be of help in several hospitals where personnel and resources are limited. METHODS: The study consisted of 2 parts. All charts of women with low risk singleton pregnancies admitted to labor room over a 5-month period were analyzed. Adjusted odds ratios were calculated to find out relative importance of each risk factor and likelihood ratios were obtained. These were prospectively applied to 1010 consecutive low risk women and the post test probability calculated. Finally the actual incidence of cesarean section was compared with posttest probability derived from predictors. RESULTS: A combination of maternal age >24 years, primiparity and height <150 cm or a combination of any 2 of the 3 variables is significantly associated with increased cesarean section rate. Individually, primiparity, height <150 cm or age >24 years also significantly increased the chances of cesarean section. CONCLUSIONS: A predictive model consisting of maternal age, parity and height can be used to identify low risk pregnant women who are likely to require cesarean section.

Adult↗
PubMed1987

Validation of the mortality prediction model for ICU patients.

We tested recently developed admission and 24-h models of hospital mortality on 1,997 consecutive admissions to a general medical/surgical ICU. This study population was independent of the group used to develop the models. The admission prediction model estimated each patient's probability of hospital mortality based on seven routinely collected admission variables. The 24-h model utilized seven variables routinely available at 24 h in the ICU. The admission model accurately described the mortality experience of the new cohort, while the 24-h model did not. Advantages of the admission model are that it is evaluable at the time of ICU admission, is independent of ICU treatment, and can be used to stratify patients by severity of illness, thereby making ICU comparisons possible. Its excellent goodness-of-fit, correct classification rate, sensitivity, and specificity suggest that this model is now ready for multihospital testing.

Aged↗
PubMed1994

Low-dose ara-C in myelodysplastic syndromes (MDS) and acute leukemia following MDS: proposal for a predictive model.

Patients with myelodysplastic syndromes (MDS) comprise an extremely heterogeneous group. There is a need for decision models both for predicting the natural course of the disease and the outcomes of different treatment alternatives. In 102 consecutive patients with MDS or acute myelogenous leukemia (AML) following MDS, pre-treatment variables were studied in relation to the response to treatment with low-dose ara-C. Thirty patients (29%) responded with either a complete remission or a significant rise in the hemoglobin level. For the remaining 71%, the treatment was ineffective and in some cases hazardous. The factors associated with a poor response to treatment could be divided into two groups: one included low platelet counts and the presence of chromosomal aberrations, both signs of progressive MDS with a short survival, and the other comprised morphological findings, indicating ineffective hemopoiesis. Patients with platelet counts > 150 x 10(9)/l had a response rate of 55%, compared to 24% in patients with subnormal platelet counts. Logistic regression identified low bone marrow cellularity, absence of ring sideroblasts and < 2 chromosomal aberrations as predictors of a favourable response in patients with platelet counts < 150 x 10(9)/l. These factors and the platelet count were combined in a predictive model which divided patients into three groups with different probabilities of response: one favourable (38% of the patients), with a response rate of > 50%; a second, intermediate group (33% of the patients), with a response rate of 24%; and a third, unfavourable group (29% of the patients) with only 3% responses.(ABSTRACT TRUNCATED AT 250 WORDS)

Acute Disease↗
PubMed1990

Fracture prediction models for osteoporosis prevention.

Fracture prediction models have been developed that can estimate an individual's cumulative or lifetime fracture risk on the basis of bone mineral measurements and other factors. Examples illustrate how estimates of cumulative fracture risk might be influenced by initial age and level of bone mass, as well as by anticipated bone loss rate. New data regarding bone loss rates, fracture costs, and fracture risk factors should be incorporated to improve the accuracy of existing models. Such models could then be used for cost-benefit analysis to explore optimal treatment strategies.

Bone Density↗
PubMed2005

Predictive modelling of the individual and combined effect of water activity and temperature on the radial growth of Fusarium verticilliodes and F. proliferatum on corn.

The major objective of this study was to develop validated models to describe the effect of a(w) and temperature on the radial growth on corn of the two major fumonisin producing Fusaria, namely Fusarium verticilliodes and F. proliferatum. The growth of these two isolates on corn was therefore studied at water activities between 0.810-0.985 and temperatures between 15 and 30 degrees C. Minimum a(w) for growth was 0.869 and 0.854 for F. verticilliodes and F. proliferatum, respectively. No growth took place at a(w) values equal to 0.831 and 0.838 for F. verticilliodes and F. proliferatum, respectively. The colony growth rates, g (mm d(-1)) were determined by fitting a flexible growth model describing the change in colony diameter (mm) with respect to time (days). Secondary models, relating the colony growth rate with a(w) or a(w) and temperature were developed. A third order polynomial equation and the linear Arrhenius-Davey model were used to describe the combined effect of temperature and a(w) on g. The combined modelling approaches, predicting g (mm d(-1)) at any a(w) and/or temperature were validated on independently collected data. All models proved to be good predictors of the growth rates of both isolates on maize within the experimental conditions. The third order polynomial equation had bias factors of 1.042 and 1.054 and accuracy factors of 1.128 and 1.380 for F. verticilliodes and F. proliferatum, respectively. The linear Arrhenius-Davey model had bias factors of 0.978 and 1.002 and accuracy factors of 1.098 and 1.122 for F. verticilliodes and F. proliferatum, respectively. The results confirm the general finding that a(w) has a greater influence on fungal growth than temperature. The developed models can be applied for the prevention of Fusarium growth on maize and the development of models that incorporate other factors important to mould growth on maize.

