SEARCH · PubMed Health

Results for “Regression Analysis”

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

PubMed2000

Prognostic modelling with logistic regression analysis: a comparison of selection and estimation methods in small data sets.

Logistic regression analysis may well be used to develop a prognostic model for a dichotomous outcome. Especially when limited data are available, it is difficult to determine an appropriate selection of covariables for inclusion in such models. Also, predictions may be improved by applying some sort of shrinkage in the estimation of regression coefficients. In this study we compare the performance of several selection and shrinkage methods in small data sets of patients with acute myocardial infarction, where we aim to predict 30-day mortality. Selection methods included backward stepwise selection with significance levels alpha of 0.01, 0.05, 0. 157 (the AIC criterion) or 0.50, and the use of qualitative external information on the sign of regression coefficients in the model. Estimation methods included standard maximum likelihood, the use of a linear shrinkage factor, penalized maximum likelihood, the Lasso, or quantitative external information on univariable regression coefficients. We found that stepwise selection with a low alpha (for example, 0.05) led to a relatively poor model performance, when evaluated on independent data. Substantially better performance was obtained with full models with a limited number of important predictors, where regression coefficients were reduced with any of the shrinkage methods. Incorporation of external information for selection and estimation improved the stability and quality of the prognostic models. We therefore recommend shrinkage methods in full models including prespecified predictors and incorporation of external information, when prognostic models are constructed in small data sets.

Age Factors↗

PubMed1992

Cytopathology of malignant mesothelioma: a stepwise logistic regression analysis.

Twenty-four cytologic features, previously reported to be useful in the distinction of malignant mesothelioma, adenocarcinoma, and benign mesothelial proliferation in serous effusions were assessed. Forty-four cases of malignant mesotheliomas, 46 cases of metastatic adenocarcinomas, and 30 cases of benign mesothelial proliferations were examined for these parameters. When these cytologic features were subjected to a stepwise logistic regression analysis, five features were selected to distinguish malignant mesothelioma from adenocarcinoma. These were true papillary aggregates, multinucleation with atypia, cell-to-cell apposition, acinus-like structures, and balloon-like vacuolation, the latter two features being characteristic of adenocarcinoma. The four variables selected to distinguish malignant mesothelioma from benign mesothelial proliferations were nuclear pleomorphism, macronucleoli, cell-in-cell engulfment, and monolayer cell groups, the latter being a feature of benign proliferations. Using these selected variables, the logistic model correctly predicted 95.4% of cases of malignant mesothelioma versus 100% of adenocarcinoma and 100% of malignant mesotheliomas versus 90% of benign mesothelial proliferations. The results of regression analysis suggest that many of the previously described cytologic features are not important diagnostic discriminators.

Adenocarcinoma↗

PubMed2003

Outcomes of an intervention to improve hospital antibiotic prescribing: interrupted time series with segmented regression analysis.

OBJECTIVES: To evaluate an intervention to reduce inappropriate use of key antibiotics with interrupted time series analysis. METHODS: The intervention is a policy for appropriate use of Alert Antibiotics (carbapenems, glycopeptides, amphotericin, ciprofloxacin, linezolid, piperacillin-tazobactam and third-generation cephalosporins) implemented through concurrent, patient-specific feedback by clinical pharmacists. Statistical significance and effect size were calculated by segmented regression analysis of interrupted time series of drug use and cost for 2 years before and after the intervention started. RESULTS: Use of Alert Antibiotics increased before the intervention started but decreased steadily for 2 years thereafter. The changes in slope of the time series were 0.27 defined daily doses/100 bed-days per month (95% CI 0.19-0.34) and pound 1908 per month (95% CI pound 1238- pound 2578). The cost of development, dissemination and implementation of the intervention ( pound 20133) was well below the most conservative estimate of the reduction in cost ( pound 133296), which is the lower 95% CI of effect size assuming that cost would not have continued to increase without the intervention. However, if use had continued to increase, the difference between predicted and actual cost of Alert Antibiotics was pound 572448 (95% CI pound 435696- pound 709176) over the 24 months after the intervention started. CONCLUSIONS: Segmented regression analysis of pharmacy stock data is a simple, practical and robust method for measuring the impact of interventions to change prescribing. The Alert Antibiotic Monitoring intervention was associated with significant decreases in total use and cost in the 2 years after the programme was implemented. In our hospital, the value of the data far exceeded the cost of processing and analysis.

Anti-Bacterial Agents↗

PubMed2006

Multiple regression analysis as a tool for the identification of relations between semi-quantitative LC-MS data and cytotoxicity of extracts of the fungus Fusarium avenaceum (syn. F. arthrosporioides).

