PubMed Health⌕ Search

SEARCH · PubMed Health

Results for “Statistical 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 73 records · Page 4Linked to original sources
PubMed1993

Evaluation of a novel point-of-care system, the i-STAT portable clinical analyzer.

We evaluated a novel system designed for rapid, point-of-care measurement of sodium, potassium, chloride, urea nitrogen, glucose, and hematocrit. The i-STAT Portable Clinical Analyzer (PCA) system is composed of a hand-held analyzer and disposable cartridges. Sample analysis takes place in the cartridge, which contains a series of thin-film electrodes microfabricated on silicon chips. The PCA was evaluated for precision, accuracy, and utility in emergency department and stat laboratory settings. Precision did not differ significantly between these two locations, the CVs being as follows: sodium, 0.46-0.89%; potassium, 1.06-1.45%; chloride, 0.69-2.76%; urea nitrogen, 2.54-6.12%; and glucose, 4.39-5.19%. The assessment of accuracy was based on comparison of patients' sample values analyzed by the PCA and the Kodak Ektachem 700 (or the Coulter ST for hematocrit). Regression statistics were acceptable for all analytes except chloride, for which the regression data were influenced by the limited range of results. A difference plot of the chloride comparison showed that the bias rarely exceeded 5 mmol/L. Mean hematocrit values significantly differed between the PCA and the Coulter ST, apparently because of different calibration procedures.

Blood Chemical Analysis↗
PubMed1998

Regression analysis of multiple protein structures.

A general framework is presented for analyzing multiple protein structures using statistical regression methods. The regression approach can superimpose protein structures rigidly or with shear. Also, this approach can superimpose multiple structures explicitly, without resorting to pairwise superpositions. The algorithm alternates between matching corresponding landmarks among the protein structures and superimposing these landmarks. Matching is performed using a robust dynamic programming technique that uses gap penalties that adapt to the given data. Superposition is performed using either orthogonal transformations, which impose the rigid-body assumption, or affine transformations, which allow shear. The resulting regression model of a protein family measures the amount of structural variability at each landmark. A variation of our algorithm permits a separate weight for each landmark, thereby allowing one to emphasize particular segments of a protein structure or to compensate for variances that differ at various positions in a structure. In addition, a method is introduced for finding an initial correspondence, by measuring the discrete curvature along each protein backbone. Discrete curvature also characterizes the secondary structure of a protein backbone, distinguishing among helical, strand, and loop regions. An example is presented involving a set of seven globin structures. Regression analysis, using both affine and orthogonal transformations, reveals that globins are most strongly conserved structurally in helical regions, particularly in the mid-regions of the E, F, and G helices.

Algorithms↗
PubMed2001

Physiological influences on off-finger glucose testing.

Products for monitoring blood glucose that allow extraction from sites other than the finger have recently been introduced. The FreeStyle Blood Glucose Monitor requires only 0.3 microL of blood, and allows extraction from the hand, arm, and leg, as well as the traditional finger site. Differences in circulatory physiology of the off-finger test sites lead to differences in the measured blood glucose concentration. The first study involved 160 clinic visits by 120 unique subjects with type 1 or type 2 diabetes. FreeStyle measurements were compared to YSI Model 2300 Stat Plus Glucose Analyzer plasma measurements using venous blood, capillary blood from the finger, and capillary blood from the arm. In a second study, the time course of glucose variation was tested by simultaneous measurements on the arm and finger taken every 15 min for 6 h. Thirteen subjects with type 1 diabetes were studied in two 6-h sessions. When FreeStyle was compared to YSI using venous samples and finger samples, the regression statistics were very similar. But when FreeStyle with arm samples was compared to YSI with finger samples, the regression equation was similar, but the scatter in the data was statistically significantly greater at the 95% confidence interval. By studying the time course of glucose changes, the difference between finger and arm measurements was attributed to a time lag in the glucose response on the arm with respect to glucose response on the finger. The lag was observed when the glucose concentration was increasing or decreasing, and the lag time varied from subject-to-subject in the range of 5-20 min. Using the Clarke Error Grid Analysis, the difference between arm and finger glucose measurements was not clinically significant. However, when the glucose concentration is decreasing rapidly into a state of hypoglycemia, the lag in measurements on the arm could delay detection of hypoglycemia. When specifically testing for hypoglycemia, the finger may be the preferable test site.