Adsorption↗
PubMed2006

Pregnancy prediction models and eSET criteria for IVF patients--do we need more information?

PURPOSE: The purpose of the present study was to evaluate statistical prediction models and simple allocation criteria, based on predictors for pregnancy, as tools to identify a good prognosis group in a possible eSET setting. METHODS: A pregnancy prediction model based on logistic regression models was generated by analysis of 1675 DET treatment cycles. The model was evaluated and compared to simple eSET allocation criteria. RESULTS: Embryo quality, patient age, and basal FSH were identified as significant predictors (at 5% significance level) of pregnancy. Although comparable to previously generated models, the predictive ability of the present model was relatively poor and practically similar to simple allocation criteria based on age and embryo quality. CONCLUSIONS: Existing prediction models, or simple allocation criteria, are limited in identifying good prognosis patients. Future studies of the applicability of improved pregnancy prediction models will need very comprehensive and detailed patient and embryo information.

Adult↗
PubMed1998

Escape of autocrine ligands into extracellular medium: experimental test of theoretical model predictions.

We have developed an experimental system for testing mathematical model predictions concerning escape of autocrine ligands into the extracellular bulk medium. This system employs anti-receptor blocking antibodies against the epidermal growth factor receptor (EGFR)/transforming growth factor alpha (TGFalpha) receptor/ligand pair. TGFalpha was expressed under the control of a tetracycline-repressed promoter, together with a constitutively expressed human EGFR in B82 mouse fibroblast cells. This expression system allowed us to vary TGFalpha synthesis rates over a roughly 300-fold range by adjusting tetracycline concentration. TGFalpha accumulation in the extracellular bulk medium was then measured as a function of cell density, TGFalpha synthesis rate, and anti-EGFR blocking antibody concentration. Consistent with model predictions, amounts of ligand in the medium on a per cell basis were found to diminish as cell density was increased but with reduced dependence on cell density at higher ligand synthesis rates. Similarly consistent with model predictions, higher ligand synthesis rates also decreased the effect of anti-receptor blocking antibodies. Our investigation has established that we can successfully analyze and understand autocrine ligand secretion behavior from the basis of our theoretical model.

Animals↗
PubMed2003

A prediction model for targeting low-cost, high-risk members of managed care organizations.

OBJECTIVE: To describe the development and validation of a predictive model designed to identify and target HMO members who are likely to incur high costs. STUDY DESIGN: Split-sample multivariate regression analysis. PATIENTS AND METHODS: We studied enrollees in a 350000-member HMO with > or = 1 claim in 1998 and 1999. The prediction model uses a combination of clinical and behavioral vaiables and 1998 and 1999 claims data. The prediction model was applied and used to rank low-cost patients (1998 cost < dollars 2000) according to their estimated probability of incurring costs > or = dollars 2000 in 1999. For prospective testing, we applied our models to data that are not available in advance. The same prediction model was applied to rank a different set of low-cost patients (1999 cost < dollars 2000) according to estimated probability of incurring costs > or = dollars 2000 in 2000. Because the predictions were used for disease management purposes, the outcomes of a randomly selected control group not intervened on for the disease management program was analyzed. The predictive accuracy of the model was tested by comparing the percentages of "targeted" vs all low-cost patients who incurred high costs in the subsequent year. RESULTS: Of the low-cost, top-ranked 1998 patients, 47.8% incurred high (> or = dollars 2000) medical expenses in 1999 vs 14.2% of randomly selected patients who were low cost in 1998. Of the top-ranked 1999 patients, 39.7% incurred high costs in 2000 vs 12.2% of the randomly selected low-ranked patients. CONCLUSIONS: The prediction model successfully identifies low-cost, high-risk patients who are likely to incur high costs in the next 12 months.

Female↗
PubMed2026

Predictive Models for Hypoglycemia Risk in Haemodialysis Patients With Diabetic Kidney Disease: Systematic Review and Meta-Analysis.

