Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Multivariate statistics”

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 721 records · Page 40Linked to original sources
PubMed2000

Construction of an algorithm for quick detection of patients with low bone mineral density and its applicability in daily general practice.

OBJECTIVE: To construct a quick algorithm to detect patients with low bone mineral density (BMD) and osteoporosis and determine its applicability in daily general practice. DESIGN: Cross-sectional study in all 9107 postmenopausal women, aged 50-80, registered at 12 general practice centers. SUBJECTS AND MEASUREMENTS: All healthy women (5303) and 25% of the remaining group (943/3804) were invited to participate. Of 6246 invited women, 4725 (76%) participated. The women were questioned (state of health, medical history, family history, and food questionnaire) and examined [weight, height, body mass index (BMI), and BMD of the lumbar spine]. STATISTICS: Multivariable, stepwise backward and forward logistic regression analyses were performed, with BMD of the lumbar spine (L2-L4, cut-off points at 0.800 g/cm(2) for osteoporosis and 0.970 g/cm(2) for low BMD) as the dependent variable. An algorithm was constructed with those variables that correlated statistically significantly and clinically relevant with the presence of both osteoporosis and low BMD. RESULTS: The prevalence of osteoporosis was 23%, that of low BMD was 65%. Only three variables (age, BMI, and fractures) were statistically significant and clinically relevant correlated with the presence of both osteoporosis and low BMD. Age (OR 2.70 for osteoporosis and OR 1.77 for low BMD) and fractures during the past five years (OR 3.60 for osteoporosis and OR 2.85 for low BMD) were found to be the key predictors. From the algorithm the absolute risks varied from 9% to 51% for osteoporosis and from 48% to 84% for low BMD. The corresponding relative risks varied from 1.0 to 5.7 and from 1.0 to 1.8. CONCLUSIONS: Using an algorithm with age, BMI, and fracture history subgroups at high risk could be identified. However, in whatever combination, many women with osteoporosis could not be identified. Despite the differences in methods, we found predictors for osteoporosis which were comparable with the results of other cross-sectional studies, meaning that the first selection of patients at high risk for low BMD can be done adequately by both specialists and general practitioners.

Aged↗
PubMed2004

A multivariate analysis of the outcome of endodontic treatment.

In the present study, multivariate analyses were performed on clinical and treatment variables that may influence the outcome of endodontic treatment. Data collected in a previous clinical-radiographic follow-up study were used. Of 810 treated, 675 roots in 498 teeth were followed for 6 months to 4 yr. Of these, 192 (the CAP group) had pre-existing, chronic apical periodontitis and 483 (the NAP group) had not. Root canal treatment followed a standard procedure with one of three sealers chosen at random. Demographic, clinical and radiographic variables were recorded at the start of, and during treatment. The periapical index (PAI) score was used to record the outcome of treatment, and applied in two different endpoint modes (END1 and END2) as the dependent variable for multivariate statistical analyses using logistic regression and the general model. The modes reflected increasing PAI scores (END1) and conventional success/failure assessment (END2). Dropouts were largely similar to the cases followed up. A total of 10 preoperative and peroperative variables were found to be significantly associated with treatment outcome by the multivariate analyses of either the total material or the NAP or CAP subgroups. Several of these were not significant in univariate analyses (e.g. the effect of sealer). Conventional success/failure analyses (END2) identified fewer of the influential variables and had low explanatory power, whereas PAI scores on an ordinal scale (END1) were most sensitive in identifying variables of influence on the treatment outcome.

Adolescent↗
PubMed2002

A Bayesian method for classification of images from electron micrographs.