Arm↗
PubMed1976

A review of the satatistical considerations involved in the treatment of isotope dilution calibration data.

The use of linear regression analysis for the reduction of isotope dilution data is reviewed. The calculation of linear regression statistics is based upon four assumptions: zero variance in the independent variable, equal variance for all values of the dependent variable, linearity and continuity. Unfortunately, isotope dilution data often violate one or more of these assumptions, which results in the calculation of an inaccurate calibration line. The inaccuracies can be avoided through careful inspection of the data, including analyses of variance and linearity. Large differences in the variances of the dependent variable require the use of a weighted linear regression. Nonlinearity necessitates either discarding data in the nonlinear portion of the calibration or the calculation and use of atom % excess and dilution instead of the sample isotope ratios.

Mass Spectrometry↗
PubMedDate not supplied

Nonalcoholic steatohepatitis (NASH) with diabetes: predictors of liver fibrosis.

INTRODUCTION: Nonalcoholic steatohepatitis (NASH) is common cause of chronic liver disease strongly associated with insulin resistance leading to fibrosis. No factors that determine increasing fibrosis have been well recognized. Liver biopsy is considered as gold standard for diagnosis and prognosis of this disease. AIM: To identify independent predictive factors of liver fibrosis in patients of NASH with diabetes. MATERIAL AND METHODS: During the year 2001 and 2002 total 36 patients of NASH associated with diabetes were included in the study. The diagnosis of NASH was based on 1) presence of steatosis, inflammation and ballooning on liver biopsy 2) Intake of alcohol < 20 gm of ethanol per week 3) Exclusion of other liver diseases. Patients were labeled as diabetic if random glucose was > 200 mg/dL or fasting glucose more than 140 mg/dL on 2 occasion or having documented use of oral hypoglycemic medications or insulin. Clinical and biochemical variables such as age, sex, obesity, hypercholesterolemia, AST, ALT and AST: ALT were examined for predictors of fibrosis using univariate and multiple regression statistical analysis. Obesity was defined as BMI > 30 for both males and females. Hypercholesterolemia was considered when fasting cholesterol level was above 95th percentile of normal on at least 2 occasions. Fibrosis was noted as present or absent on histology. RESULTS: Of 36 patients 17 were females and 19 males with age range of 25 to 75 years, mean age 50.8 years. Fibrosis was present in 11 (30.5%) and absent in 25 (69.4%) patients. Univariate and multiple correlations co-efficient failed to detect significant association of fibrosis with above mentioned variables. However multiple regression and logistic regression analysis (MLR) detected statistical significance for AST, ALT levels and AST: ALT ratio between fibrosis and no fibrosis in 80.6% patients. CONCLUSION: There is no definite noninvasive test that helps to predict liver fibrosis however AST, ALT levels and AST: ALT ratio may help to determine the fibrosis in patients of NASH with diabetes in majority of cases.

Adult↗
PubMed1991

Function learning: induction of continuous stimulus-response relations.

This research investigates how people learn to select particular response magnitudes along one physical dimension (duration) when given stimulus magnitudes from another dimension (spatial extent). Stimuli and correct responses were related by either a power function, a logarithmic function, or a linear function with a positive intercept. The power function was learned quickly and accurately. In contrast, systematic response biases occurred during the early phases of learning the logarithmic and linear functions. As practice progressed, however, the biases gradually disappeared. These results support an adaptive regression model. According to it, people learn functions through a subjective process analogous to statistical regression. There is assumed to be an initial constraint that treats stimulus-response pairs as if a power function characterizes them.

Female↗
PubMed2000

Noninvasive measurement of total serum bilirubin in a multiracial predischarge newborn population to assess the risk of severe hyperbilirubinemia.

