SEARCH · PubMed Health

Results for “Linear Regression”

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 199 records · Page 11Linked to original sources

PubMed2006

Linear and non-linear regression analysis for the sorption kinetics of methylene blue onto activated carbon.

Batch kinetic experiments were carried out for the sorption of methylene blue onto activated carbon. The experimental kinetics were fitted to the pseudo first-order and pseudo second-order kinetics by linear and a non-linear method. The five different types of Ho pseudo second-order expression have been discussed. A comparison of linear least-squares method and a trial and error non-linear method of estimating the pseudo second-order rate kinetic parameters were examined. The sorption process was found to follow a both pseudo first-order kinetic and pseudo second-order kinetic model. Present investigation showed that it is inappropriate to use a type 1 and type pseudo second-order expressions as proposed by Ho and Blanachard et al. respectively for predicting the kinetic rate constants and the initial sorption rate for the studied system. Three correct possible alternate linear expressions (type 2 to type 4) to better predict the initial sorption rate and kinetic rate constants for the studied system (methylene blue/activated carbon) was proposed. Linear method was found to check only the hypothesis instead of verifying the kinetic model. Non-linear regression method was found to be the more appropriate method to determine the rate kinetic parameters.

Adsorption↗

PubMedDate not supplied

Non-linear regression and variance ratio analysis of time based NMR data.

Biomedical NMR experiments rely frequently on data obtained sequentially over time. A method is presented for analysis of time based NMR data, which allows modelling of continuous and discontinuous functions to observed intensity changes by non-linear regression and which uses variance ratio analysis to compare these models statistically. The method eliminates many of the usual problems in the parametric analysis of experimental values obtained at discrete time points and of comparison of the coefficients of model functions which require unsubstantiated assumptions about the distribution of parameters and ignore internal correlations which may exist between such parameters. The variance ratio method is illustrated for multiple time courses obtained with 23Na NMR of perfused rat kidney undergoing hypoxic perturbation in the presence of different treatments.

Amino Acids↗

PubMedDate not supplied

Prediction of the left ventricular ejection fraction response to doxorubicin using a multiple linear regression model.

The aim of this study was to develop a model for predicting clinically significant deterioration in the left ventricular ejection fraction due to chronic doxorubicin administration. Twenty-six patients were monitored during their courses of doxorubicin chemotherapy with serial gated equilibrium radionuclide angiography. Multiple linear regression analysis was used to derived the best combination of clinical and radionuclide angiographic predictors of resting left ventricular ejection fraction at any point during the course of chemotherapy. The final model consisted of five variables: left ventricular ejection fraction at the previous monitoring point; cumulative dose of doxorubicin achieved at the previous monitoring point; increment in dose from the previous monitoring point; age of the patient; and time to peak left ventricular emptying at the previous monitoring point. The cumulative dose, the ejection fraction at the previous monitoring point and the final model, respectively, explained 11%, 33% and 53% of the variability in ejection fraction determinations during the 26 patient courses. The final model also forecast a potentially very low resting left ventricular ejection fraction (less than 35%) at the cumulative doses of doxorubicin which provoked serious clinical cardiotoxicity in two patients. A multivariate model is a useful aid in timing discontinuation of doxorubicin prior to the development of a clinically significant deterioration in left ventricular ejection fraction.

Computer Simulation↗

PubMed2004

Navy Asbestos Medical Surveillance Program (1991-1999): linear regression analysis for the effect of asbestos exposure on pulmonary function testing.

The effect of asbestos exposure on pulmonary function was studied using data from the Navy Asbestos Medical Surveillance Program. Records were selected for Caucasian men from 1991 to 1999 (N = 89,318) and were analyzed using a cross-sectional, linear regression model. Dependent variables were forced expiratory volume in 1 s (FEV1) and forced vital capacity (FVC), with independent continuous variables of age, height, weight, smoking, and asbestos history. Overall, the continuous variable for asbestos exposure demonstrated significant protection of +1.1 cm3/year (t = 3.278, p = 0.001) for FEV1 and +1.6 cm3/year (t = 4.225, p = 0.000) for FVC. There was significant interaction between asbestos exposure and smoking history (FEV1, -0.09 cm3/year2, t = -6.467, p = 0.000; FVC, -0.097 cm3/year2, t = -5.663, p = 0.000). This study suggests that workers within the program demonstrated minimal additional pulmonary function changes during the period, particularly if they do not smoke tobacco. The study also supports continuing smoking cessation efforts for all asbestos-exposed workers.