The cytotoxicity of methanolic extracts from rice cultures of 53 Fusarium avenaceum strains, which had been isolated from different host organisms in Northern Europe, Canada and Australia/New Zealand, was investigated in a rat hepatoma (H4IIE-W), porcine epithelial kidney (PK-15), foetal feline lung fibroblast, dog lymphoblast (D3447), and a human hepatocarcinoma (Hep G2) cell line using the Alamar Bluetrade mark assay. All extracts were screened for known fungal metabolites using high-performance liquid chromatography with photodiode array and mass spectrometric detection, and both known and unknown metabolites were semi-quantified. Known metabolites that were determined in the cultures include acuminatopyrone, 2-amino-14,16-dimethyloctadecan-3-ol (2-AOD-3-ol), antibiotic Y, aurofusarin, chlamydosporol, chlamydospordiol, enniatins, fusarin A and C, and moniliformin. Multiple regression analysis was used in order to relate fungal metabolites to the cytotoxicity of the extracts. Separate linear regression models were constructed for each cell line. Eleven different fungal metabolites were related to the cytotoxicity (P<0.05). Out of these, nine metabolites were siginificantly related to the cytotoxicity in only one of the five models, while two, namely enniatins and 2-AOD-3-ol, were significant contributors in three or four regression models, respectively. This paper describes how multiple regression analysis may be applied for the assignment of bioactivity/toxicity to the constituents of a multi-component mixture.

Animals↗

PubMed1992

PROBIT: weighted probit regression analysis for estimation of biological activity.

PROBIT estimates biologic activity by weighted probit regression analysis and comparison with a known standard. The program is intended to speed up calculation of the concentration of cytokines in large numbers of cell supernatants. Data are input from either sequential or free-form text files created with a spreadsheet or text editor. This allows transfer of data from a beta-counter equipped with a suitable terminal without retyping. Results are displayed and/or printed as relative potency (% of standard) and 50% effective dose (ED50).

Cytokines↗

PubMed1981

[Use of regression analysis to construct mathematical models of the diameters of arteries].

By means of regressive analysis, using the cervicocerebral arteries as an example, the method to prognosticate the diameter in large arteries by measuring easily accessible external body parts was demonstrated. In 10 dogs body weight, diameter and length of the neck, height of the forelegs and diameter of the cervical arteries (in the roentgenograms) were measured. Statistically significant mathematical models on the diameter of the arteries studied were represented. The degree of influence of each of the four factors was stated for various cervical arteries. The data obtained could be explained from biological point of view.

Anatomy↗

PubMed2005

The influence of assay error weight on gentamicin pharmacokinetics using the Bayesian and nonlinear least square regression analysis in appendicitis patients.

The purpose of this study was to determine the influence of weight with gentamicin assay error on the Bayesian and nonlinear least squares regression analysis in 12 Korean appendicitis patients. Gentamicin was administered intravenously over 0.5 h every 8 h. Three specimens were collected at 48 h after the first dose from all patients at the following times, just before regularly scheduled infusion, at 0.5 h and 2 h after the end of 0.5 h infusion. Serum gentamicin levels were analyzed by fluorescence polarization immunoassay technique with TDxFLx. The standard deviation (SD) of the assay over its working range had been determined at the serum gentamicin concentrations of 0, 2, 4, 8, 12, and 16 microg/mL in quadruplicate. The polynominal equation of gentamicin assay error was found to be SD (microg/mL) = 0.0246 - (0.0495C) + (0.00203C2). There were differences in the influence of weight with gentamicin assay error on pharmacokinetic parameters of gentamicin using the nonlinear least squares regression analysis but there were no differences on the Bayesian analysis. This polynominal equation can be used to improve the precision of fitting of pharmacokinetic models to optimize the process of model simulation both for population and for individualized pharmacokinetic models. The result would be improved dosage regimens and better, safer care of patients receiving gentamicin.

Anti-Bacterial Agents↗

PubMed2005

The influence of weight with assay error on gentamicin pharmacokinetics using the Bayesian and nonlinear least square regression analysis in appendicitis patients.