BACKGROUND: Jaundice in near-term and term newborns is a frequent diagnosis that may prompt hospital readmission in the first postnatal week. Hyperbilirubinemia, when excessive, can lead to potentially irreversible bilirubin-induced neurotoxicity. Predischarge risk assessment (at 24-72 hours of age) for subsequent excessive hyperbilirubinemia is feasible by a laboratory-based assay of total serum bilirubin (TSB). Hypothesis. Noninvasive, transcutaneous, point-of-care measurement of transcutaneous bilirubin (TcB) predischarge by multiwavelength spectral analysis, using a portable BiliCheck device (SpectRx Inc, Norcross, GA), is clinically equivalent to measurement of TSB in a diverse, multiracial term and near-term newborn population and predictive of subsequent hyperbilirubinemia. METHODOLOGY: We evaluated a hand-held device that uses multiwavelength spectral reflectance analysis to measure TcB (BiliCheck). The study population (490 term and near-term newborns) was racially diverse (59.1% white, 29.5% black, 3.46% Hispanic, 4.48% Asian, and 3.46% other) and was evaluated at 2 separate institutions using multiple (11) devices. The postnatal age ranged from 12 to 98 hours and the ranges of birth weights and gestational ages were 2000 to 5665 g and 35 to 42 weeks, respectively. All transcutaneous evaluations were performed contemporaneously and paired with a heelstick TSB measurement. All TSB assays were performed by high performance liquid chromatography, as well as by diazo dichlorophenyldiazonium tetrafluoroborate techniques. RESULTS: TSB values ranged from .2 to 18.2 mg/dL (mean +/- standard deviation: 7.65 +/- 3.35 mg/dL). The overall correlation of TSB (by high performance liquid chromatography technique) to TcB (by BiliCheck devices) was linear and statistically significant (r =.91; r(2) =.83; TcB =.84; TSB = +.75; standard error of regression line = 1.38; P <.001; n = 490 infants; 1788 samples). Similar regression statistics were evident in subset populations categorized by race (white: r =.91 [n = 289 infants]; black: r =.91 [n = 145 infants]) as well as by gestation (term: r =. 91 [n = 1625 samples]; near-term: r =.89 [n = 163 samples]). Intradevice precision was determined to be.59 mg/dL (2-3 measurements per infant with 1 device; n = 210 infants; 510 samples in a separate subset). Interdevice evaluation of 11 devices determined the precision to be.68 mg/dL (2-4 devices used for measurements per patient). In 23 of 419 of the study population infants who were in the 24- to 72-hour age range, the predischarge TSB values designated them to be at high risk for subsequent excessive hyperbilirubinemia (above the 95th percentile track on the hour-specific bilirubin nomogram). For these infants, the paired BiliCheck TcB values were all above the 75th percentile track (negative predictive value = 100%; positive predictive value = 32. 86%; sensitivity = 100%; specificity = 88.1%; likelihood ratio = 8. 43). CONCLUSIONS: Our data demonstrate the accuracy and reproducibility of the predischarge BiliCheck measurements in term and near-term newborn infants of diverse races and ethnicities. Infants with predischarge BiliCheck values above the 75th percentile of hour-specific TSB values on the bilirubin nomogram may be considered to be at high risk for subsequent excessive hyperbilirubinemia. Further studies are needed to assess the efficacy of this technique in preterm infants, those undergoing phototherapy, and those with TSB values of >/=15 mg/dL (>/=256 micromol/L).

Analysis of Variance↗
PubMed2000

[Hospital capacity and waiting time for treatment--is there a connection?].

BACKGROUND: The association between hospital capacity and waiting time for treatment is uncertain. MATERIAL AND METHODS: Waiting times for patients on waiting lists for inpatient treatment in 1998 were analysed to disclose possible associations with the hospitals' treatment resources, i.e., general costs, number of beds, doctors or nurses in relation to the population of its catchment area, and the relation between acute and elective admissions. Waiting times were calculated from the National Patient Register, which collects information on hospital stays. Resource data and data on acute admissions were taken from the SAMDATA publications for 1998. RESULTS: Median waiting time varied from 50 to 300 days among the hospitals. Statistical regression models were, however, unable to explain the variation in waiting time on the basis of any variable related to hospital resources or acute admissions that may influence the capacity for elective admissions. INTERPRETATION: To avoid breaches of the guarantees for patients guaranteed a maximum of three months on the waiting list, the median waiting time should be below 12-15 days. This goal may, however, be much too ambitious in view of the fact that the median waiting time for patients with mammary or colon cancer is about 30 days.