Particle classification is an important component of multivariate statistical analysis methods that has been used extensively to extract information from electron micrographs of single particles. Here we describe a new Bayesian Gibbs sampling algorithm for the classification of such images. This algorithm, which is applied after dimension reduction by correspondence analysis or by principal components analysis, dynamically learns the parameters of the multivariate Gaussian distributions that characterize each class. These distributions describe tilted ellipsoidal clusters that adaptively adjust shape to capture differences in the variances of factors and the correlations of factors within classes. A novel Bayesian procedure to objectively select factors for inclusion in the classification models is a component of this procedure. A comparison of this algorithm with hierarchical ascendant classification of simulated data sets shows improved classification over a broad range of signal-to-noise ratios.

Algorithms↗
PubMed1990

A critical review of clinical applications of topographic mapping of brain potentials.

A brief review of several aspects of topographic mapping of brain potentials is presented. Some general factors of importance for the construction of brain maps are considered, namely, how to sample in space and how to deal with the influence of the reference electrode. The issue of what is a representative map is examined for the cases of both steady and dynamic states. Special attention is paid to the methodology of assessing whether a brain map may be considered "normal" or "abnormal" on the basis of multivariate statistical considerations. Besides being used to visualize statistical comparisons between sets of EEG data, brain maps may be used as a first step for a model-based analysis in order to estimate the localization of equivalent dipolar sources of the electrical activity within the brain. The question of whether brain maps may yield added value to multivariate quantitative EEG analysis is examined for the following clinical cases: (1) developmental disorders; (2) space-occupying lesions; (3) epilepsy; (4) cerebrovascular diseases; (5) dementia; and (6) psychiatric diseases. It is concluded that brain mapping may be valuable as a comprehensive and attractive form of data reduction to represent a complex set of multichannel (optimally at least 64) EEG signals, but only if the user is well aware of the pitfalls involved. Brain mapping might supplement quantitative topographic analysis if it can be used as the basis for solving the inverse problem in EEG. More extensive and well-controlled studies are necessary in order to establish the clinical value of complementing a routine EEG examination with brain mapping.

Brain↗
PubMed1992

Prognosis after therapy discontinuation in children with epilepsy.

Anticonvulsant therapy was stopped in 191 epileptic children (109 males and 82 females) after a seizure-free period of at least 2 years, independent of EEG findings, followed up for a minimum of 2 years after withdrawal. Overall, 43 (22.5%) had recurrence of seizures; of these, 79% relapsed within the 1st year after drug stopping and 93% within the 2nd year. The probability of remaining seizure free was 97% at the end of the withdrawal period, 82% 1 year later, 79% at 2 years and 77% at 5 years. In order to evaluate the risk of recurrence and the predictive factors of relapse, several parameters were investigated by univariate and multivariate statistical analysis. At univariate analysis, the factors which proved to be significantly correlated to relapses were: age at onset over 4 years, seizure-free time less than 2 years, sudden drug discontinuation, pathological EEG records during seizure-free time and paroxysmal responses to intermittent photic stimulation (IPS). At multivariate analysis, only age at onset, seizure-free time and sudden discontinuation were the factors indicating a significantly higher relapse risk; paroxysmal IPS responses, when analyzed in association with these variables, proved to be significant and increased the predictive value for prognosis of associated factors.

Adolescent↗
PubMed1998

Do QEEG topographic map data have a multivariate normal distribution? Implications for tests of individual maps.

The study was designed to test whether topographic map data expressed in log power units have a multivariate normal (MVN) distribution in a healthy population, and to determine if any deviation from normality poses serious difficulties for the use of standard multivariate statistical tests in assessing the significance of deformations of the map in individual cases. Data on 361 healthy adults using 20 electrodes in the classic 10/20 configuration were recorded in six frequency bands. The log-transformed power data were shown to deviate markedly from MVN. The actual distributions of multivariate tests were computed for the sample using the 'jackknife' method, and shown to deviate markedly from the F-distributions that would be expected for MVN data. These 'jackknife' sampling distributions were then used to demonstrate significant deformations in the topographic maps of a patient who had sustained traumatic head injury.

Adolescent↗
PubMed2002

Poisson and multinomial mixture models for multivariate SIMS image segmentation.