The purpose of this study was to determine the influence of weight with gentamicin assay error on the Bayesian and nonlinear least squares regression analysis in 12 Korean appendicitis patients. Gentamicin was administered intravenously over 0.5 h every 8 h. Three specimens were collected 48 h after the first dose from all patients at the following times, just before the regularly scheduled infusion, at 0.5 h and 2 h after the end of the 0.5 h infusion. Serum gentamicin levels were analysed by fluorescence polarization immunoassay technique with TDxFLx. The standard deviation (SD) of the assay over its working range had been determined at the serum gentamicin concentrations of 0, 2, 4, 8, 12 and 16 microg/ml in quadruplicate. The polynominal equation of gentamicin assay error was found to be SD (microg/ml) = 0.0246-(0.0495C) + (0.00203C(2)). There were differences in the influence of weight with gentamicin assay error on pharmacokinetic parameters of gentamicin using the nonlinear least squares regression analysis but there were no differences on the Bayesian analysis. This polynominal equation can be used to improve the precision of fitting of pharmacokinetic models to optimize the process of model simulation both for population and for individualized pharmacokinetic models. The result would be improved dosage regimens and the better, safer care of patients receiving gentamicin.

Anti-Bacterial Agents↗

PubMedDate not supplied

Use of regression analysis and flow cytometry for determining levels of mixed semiallogeneic immune chimerism.

It has been shown that tolerance or specific immunologic nonresponsiveness in various lymphohemopoietic transplant models can be associated with the development of mixed lymphoid chimerism. As a specific example, composite tissue (limb) allografts were studied as a model for vascularized bone marrow transplantation (VBMT) and it was demonstrated that development of stable cellular immune chimerism is associated with long-term allograft survival. Recently, studies were initiated using a new parental to hybrid VBMT model, but the detection of donor cells is complicated, due to the fact that they share one parental allotypic determinant. Therefore, regression analysis with a flow cytometric immunofluorescent staining assay was evaluated for the assessment of cellular lymphoid chimerism in donor parental to hybrid (P-->F1) lymphohemopoietic transplant models. Standard curves consisting of known mixed populations of parental donor (Lewis, LEW) and hybrid host F1 (Lew x BN, LBN) lymphocytes were established. Standard curves were analyzed by linear regression statistics and excellent coefficients of determination (r > .881) were obtained for all standard curves. A highly statistically significant (p < .016) linear relationship between level of donor cell chimerism (independent variable) and percent stained (dependent variable) was determined. The technique was then evaluated using the parental to hybrid VBMT model. Levels of donor LEW lymphoid chimerism in all VBMT LBN recipients were successfully assessed by regression analysis and inverse prediction using distinct recipient allodeterminant markers. In conclusion, this technique was proven to be reliable and accurate for the detection of of chimerism in parental to F1 lymphohemopoietic allograft models.

Animals↗

PubMed1986

Regression analysis of aggregate accident data: some methodological considerations and practical experiences.

In empirical traffic safety studies regression analysis is perhaps the most widely used statistical technique. In the great majority of cases the regression model is applied in its standard form. If, however, highly aggregate accident data are analyzed, usually most of the assumptions of the standard model are violated. Referring to a practical study it is demonstrated how it may become difficult to decide whether or not a certain variable can be regarded as truly "explanatory." It appears that even in seemingly unproblematic situations contradictory results are obtained from only slightly different model specifications. It is recommended that the use of highly aggregate accident data be avoided whenever possible. Some hints are given concerning the use of appropriate extensions of the classical linear regression model in road accident studies.

Accidents, Traffic↗

PubMed1999

Multiple regression analysis of interference effects from a hemoglobin-based oxygen carrier solution.

The use of hemoglobin-based oxygen carrier solutions in patients requiring blood transfusion will necessitate that clinical laboratories have mechanisms in place to evaluate the potential interference effect of these substances on testing methods. Because these oxygen carrier solutions contain acellular hemoglobin, but do not contain many of the intracellular enzymes and ions present in erythrocytes, interference effects from blood substitutes may be quite different when compared to in vivo or in vitro lysis of erythrocytes. We evaluated the potential interference effect of Diaspirin Cross-linked Hemoglobin on 29 different clinical laboratory analytes. Various combinations of these analytes were tested using the Hitachi 747 and 911 systems, a Beckman CX3, an Abbott AxSym, a Bayer Immuno I, and a Dade ACA IV; a total of 60 analyte/instrument combinations. We used the method of multiple regression analysis to classify interferences as analyte-dependent, analyte-independent, or a combination of the first two types. The presence of clinically significant test interference was derived by using the criteria for maximum allowable error specified in the Clinical Laboratory Improvement Amendments of 1988. Using these criteria, we found significant interference from Diaspirin Cross-linked Hemoglobin with 13 of 29 analytes tested. Interference was noted with the Hitachi 747 and 911 methods for albumin, alkaline phosphatase, total and conjugated bilirubin, cholesterol, total carbon dioxide, iron, lactate dehydrogenase, magnesium, total protein, and triglyceride. In addition, Diaspirin Cross-linked Hemoglobin interfered with measurement of L-lactate using the ACA IV and minor interference was noted with glucose measured using the Beckman CX3. Data from the interference studies was graphically displayed in the form of interference plots. These plots show the maximum allowable test error, due to Diaspirin Cross-linked Hemoglobin, as a function of analyte and interferent concentrations. Evaluation of the potential interference effect of hemoglobin-based oxygen carrier solutions with use of multiple regression analysis and graphical display of the resultant data in the form of interference plots allows for more reliable reporting of test results from specimens containing these products.