AIM: To provide evidence for selecting and developing reliable clinical assessment tools for hypoglycemia in diabetic kidney disease patients during haemodialysis. DESIGN: Review. METHODS: Systematic searches were performed in 9 Chinese and English databases to collect literature regarding the development of hypoglycemia risk prediction models in haemodialysis patients with diabetic kidney disease. Two reviewers independently performed literature screening, data extraction, risk-of-bias assessment, and applicability evaluation. The Prediction Model Risk of Bias Assessment Tool was used to assess the risk of bias and applicability of the included studies. Meta-analysis was conducted using R software. DATA SOURCES: CNKI, Wanfang, VIP, CBM, PubMed, Cochrane Library, EMbase, Web of Science, and CINAHL. The search period covered from the establishment date of each database to December 2025. RESULTS: Six studies, comprising six prediction models, were included. Two studies performed internal validation, and three conducted external validation. All models reported the area under the curve, ranging from 0.813 to 0.866, and calibration measures. Four studies were rated as having a high risk of bias, while all six demonstrated good overall applicability. The meta-analysis showed that the pooled AUC value of the six studies was 0.846 (95% CI: 0.823-0.867). CONCLUSION: Research on hypoglycemia risk prediction models in haemodialysis patients with diabetic kidney disease remains in the developmental stage. Although the included prediction models exhibited satisfactory apparent discriminatory ability and clinical applicability, most of the original studies suffered from a high risk of bias and lacked adequate validation. The true predictive performance and clinical application value of these models remain to be further verified. Accordingly, routine and unconditional clinical application is not recommended at this stage. Future studies should include more high-quality, multicenter external validation and develop models with high generalizability, favourable clinical applicability, and robust predictive performance to facilitate early identification of hypoglycemia risk in this population. IMPACT: This study systematically evaluated the hypoglycemia risk prediction models for diabetic kidney disease patients during haemodialysis, and the research on hypoglycemia risk prediction models for maintenance haemodialysis patients during dialysis is still in the development stage. This study provides a reference for clinical medical staff to select or develop hypoglycemia risk prediction and assessment tools for diabetic kidney disease patients during haemodialysis. REPORTING METHOD: This study was conducted in accordance with the relevant guidelines of the EQUATOR Network and followed the TRIPOD-SRMA Checklist. PATIENT OR PUBLIC CONTRIBUTION: No patient or public contribution. TRIAL REGISTRATION: PROSPERO: CRD420251243352.

Humans↗
PubMed2002

Development of predictive models for long-term cardiovascular risk associated with systolic and diastolic blood pressure.

Most existing risk prediction models have not considered the joint contribution of systolic and diastolic blood pressure to cardiovascular risk, and some suggest that there are thresholds below which further reductions of blood pressure yield no additional benefit. We developed multivariate risk prediction models that quantify the risk associated with both systolic and diastolic blood pressure and that can be used to infer the benefits of antihypertensive therapy in populations. Two large clinical trial cohorts, the Physicians' Health Study, composed of 22 071 males (mean age, 53.2 years; median follow-up, 13.0 years), and the Women's Health Study, composed of 39 876 females (mean age, 53.8 years; median follow-up, 6.2 years), were used to develop gender-specific predictive models via Cox regression. End points included myocardial infarction, stroke, coronary artery bypass, angioplasty, and cardiovascular death. Risk reduction estimates were derived by computing reductions associated with incremental lowering of systolic and diastolic blood pressures. In both populations, lower levels of blood pressure predicted lower event rates, with no evidence of a plateau or a J-shaped curve. In males, both systolic and diastolic blood pressures were significantly associated with events (P<0.001), whereas in females, only systolic blood pressure (P<0.001) predicted outcome after multivariate adjustment. Correction for measurement error in blood pressure increased risk estimates by approximately 50%. Differences in systolic blood pressure yielded greater relative risk reductions than did differences in diastolic blood pressure in a combined population of males and females. These predictive models may be useful for risk estimation associated with hypertension in similar populations and may also be used to infer the benefits of antihypertensive therapy.

Adult↗
PubMed2005

A hybrid generative and predictive model of the motor cortex.

We describe a hybrid generative and predictive model of the motor cortex. The generative model is related to the hierarchically directed cortico-cortical (or thalamo-cortical) connections and unsupervised training leads to a topographic and sparse hidden representation of its sensory and motor input. The predictive model is related to lateral intra-area and inter-area cortical connections, functions as a hetero-associator attractor network and is trained to predict the future state of the network. Applying partial input, the generative model can map sensory input to motor actions and can thereby perform learnt action sequences of the agent within the environment. The predictive model can additionally predict a longer perception- and action sequence (mental simulation). The models' performance is demonstrated on a visually guided robot docking manoeuvre. We propose that the motor cortex might take over functions previously learnt by reinforcement in the basal ganglia and relate this to mirror neurons and imitation.

Brain Mapping↗
PubMed2005

A simple predictive model for blood-brain barrier penetration.

A simple two-descriptor model to predict blood-brain barrier penetration is derived from a training set of 79 compounds: log BB = - 13.31V2 + 9.601V - 2.231PSA - 0.5290 (n = 79, r2 = 0.83) where log BB is the logarithm of the ratio of the steady-state concentration of the compound in the brain to in the blood, V (nm3) is the molecular volume, PSA (nm2) is the polar surface area which is defined as the sum of the van der Waals surface areas of oxygen atoms, nitrogen atoms, and attached hydrogen atoms in a molecule, n is the number of compounds, and r is the correlation coefficient. The model is validated by a leave-one-out procedure and an external test set (25 compounds). The results indicate that the model developed is statistically sound and is sufficiently reliable and robust for predictive use. The descriptors in the model can be easily computed and it is suitable for the rapid prediction of the blood-brain barrier penetration for a wide range of drug candidates.

Blood-Brain Barrier↗