Multivariate statistical methods have been advocated for analysis of spectral images, such as those obtained with imaging time-of-flight secondary ion mass spectrometry (TOF-SIMS). TOF-SIMS images using total secondary ion counts or secondary ion counts at individual masses often fail to reveal all salient chemical patterns on the surface. Multivariate methods simultaneously analyze peak intensities at all masses. We propose multivariate methods based on Poisson and multinomial mixture models to segment SIMS images into chemically homogeneous regions. The Poisson mixture model is derived from the assumption that secondary ion counts at any mass in a chemically homogeneous region vary according to the Poisson distribution. The multinomial model is derived as a standardized Poisson mixture model, which is analogous to standardizing the data by dividing by total secondary ion counts. The methods are adapted for contextual image segmentation, allowing for spatial correlation of neighboring pixels. The methods are applied to 52 mass units of a SIMS image with known chemical components. The spectral profile and relative prevalence for each chemical phase are obtained from estimates of model parameters.

Models, Statistical↗
PubMed1987

Angular reconstitution: a posteriori assignment of projection directions for 3D reconstruction.

In computerized tomography as well as in most problems of three-dimensional reconstruction from projections, one knows from the experimental set-up the angular relationships between the projections from which the reconstruction is to be calculated. A serious difficulty is encountered when the angles are not known. In this paper, a method of "angular reconstitution" is described, which allows the a posteriori determination of the relative angular orientations of the projections and thus enables the three-dimensional reconstruction of the object to be calculated. For asymmetric objects, a minimum of three projections is required, which should not be related by a tilt around a single rotation axis. The method can be applied to determine the three-dimensional structure of biological macromolecules based on electron micrographs of randomly oriented individual molecules. Angular reconstitution, in combination with multivariate statistical techniques to classify and average the characteristic views of a molecule forms a complete, self-contained methodology for molecular structure analysis by electron microscopy.

Macromolecular Substances↗
PubMed1995

Multiple expression patterns of biopathological markers in primary invasive breast carcinoma: a useful tool for elucidating its biological behaviour.

BACKGROUND: Commonly used clinical and morphologic criteria have been reported to be of limited value in predicting the outcome of malignant tumours of the breast. Integrated information from the quantitative analysis in tumour tissue of biological parameters such as oestrogen and progesterone receptors (ER and PGR), proliferative activity, and proto-oncogene p53, c-erB2, and bcl-2 expression, may be useful for defining the biology of growth of breast carcinoma and to plan effective therapeutic strategies. PATIENTS AND METHODS: Immunohistochemistry with antibodies recognizing ER, PGR, Ki-67, and the p53, c-erbB2, and bcl-2 encoded proteins was performed on 291 primary breast carcinomas. Results were integrated with clinico-pathological indicators and examined with multivariate statistical procedures and modeling. RESULTS: P53, c-erbB2, and bcl-2 gene products were detected, respectively, in 30.6%, 31.6%, and 85.9% of the examined invasive breast carcinomas, revealing variable associations with cellular differentiation and proliferation as defined by ER/PGR status, Ki-67, tumour mass and histologic and nuclear grading. A multivariate graphical display on a subset of the most informative cases revealed that bcl-2 expression parallels ER/PGR status and is of importance in separating tumour clusters with different degrees of aggressiveness. CONCLUSIONS: The results of this study indicate that multivariate explorative analyses conducted on biological and clinico-pathological parameters might constitute an integrated approach to data analysis useful for distinguishing different biological behaviours and therapeutic groups in breast carcinoma. Our findings also suggest that bcl-2 expression may play a pivotal role in tumours lacking ER-mediated growth regulation.

Adult↗
PubMed1996

Prognostic factors for survival in patients with transitional cell carcinoma of the bladder: evaluation by histopathologic grade, pathologic stage and flow-cytometric analysis.