Adult↗

PubMed1995

Simulation of the 13C nuclear magnetic resonance spectra of trisaccharides using multiple linear regression analysis and neural networks.

Predictive models are developed for the 13C NMR chemical shifts of the carbon atoms comprising the central rings of 46 trisaccharide compounds. Thirty-nine trisaccharides are used as a training set for development of models using regression analysis and computational neural networks, and seven compounds are used as an external prediction set. The descriptors used in the models are developed directly from the molecular structures of the trisaccharides. Three different methods of descriptor selection are compared. The dependence of the models on the geometries of the trisaccharides is explored. The models developed with geometric descriptors are better than those developed without geometric descriptors, although the latter models are still of a comparable quality. Overall, the best model found is a neural network based on descriptors selected by multiple linear regression.

Algorithms↗

PubMed2000

A linear regression model to predict the pH of neonatal parenteral nutrition solution.

INTRODUCTION: Providing the high calcium intake necessary for normal bone mineralization in rapidly growing very low birth weight infants is difficult because calcium and phosphorus solubility is limited in the range of parenteral nutrition pH. A major determinant of calcium and phosphorus solubility in vitro is solution pH. The objective of this study was to develop and assess the accuracy of a method to predict the final parenteral solution pH as a linear function of the individual parenteral component concentrations. METHODS: pH values were measured for 205 neonatal parenteral nutrition solutions prepared during a 5-week period. Concentrations of the 13 components used to synthesize parenteral nutrition were determined for each solution. Data from 135 samples were used to develop a linear regression coefficient model with pH as the dependent variable. From the regression model the pH was predicted for the remaining 70 samples using the seven significant solution component concentrations, and the predicted and measured solution pH values were compared. RESULTS: The mean measured parenteral nutrition pH for all solutions was 5.364 +/- 0.110 (mean +/- SD, range 5.03-5.73). The absolute mean pH difference between the predicted and measured value for the 70 test samples was 0.04 +/- 0.04. pH estimated with the model correlated highly with measured pH (r2 = 0.77). The seven components in the regression model accounted for 81% of the pH variance. CONCLUSION: The pH of neonatal parenteral nutrition solutions can be predicted accurately as a linear function of the solution concentrations of the following seven components: sodium acetate, sodium phosphate, potassium phosphate, potassium acetate, magnesium sulphate, amino acid solution and dextrose. The absolute mean difference between measured pH and predicted pH was 0.04. Applying this method to estimate pH with the interactive properties of computer-based ordering systems could enhance calcium and phosphorus administration to very low birth weight infants.

Calcification, Physiologic↗

PubMed1978

Discrimination between descriptive models of L-glutamate uptake by the retina using non-linear regression analysis.

1. The uptake of labelled L-glutamate by the isolated rat retina was measured over a large range of external concentrations (1 micron to 1 mM). 2. The results obtained from measurements of the initial velocity of L-glutamate uptake at different concentrations did not follow simple hyperbolic kinetics. 3. The error structure of replicate velocity measurements was examined and found to be normally distributed and heteroscedastic. 4. Descriptive models were fitted directly to data, weighted by the invariance, using non-linear regression analysis. 5. The most suitable suitable descriptive model consisted of a saturable hyperbola (Vm = 285 n-mole.(g wet wt.)-1.min-1, Km = 252 micron) and a linear term (b = 0.45 min-1).

Animals↗

PubMed1997

Mental chronometry with simple linear regression.

Typically, mental chronometry is performed by means of introducing an independent variable postulated to affect selectively some stage of a presumed multistage process. However, the effect could be a global one that spreads proportionally over all stages of the process. Currently, there is no method to test this possibility although simple linear regression might serve the purpose. In the present study, the regression approach was tested with tasks (memory scanning and mental rotation) that involved a selective effect and with a task (word superiority effect) that involved a global effect, by the dominant theories. The results indicate (1) the manipulation of the size of a memory set or of angular disparity affects the intercept of the regression function that relates the times for memory scanning with different set sizes or for mental rotation with different angular disparities and (2) the manipulation of context affects the slope of the regression function that relates the times for detecting a target character under word and nonword conditions. These ratify the regression approach as a useful method for doing mental chronometry.