Aspirin↗

PubMed1985

A multiple regression analysis of hemoglobin values and iron status in Japanese farmers.

The authors performed multiple regression analysis of hemoglobin values and the parameters related to iron dynamics. The subjects consisted of 37 farmers whose Hb, Ht, GB, SI, %-St, SFR and FEP was examined once a month for one year. 19.2% of the female farmers had anemia of which 73.8% was iron-deficiency anemia, 13.7% being latent iron deficiency. Hb, GB and FEP in males and SI, %-St, SFR, FEP in females revealed clear seasonal variations. Multiple correlation coefficient between Hb and SI, %-St, SFR, FEP as independent variables was significant in females but not in males, and the ratio of contribution (38.7%) was not very high, however, the multiple correlation coefficient was significantly high (0.622). Standard regression coefficients between Hb and SFR, FEP were significantly high (p less than 0.01). When Ht and GB were added to these 4 variables, multiple correlation coefficients (male, 0.906; female, 0.957) were remarkably high, which means serum-related variables such as Ht and GB have an undeniably important role as explanatory variables of hemoglobin levels. These standard regression coefficients showed seasonal changes.

Adult↗

PubMed1984

Multiple-dose nonlinear regression analysis program for the microcomputer.

An adaptation of a previously published program for nonlinear regression analysis of serum drug concentrations which can utilize data obtained during multiple-dose administration is described. Specific programs have been developed for aminoglycosides, digoxin, and theophylline. The programs provide initial parameter estimates (for a one-compartment linear model) for each drug based on mean population parameters and refined estimates based on observed concentration data. Clinical examples which demonstrate the flexibility of these programs are provided. The programs may also be employed for fitting data for any drug that can be adequately described by a one-compartment linear model.

Adult↗

PubMedDate not supplied

A multiple regression analysis for predicting local control of esophageal carcinoma treated by intracavitary irradiation.

A multiple regression analysis for predicting local control of esophageal carcinoma treated by intracavitary irradiation is reported. In phase I, the value of predicting local control (VPLC) was determined by five parameters: VPLC = 1.38950 - 0.01571 X (age) + 0.04517 X (tumor length) + 0.62167 X (stenosis) + 0.94811 X (deep ulcer) - 0.02969 X (Total dose of intracavitary irradiation). From correlations between VPLC and the local control observed in 19 phase I patients, it was determined that a VPLC value of 0.5 or more predicted local failure and a VPLC of less than 0.5 signified successful local control. In phase II, prediction of local failure or local control was made for another 24 patients by calculating VPLC, and 22 of the 24 cases (91.7%) were correctly predicted. These results indicate that VPLC is a useful index for predicting of local control after intracavitary irradiation in cases with esophageal carcinoma.

Brachytherapy↗

PubMed1985

Clinical factors predisposing to major ulcer bleeding. A logistic regression analysis.

In a consecutive series of patients with bleeding from peptic ulcer, a logistic regression analysis was used to evaluate clinical factors as predictors of major hemorrhage. A hemodynamic assessment distinguished between self-limiting hemorrhage requiring no treatment (190 patients) and major bleeding necessitating emergency surgery (86 patients). Most data, including patient age and sex, previous bleeding, and pulse rate, blood pressure and hematocrit on admission, were poor predictors of major ulcer bleeding. A few variables--onset of hemorrhage, appearance of hematemesis and site of ulcer--were statistically of predictive importance, but clinically only in combination. When one of the selected predictor variables was present, the risk of major hemorrhage did not statistically exceed the prior probability. When combined in pairs, the probability increased, primarily when duodenal ulcer or red hematemesis was present. When the patients presented with three of the selected predictor variables, the probability of major ulcer hemorrhage was markedly increased. However, only 8% of the patients presented with three of the selected variables. Clinical factors, therefore, contribute in general little in prediction of major hemorrhage form peptic ulcer.