OBJECTIVE AND METHODS: Univariate and multivariate statistical analyses of prognostic factors were performed on 93 patients with transitional cell carcinoma of the bladder who had undergone cystectomies between 1978 and 1988. All patients underwent DNA flow cytometry by way of paraffin-embedded specimens. In univariate analysis, ploidy status, proliferation index, T stage, lymph node status and histopathologic grade were significant prognostic factors. However, in multivariate analysis, only ploidy status (p = 0.0000) and T stage (p = 0.0254) were significant as prognostic factors. With these two variables, we stratified the patients into three distinct groups with varying degrees of prognosis. RESULTS: The mean survival time was 91.3 months for group 1 (diploid T3a or less), 71.3 months for group 2 (diploid T3b, T4/aneuploid T3a or less) and 25.7 months for group 3 (aneuploid T3b, T4). Group 1 patients had a significantly longer survival rate than did the group 2 patients. The difference in survival between groups 2 and 3 was also statistically significant. CONCLUSION: This study shows that DNA ploidy together with T stage may be a useful prognostic factor when identifying risk groups and planning management of patients with transitional cell carcinoma of the bladder.

Adult↗
PubMed1999

A multivariate analysis of lower urinary tract ageing and urinary symptoms: the role of fibrosis.

An investigative trial including 72 patients who underwent open surgery for benign prostatic hypertrophy (BPH) induced urinary symptoms was carried out with the purpose to obtain a deeper insight in the pathophysiology of this clinical picture. Prostate weight, stroma to parenchyma ratio, bladder wall fibrosis, I-PSS score, residual urine and uroflow obtained from these patients were processed by statistical multivariate analysis. The results point out the pivotal impact of prostate and bladder wall fibrosis in conditioning biological and chronological ageing of the lower urinary tract and relative symptoms.

Age Factors↗
PubMed2000

Pulmonary complications of liver transplantation: radiological appearance and statistical evaluation of risk factors in 300 cases.

The aim of this study was to evaluate the incidence, radiographic appearance, time of onset, outcome and risk factors of non-infectious and infectious pulmonary complications following liver transplantation. Chest X-ray features of 300 consecutive patients who had undergone 333 liver transplants over an 11-year period were analysed: the type of pulmonary complication, the infecting pathogens and the mean time of their occurrence are described. The main risk factors for lung infections were quantified through univariate and multivariate statistical analysis. Non-infectious pulmonary abnormalities (atelectasis and/or pleural effusion: 86.7%) and pulmonary oedema (44.7%) appeared during the first postoperative week. Infectious pneumonia was observed in 13.7%, with a mortality of 36.6%. Bacterial and viral pneumonia made up the bulk of infections (63.4 and 29.3%, respectively) followed by fungal infiltrates (24.4 %). A fairly good correlation between radiological chest X-ray pattern, time of onset and the cultured microorganisms has been observed in all cases. In multivariate analysis, persistent non-infectious abnormalities and pulmonary oedema were identified as the major independent predictors of posttransplant pneumonia, followed by prolonged assisted mechanical ventilation and traditional caval anastomosis. A "pneumonia-risk score" was calculated: low-risk score ( < 2.25) predicts 2.7% of probability of the onset of infections compared with 28.7% of high-risk (> 3.30) population. The "pneumonia-risk score" identifies a specific group of patients in whom closer radiographic monitoring is recommended. In addition, a highly significant correlation (p < 0.001) was observed between pneumonia-risk score and the expected survival, thus confirming pulmonary infections as a major cause of death in OLT recipients.

Female↗
PubMedDate not supplied

The use of conscious sedation for pain control during forearm fracture reduction in children: does race matter?