Attention↗

PubMed1996

Appropriate use of the Glasgow Coma Scale in intubated patients: a linear regression prediction of the Glasgow verbal score from the Glasgow eye and motor scores.

UNLABELLED: The Glasgow Coma Scale (GCS) has been shown to be a valuable tool in assessing the neurologic and physiologic status of critically ill patients. Unfortunately, the GCS requires assessment of the verbal response of the patient and this can be blocked by intubation. The purpose of this study was to assess the ability of a regression model based upon the eye and motor components of the GCS to accurately predict the verbal response of the GCS. The primary hypothesis was that the verbal response could be derived from the motor and eye responses of the GCS. METHODS: Data were collected prospectively in an intensive care unit computer data base. Patients were divided into training and test data sets. Linear regression was used to derive a model of verbal score from the motor and eye scores of the GCS in the training data set. Correlation between the actual and the predicted verbal scores was calculated. RESULTS: A total of 2,521 GCS assessments were available for analysis. The second order multiple regression model was an accurate predictor of the verbal score (Pearson's Correlation r = 0.9, R2 = 0.8, p = 0.0001) in 1,463 observations in the training data set. Second Order Multiple Regression Model: Estimated GCS Verbal = (2.3976) + [GCS Motor x (-0.9253)] + [GCS Eye x (-0.9214)] + [(GCS Motor)2 x (0.2208)] + [(GCS Eye)2 x (0.2318)] where r = 0.91, R2 = 0.83, and p = 0.0001. The accuracy of this model was confirmed by comparing the predicted verbal score to the actual verbal score in the test data set (n = 736, r = 0.92, R2 = 0.85, p = 0.0001) CONCLUSIONS: The GCS is a useful tool in the intensive care unit and a critical part of the APACHE II assessment of patient acuity. GCS has been shown to be a useful tool in its own right as a predictor of outcome in the critically ill. Its use is limited with intubation. (See Segatore M, Way C: Heart Lung 21:548, 1992; and Lieh-Lai MW, Theodorou AA, Sarnaik AP, et al: J Pediatr 120:195, 1992.) The present study demonstrates that a relatively simple regression model can use the eye and motor components of the GCS to predict the expected verbal component of the GCS, thus allowing the calculation of the GCS sum score in intubated patients.

Adult↗

PubMed1987

A non-linear regression program in BASIC for estimating Km and Vmax.

This paper gives a program in BASIC for calculating the kinetic parameters Km and Vmax for an enzyme reaction from a set of paired values of reaction velocity at given substrate concentrations. An initial estimate of the two parameters is made using a weighted linear regression, these values are then used in an iterative process to fit the data to the Michaelis-Menten equation and give final values of Km and Vmax with their associated standard errors.

Algorithms↗

PubMedDate not supplied

[Body mass index and its relationship to nutritional and socioeconomic variables: a linear regression approach to a Brazilian adult sub-population].

This paper focuses on the relationship between body mass index (BMI) and family energy intake, occupational energy expenditure, per capita family expenditure, sex, age, and left arm circumference for a group of Brazilian adults randomly selected among those interviewed for a survey on food consumption and family budgets, called the National Family Expenditure Survey. The authors discuss linear regression methodological issues related to treatment of outliers and influential cases, multicollinearity, model specification, heteroscedasticity, as well as the use of two-level variables derived from samples with complex design. The results indicate that the model is not affected by outliers and that there are no significant specification errors. They also show a significant linear relationship between BMI and the variables listed above. Although the hypothesis tests indicate significant heteroscedasticity, its corrections did not significantly change the model's parameters, probably due to the sample size (14,000 adults), making hypothesis tests more rigorous than desired.

Adult↗

PubMed1982

A computer linear regression model to determine ventilatory anaerobic threshold.