Catchment Area, Health↗
PubMed1975

Characteristics of mentally-impaired aged profiting from individualized treatment.

This study investigated subjects' characteristics which related to improvements from individualized treatment of mentally impaired aged. Project variables (158) were reduced to 12 Q-type factors. The excess disability (ED) criteria were partialled into experimental/control influences (EDe/c) and residual ED variance. Statistical regressions were run on the EDe/c criteria vs. age and the 12 Q-factors. Age significantly predicted post-treatment ED functioning in both conditions. Younger subjects (ages 65-82) profited most in both groups. After the statistical removal of age in the regression, Factor Q-3 still retained a significant relationship to EDe/c for the experimentals. The correlations indicate the importance of social capabilities, irrespective of age level, with regard to improvements due to ED treatments. For the control subjects, Q-12 and Q-5 showed significant independent relationships to the criteria. Subjects' competence appeared most important for EDe/c gains made without treatment.

Activities of Daily Living↗
PubMed2005

A multivariate analysis of otological, surgical and patient-related factors in determining success in myringoplasty.

OBJECTIVES: The aim of this study was to evaluate the effects of smoking and other prognostic factors on long-term surgical outcome and hearing results in myringoplasty. DESIGN: Retrospective cohort. SETTING: Tertiary medical centre. PARTICIPANTS: Eighty myringoplasty operations were analysed retrospectively from the charts of 74 patients who underwent myringoplasty between January 2000 and November 2002 at the First ENT Clinic of Ataturk Training and Research Hospital, using the chi-squared test, Fisher's exact test, t-test and multiple logistic regression statistical analysis. MAIN OUTCOME MEASURES: The effects of prognostic factors and smoking on myringoplasty were investigated by variables noted from patients' files such as smoking status, operation type, duration of dry period of the diseased ear, peroperative otorrhoea, status of the opposite ear, presence of perforation at admission, place of perforation, size of perforation, status of middle ear mucosa, presence of cholesteatoma and tympanosclerosis, ossicular status, previous and related surgery, experience of the surgeon, presence of anterior overhang, presence of valsalva manoeuvre, postoperative antiobiotic cover and income per day. RESULTS: The overall take rate of the myringoplasty graft was 71%. Based on the univariate analysis, smoking status of the patients (P = 0.008), status of the opposite ear (P = 0.01), size of perforation (P = 0.009) and the experience of the surgeon (P = 0.002) were found to be statistically significant (P < 0.01) prognostic factors influencing the success rate of the operations. Multiple logistic regression analysis was subsequently carried out on these prognostic factors and yielded the following classification (95% CI): senior surgeon (OR = 14.3), non-smoking patient (OR = 11.4), longer duration of dry period of the diseased ear (OR = 5.3), normal opposite ear (OR = 5.0) and small perforation (OR = 4.8). CONCLUSIONS: A non-smoking patient, a longer dry ear, a healthy opposite ear, a relatively smaller perforation and a senior surgeon were found to be significant prognostic factors positively influencing the success rate of myringoplasty.

Adolescent↗
PubMed1991

[Croup and air pollutants: results of a two-year prospective longitudinal study].

The Reutlingen/Tübingen/Rottenburg region in Baden-Württemberg is characterized by medium and small sized towns and rural areas. In 1986/87 875 cases of croup were registered there by the treating physicians during a 24-months period. In consideration of meteorological and virological "disturbing variables" the influence of the measured air pollution by SO2, NO, NO2, CO, ozone and dust on croup frequency was computed by means of statistical regression methods. For the months September till March, the main manifestation period of croup, weak but statistically significant influences of the daily means of NO and NO2 were found, for the whole year influences of NO, NO2, and CO. During the winter months temperature correlates positively and velocity of wind negatively, both statistically with significance, to croup frequency. The essential conditions of croup are individual and familiar disposition on the one hand, virus infections on the other. Air pollution of a concentration like given in the investigated region was found to be a weak additional factor that favours the manifestation of croup.

Air Pollutants↗
PubMed2005

Predictors of remission after electroconvulsive therapy in unipolar major depression.