Recent reports in the medical literature indicate that certain racial disparities have been identified in healthcare. The authors sought to identify the potential relationship between the use of pain medications in African-American and Caucasian children undergoing forearm fracture reduction. This retrospective cohort study was performed at a university-affiliated tertiary care children's hospital emergency department. All Caucasian and African-American patients who underwent a closed reduction of a fractured ulna or radius over the 2-year observational period were enrolled. Patients were excluded from the study if they were admitted to the hospital for an open reduction or had multiple injuries. The relationship between race, gender, insurance status, time of admission, length of stay in the emergency department, fracture characteristics, and the use of conscious sedation was analyzed. t tests, chi-square tests, and stepwise logistic regression were used for data analysis. A total of 503 patients were included, 83% Caucasian and 17% African-American. Four hundred four patients received conscious sedation as part of their fracture reduction procedure and 99 did not. Univariate analysis showed that African-American and Caucasian children had different forearm fracture patterns (P = 0.0116) and different severities of angulation (P = 0.0094). Multivariate statistical analysis revealed that higher amounts of fracture translation (P < 0.0001) and angulation (P < 0.0027) and younger age of the patient (P = 0.0059) were significant predictors of conscious sedation use. Race was not found to be significantly associated with the use of conscious sedation (P = 0.0606 in univariate analysis, P = 0.1678 in multivariate analysis). The authors found that the decision to use conscious sedation for pediatric forearm fractures was not influenced by race, but was influenced by certain fracture characteristics and patient age.

Adolescent↗
PubMed2006

Coagulopathy does not fully protect hospitalized cirrhosis patients from peripheral venous thromboembolism.

OBJECTIVE: Despite the endogenous coagulopathy of cirrhosis, some patients with cirrhosis experience thrombophilic states. This study aims to determine the incidence and predictors of venous thromboembolism (VTE), such as deep vein thrombosis (DVT) and pulmonary embolism, in hospitalized patients with cirrhosis. METHODS: A retrospective case-control study was performed in a tertiary-care teaching hospital over an 8-yr period. A total of 113 hospitalized patients with cirrhosis with a documented new VTE were compared to controls. Risk factors for VTE were determined using univariate and multivariate statistical analyses. RESULTS: Approximately 0.5% of all hospitalized patients with cirrhosis had a VTE. Traditional markers of coagulation such as INR and platelet count were not predictive of VTE. In the univariate analysis, serum albumin level was significantly lower in cases than controls (2.85 vs. 3.10 g/dL, respectively, p = 0.01). In the multivariate analysis, serum albumin remained independently predictive of VTE, with an odds ratio of 0.25 (95% CI 0.10-0.56). CONCLUSIONS: Approximately 0.5% of admissions involving cirrhosis patients resulted in a new thromboembolic event. Low serum albumin was strongly predictive of increased risk for developing VTE, independent of international normalized ratio or platelet count. Serum albumin deficiency may indicate low levels of endogenous anticoagulants.

Blood Coagulation Disorders↗
PubMed1998

The prognostic significance of the biomarkers p21WAF1/CIP1, p53, and bcl-2 in laryngeal squamous cell carcinoma.

BACKGROUND: The clinical course of laryngeal squamous cell carcinoma (LSCC) varies considerably among patients. New biologic markers are needed to facilitate the stratification of individual patients within the conventional clinicopathologic stages of LSCC. METHODS: Eighty-three LSCCs from an equal number of patients who received at least 10 years of follow-up were investigated for p53, p21WAF1/CIP1, and bcl-2 protein expression by immunohistochemical techniques. The results were correlated with various clinicopathologic parameters, DNA content, and patient outcome by univariate and multivariate statistical analyses. RESULTS: Stage IV disease, large tumor size (>3 cm), positive lymph node status, extranodal extension, and p53 overexpression (in > 75% of cells) correlated significantly with prognosis in univariate analysis. There was no correlation between patient outcome and age, gender, race, histologic differentiation, or expression of bcl-2 or p21WAF1/CIP1. In multivariate analysis, lymph node status and p53 overexpression were the only factors significantly associated with survival. CONCLUSIONS: High p53 expression and positive lymph node status were independent predictors of the outcomes of patients with LSCC. These factors may assist in prognostication and better classification of patients for treatment.