Adolescent↗

PubMed1989

[Growth hormone response to arginine administration in diabetics--with special reference to the multiple regression analysis in association with diabetic retinopathy].

In 107 patients with non-insulin dependent diabetes(NIDDM), plasma growth hormone(GH) responses during standard arginine test (0.5 g/kg of body weight) were studied and analyzed in comparison with those in 17 normal subjects. The indices of the responsiveness of GH, peak value of GH, sum of GH values(sigma GH), area of GH curve(integral of GH), sum of GH values above fasting level(sigma delta GH) and area of GH curve above fasting level(integral of delta GH) during the test (2 hr) were calculated. Data were also analyzed with multiple regression analysis using stepwise method for variable selection. Basal level of GH was significantly higher in diabetic patients than in normal subjects (2.1 +/- 1.7 vs. 1.6 +/- 0.5 ng/ml, mean +/- SD, p less than 0.05), and sigma GH and integral of GH were also higher in diabetic patients. There was a significantly positive correlation between fasting plasma glucose(FPG) and basal level of GH (r = 0.24, n = 107, p less than 0.05), and the indices of GH responses except delta GH and GH peak value (r = 0.24 to 0.31, p less than 0.05 to 0.01). Some indices of GH responses (sigma delta GH, sigma GH, integral of delta GH and integral of GH) were significantly higher in the poor control group (patients with FPG above 180 mg/dl, n = 29) of diabetic patients than in the good control group (patients with FPG below 140 mg/dl, n = 59), or in the group with no abnormal findings of retinopathy (n = 46). During the follow-up of retinopathy for 2.5 years on the average, progression of retinopathy was found in 21 out of 107 patients. Significantly higher GH, and GH in the patients with increasing severity of retinopathy were revealed retrospectively compared to the patients without it. However, there were no significant differences in these parameters between both groups matched by FPG or severity of retinopathy. Multiple regression analysis to the basal GH level and GH responses during arginine infusion as criterion variables of various predictor variables (total 44 factors: biochemical laboratory data, indices of glucose and insulin response to oral glucose load, indices of glucose response to arginine, age, age of the onset, obese index, duration of retinopathy, neuropathy, and therapy) were performed in 86 patients using forward and backward method for variable selection. Basal plasma level of GH showed close positive association with therapy and proteinuria and negative association with age and obesity. Five of 6 indices of GH responsiveness showed significant relationship with retinopathy.(ABSTRACT TRUNCATED AT 400 WORDS)

Adult↗

PubMed1995

An investigation of the relationship between antioxidant vitamin intake and coronary heart disease in men and women using logistic regression analysis.

Antioxidant vitamin intake (C,E and carotene) is assessed from a food frequency questionnaire applied to 10,359 middle-aged men and women participating in the Scottish Heart Health Study. Logistic regression analysis is then used to quantify the relationship between antioxidant vitamin consumption and prevalent coronary heart disease (CHD), analysing diagnosed and undiagnosed cases separately. For men, there is a protective effect of all three antioxidants, before and after adjustment for a comprehensive set of confounding variables. For women the picture is less clear, only vitamin C is negatively associated with CHD, but the effect is removed by adjustment. The logistic regression model is also used to determine classification rules for deciding whether or not an individual has CHD. The classification error rates using the antioxidants are found to be very similar to those found using smoking, blood pressure and serum total cholesterol as classification variables. Significant interactions are found for the antioxidants with smoking, cholesterol and age. It is concluded that antioxidant vitamin intake protects against CHD for men. Logistic regression analysis is compared with discriminant analysis, and is found to have important advantages as an epidemiological tool.

Adult↗

PubMed2006

Nonlinear regression analysis of kinetics of the photocatalytic decolorization of an azo dye in aqueous TiO2 slurry.

The kinetics of decolorization of an anionic monoazo dye of acid class named C.I. Acid Red 27 (AR27) was investigated in the UV/TiO2 process with nonlinear regression analysis. The experimental results indicated that the kinetics of decolorization of AR27 in this process fit well by pseudo-first order kinetics. With nonlinear regression analysis a model was developed for pseudo-first order rate constant (k(ap,UV/TiO2)) as a function of operational parameters such as TiO2 dosage, initial concentration of AR27, concentration of O2 and UV-light intensity (I0) as following: k(ap,UV/TiO2) = 0.0025 [TiO2](0.65) [AR27]0(-0.96) [O2](0.16)I0. This rate expression can be used for predicting k(ap,UV/TiO2) at different conditions.

Journal Article↗