The anaerobic threshold has generally been determined by simple visual inspection of ventilation or other gas-exchange data obtained during incremental exercise. To establish objective criteria for the determination of anaerobic threshold, a computer algorithm has been developed that models the ventilatory response to exercise using multisegment linear regression. The best-fit regression model is chosen by minimizing the pooled residual sum of squares . The anaerobic threshold is reported as the first break point in that model. The computer-determined anaerobic threshold values for 37 subjects were compared with subjectively determined values as chosen by four independent observers. The observers' estimates, when pooled to yield a single a single value for each subject, gave a mean value for the gas-exchange anaerobic threshold of 2.26 +/- 0.69 l/min. The estimates by the computer method averaged 2.21 +/- 0.65 l/min. The correlation coefficient for these two methods was 0.94.

Adult↗

PubMed2004

Predicting range of motion after revision total knee arthroplasty: clustering and log-linear regression analyses.

The purpose of this study was to determine which variables affected the range of motion following revision total knee arthroplasty. These variables included preoperative flexion, intraoperative flexion, preoperative alignment, patient demographics, type of posterior soft-tissue release, previous prosthesis type, and prosthesis type used for revision of 355 total knee arthroplasties. Clustering and log-linear regression analyses were used to determine which variables were significantly related to the postoperative flexion. The mean preoperative and postoperative flexion were 100.5 degrees and 104.6 degrees. Low preoperative (<103 degrees) and intraoperative flexion (<117 degrees), young age (<44 years), and constrained and hinged prosthesis types were associated with diminished flexion. Higher preoperative and intraoperative flexion resulted in higher postoperative flexion. When comparing the results of this study to the results of a similar study of primary total knee arthroplasties, flexion improved less following revision than following primary total knee arthroplasty.

Aged↗

PubMed2001

Comparison of antimicrobial in vitro activities against Streptococcus pneumoniae independent of MIC susceptibility breakpoints using MIC frequency distribution curves, scattergrams and linear regression analyses.

Comparing in vitro activities of antimicrobial agents against Streptococcus pneumoniae using per cent susceptible to recommended MIC breakpoints is not optimal. In this study, MICs of penicillin G, ampicillin/sulbactam, ceftriaxone, cefuroxime, erythromycin, tetracycline, trimethoprim/sulfamethoxazole, ciprofloxacin, levofloxacin, trovafloxacin and moxifloxacin were determined for 646 strains. Drug activities were compared using MIC frequency distribution curves, scattergrams and linear regression analyses of MICs (log2). MIC frequency distributions did not always correspond to recommended breakpoints for distinguishing susceptible, intermediate and resistant strains. Penicillin G, ampicillin/sulbactam and ceftriaxone had similar activities and were each c. 1-1.5 dilution steps more active than cefuroxime. For all beta-lactam drug pairs, there was a high correlation of MICs with regression line slopes (a approximately equal to 1) and coefficients of determination (R(2) = 0.90-0.97). Although beta-lactam-resistant strains were more likely to be resistant to erythromycin, tetracycline and/or trimethoprim/sulfamethoxazole than were beta-lactam-susceptible strains, MIC correlations were relatively poor (R(2) = 0.14-0.46), as they were when the non-beta-lactam drugs were compared with each other (R(2) = 0.10-0.25). Trovafloxacin and moxifloxacin were each c. 2.5 dilution steps more active than ciprofloxacin and levofloxacin. There was no correlation of quinolone MICs with MICs of any other drug class (R(2) 0.02). Among the quinolones, however, there was a high correlation of MICs with a approximately equal to 1 and R(2) = 0.81-0.92. With the quinolone drug pairs, lines of best fit were second-order polynomial equations, consistent with a dissociation of low level resistance mechanisms. In summary, beta-lactam and quinolone MICs were predictable within drug classes and testing multiple derivatives within each class is probably not necessary. Although there was some relationship between beta-lactam, erythromycin, tetracycline and trimethoprim/sulfamethoxazole MICs, predictability of MICs between drug classes was poor. There was no relationship between quinolone MICs and MICs of any of the other drugs tested.

4-Quinolones↗

PubMed2006

Neck-focused panic attacks among Cambodian refugees; a logistic and linear regression analysis.