Adult↗
PubMed1996

Predictors of smoking among male junior secondary school students in Riyadh, Saudi Arabia.

OBJECTIVE: To determine the prevalence and determinants of cigarette smoking among intermediate (junior secondary) schoolboys in Riyadh, Saudi Arabia. PARTICIPANTS: A sample of 1382 students (ages 12-19 years) in 45 classes randomly selected from 15 schools, using a two-stage stratified cluster sampling scheme. DESIGN: Students in the selected classes were requested to complete an anonymous questionnaire, under the supervision of trained interviewers. Univariate and multivariate statistical analyses of potential risk factors were performed. SETTING: Intermediate schools in Riyadh, Saudi Arabia. MAIN OUTCOME MEASURES: Association between current smoking and socio-demographic variables, history of smoking, age of smoking initiation, smoking behaviour among family members, knowledge of the harmful effects of smoking, and whether smoking is allowed in the presence of relatives and acquaintances. RESULTS: The prevalence of current smokers was 13.2% overall, ranging from 3.2% in those 12-13 years old to 31.1% in those aged 18-19. Some of the variables (nationality, father's education, and smoking allowed in the presence of parents or teachers) found to be associated with current smoking in a univariate analysis were no longer significantly associated with smoking in the multivariate analysis. By multivariate analysis, knowledge of the harmful effects of smoking, age, smoking allowed in the presence of friends or brothers, and previous smoking were statistically significant determinants of current smoking. CONCLUSIONS: Current health education activities against smoking should be continued and extended to the young population to further reduce the prevalence of smoking and its health consequences. Religious antipathy toward smoking should be emphasised in any local anti-smoking campaigns.

Adolescent↗
PubMed1998

Psychiatric epidemiology.

Statisticians working in psychiatric epidemiology regularly confront a variety of problems that are rare in other branches of epidemiology. These problems range from the frequent absence of 'objective' biological markers and consequent reliance on informant and self-reports of considerable fallability, to the perhaps corresponding frequent use of multivariate statistical methods such as latent variable modelling. Equally, such difficulties can be regarded as challenges that offer the statistician a critical role.

Adult↗
PubMedDate not supplied

Left ventricular diastolic dysfunction secondary to hyperglycemia in patients with type II diabetes.

Diabetes mellitus type II, a cause of preclinical left ventricular dysfunction that can progress to cardiac insufficiency ventricular dysfunction in diabetic patients, is attributed to systemic arterial hypertension, or ischemic cardiopathy. Diastolic ventricular dysfunction takes place during the course of diabetes mellitus. The purpose of the present article is to report on the influence of hyperglycemia on the left ventricular diastolic dysfunction independently of dyslipidemia, obesity, and systemic arterial hypertension, usually present in diabetic patients. Left ventricular diastolic function was studied by Doppler echocardiography in asymptomatic type II diabetic patients without ischemic or valvular cardiopathies, cardiomegaly, or systemic arterial hypertension. Two groups of patients were integrated: patients with and without left ventricular diastolic dysfunction, i.e., groups A and B, respectively. Glycemia, cholesterol, triglycerides, and body mass index (BMI) were determined in each subject. Bivariate statistical tests (Student t, chi-square, or Mann-Whitney U tests) were applied to study the influence of the previously mentioned variables on the ventricular diastolic function. To evaluate the influence of hyperglycemia on ventricular diastolic function separately from dyslipidemia, systemic arterial hypertension, and the influence of obesity, logistic regression, and multivariate statistical analysis were applied. Independently of dyslipidemia and obesity, a relationship was found between hyperglycemia and diastolic dysfunction of the left ventricle in patients belonging to group A (p <0.05, odds ratio [OR] 12.1). No statistical significance was found between glycemia and the diastolic function of the left ventricle in group B patients. Even in type II diabetic patients without cardiopathy, uncontrolled hyperglycemia provokes diastolic left ventricular dysfunction.

Adult↗