Consecutive Cambodian refugees attending a psychiatric clinic were assessed for the presence and severity of current--i.e., at least one episode in the last month--neck-focused panic. Among the whole sample (N=130), in a logistic regression analysis, the Anxiety Sensitivity Index (ASI; odds ratio=3.70) and the Clinician-Administered PTSD Scale (CAPS; odds ratio=2.61) significantly predicted the presence of current neck panic (NP). Among the neck panic patients (N=60), in the linear regression analysis, NP severity was significantly predicted by NP-associated flashbacks (beta=.42), NP-associated catastrophic cognitions (beta=.22), and CAPS score (beta=.28). Further analysis revealed the effect of the CAPS score to be significantly mediated (Sobel test [Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1182]) by both NP-associated flashbacks and catastrophic cognitions. In the care of traumatized Cambodian refugees, NP severity, as well as NP-associated flashbacks and catastrophic cognitions, should be specifically assessed and treated.

Adult↗

PubMed1988

Measurement error and its impact on partial correlation and multiple linear regression analyses.

In studies examining associations between dietary factors and biomedical risk factors, the relations, if they exist, are frequently attenuated by measurement error. Measurement error may be due to a large intraindividual variation and an inadequate number of measurements or to an inaccurate measuring instrument. This paper evaluates the impact of measurement error on partial correlation and multiple linear regression analyses. Quantitative methods are derived to estimate the potential attenuation of associations. The results indicate that when the controlled variables do not have measurement error, but the correlated variables do, the attenuation of the partial correlation coefficient (or multiple regression coefficient) is greater than that of the simple correlation (or regression) coefficient. When both the correlated variables and the controlled variables have measurement error, the partial correlation (or the regression) coefficients can be either increased or decreased.

Cholesterol, Dietary↗

PubMed2003

Predicting range of motion after total knee arthroplasty. Clustering, log-linear regression, and regression tree analysis.

BACKGROUND: Range of motion is a crucial measure of the outcome of total knee arthroplasty. The purpose of this study was to determine which factors are predictive of the postoperative range of motion. METHODS: We retrospectively studied 3066 patients (4727 knees) who had a primary total knee arthroplasty with the same type of implant at the same center between 1983 and 1998. Statistical clustering analysis paired with log-linear regression was used to determine groupings along continuous variables. Regression tree analysis was used to characterize the combinations of variables influencing the postoperative range of motion. The variables considered were preoperative and intraoperative flexion and extension, preoperative alignment, age, gender, and soft-tissue releases. RESULTS: Preoperative flexion was the strongest predictor of the postoperative flexion regardless of preoperative alignment. Other factors that were significantly related to reduced flexion were intraoperative flexion (p < 0.0001), gender (p < 0.0001), preoperative tibiofemoral alignment (p = 0.0005), age (p < 0.0001), and posterior capsular release (p < 0.0001). The removal of posterior osteophytes was related to the greatest increase in postoperative flexion in the group of patients with a varus tibiofemoral alignment preoperatively. CONCLUSIONS: The principal predictive factor of the postoperative range of motion was the preoperative range of motion. Removal of posterior osteophytes and release of the deep medial collateral ligament, the semimembranosus tendon, and the pes anserinus tendon in patients with large preoperative varus alignment and the attainment of a good intraoperative range of motion improved the likelihood that a good postoperative range of motion would be achieved.

Adolescent↗

PubMed2005

Imputation and variable selection in linear regression models with missing covariates.

Across multiply imputed data sets, variable selection methods such as stepwise regression and other criterion-based strategies that include or exclude particular variables typically result in models with different selected predictors, thus presenting a problem for combining the results from separate complete-data analyses. Here, drawing on a Bayesian framework, we propose two alternative strategies to address the problem of choosing among linear regression models when there are missing covariates. One approach, which we call "impute, then select" (ITS) involves initially performing multiple imputation and then applying Bayesian variable selection to the multiply imputed data sets. A second strategy is to conduct Bayesian variable selection and missing data imputation simultaneously within one Gibbs sampling process, which we call "simultaneously impute and select" (SIAS). The methods are implemented and evaluated using the Bayesian procedure known as stochastic search variable selection for multivariate normal data sets, but both strategies offer general frameworks within which different Bayesian variable selection algorithms could be used for other types of data sets. A study of mental health services utilization among children in foster care programs is used to illustrate the techniques. Simulation studies show that both ITS and SIAS outperform complete-case analysis with stepwise variable selection and that SIAS slightly outperforms ITS.

Algorithms↗