CONTEXT: Electroconvulsive therapy (ECT) is the most effective biological treatment for major depression. However, there is little agreement about clinically useful predictors of acute ECT outcomes. OBJECTIVE: To assess whether age, sex, burden of comorbid physical illness, age at onset, history of recurrence, episode duration, chronic depression or comorbid dysthymia, melancholic features, episode severity, and medication resistance are predictors of remission after an acute course of ECT. DESIGN: We performed an analysis using data gathered prospectively in 328 patients with unipolar major depression (according to Research Diagnostic Criteria) treated with ECT. The study was conducted from 1993 through 1999. Patients had a pretreatment score of 21 or higher on the 24-item Hamilton Rating Scale for Depression (HAM-D). Treatment history was assessed using the Antidepressant Treatment History Form. Remission was defined as a 24-item HAM-D score of 10 or less and a 60% or more relative reduction of the HAM-D score. RESULTS: On univariate logistic regression, statistically significant predictors of nonremission were chronic depression/dysthymia, medication resistance, longer episode duration, and younger age. On backward elimination logistic regression, only medication resistance (OR = 1.67, 95% CI = 1.05 to 2.67) and chronic depression/ dysthymia (OR = 1.84, 95% CI = 1.06 to 3.21) were statistically significant predictors of nonremission. CONCLUSIONS: In patients with major depression, lower rates of remission after acute ECT are associated with medication resistance and chronicity, but not with age or burden of physical illness.

Age of Onset↗
PubMed2005

Measuring parallelism, linearity, and relative potency in bioassay and immunoassay data.

There is often a need to determine parallelism or linearity between pairs of dose-response data sets for various biological applications. This article describes a technique based on a modification of the well-known extra-sum-of-squares principle of statistical regression. The standard extra-sum-of-squares method uses an F-distributed ratio as a statistic and an F-test based on this statistic as the parallelism test. It is shown here that this metric does not directly measure the parallelism between the two curves and can often vary in opposition to actual parallelism. To overcome this problem, a metric based on a chi-square test applied directly on the chi-square-distributed extra-sum-of-squares statistic is developed, which is shown to correspond directly to parallelism. This parallelism metric does not suffer from the shortcomings of the conventional F-test-based metric, and is a more reliable and appropriate measure of parallelism. The article also shows that the choice of curve model has a large effect on the sensitivity of either metric, and that using an asymmetric model, such as the asymmetric five-parameter logistic function, a generalization of the commonly used symmetric four-parameter logistic function, is necessary when working with asymmetric dose-response data. The effect of noise, as well as the importance of correct weighting on the parallelism metrics and the relative potency, is also studied.

Algorithms↗
PubMed2003

[Ultrastructural analysis of the role of microtubules in the transport and exocytosis of protein secretion in secretory cells of mouse mammary glands].

Morphometrical analysis of microtubules (MT) and coated vesicles (CV) was done on electron micrographs of the murine mammary gland sections at the final stages of functional maturation of secretory cells (SC), as well as at different stages of the cell secretory cycle (CSC). The results obtained allow to make the following conclusions: 1) dynamics of changes in the MT count in different subplasmalemmal cell compartments is similar, being directly associated with heterochronical character of CSC processes; 2) the maximum MT count in SC was seen during formation of the secretion product, this increase occurring at the expense of short MT, which are disassembled before and during exocytosis of secretory vesicle (SV) contents; 3) changes in the number of MT and CV are of similar character, which was revealed by morphometrical analysis of SC in the course of maturation. The number of MT and CV in SC increased rapidly and significantly on the 1st and 2nd days after parturition, compared to that observed 1-2 days before parturition. However, the number of MT and CV decreased by the 10th day of lactation, when the mammary secretory activity reached its maximum. Both correlation and regression statistical analyses made during the CSC development point to a linear relation between MT (x) and CV (y) numbers. Regression of y on x is: y = -0.89 + 12.43x. A hypothesis about the possibility of MT participation in CV and SV transport and in the formation of a barrier on the path leading to exocytosis of SV contents is suggested.

Animals↗
PubMed1989

Models of spinal trabecular bone loss as determined by quantitative computed tomography.

Noninvasive bone densitometry techniques have significantly improved our understanding of the pattern and magnitude of bone loss over the life span. Quantitative computed tomography (QCT) is capable of selectively measuring highly labile trabecular bone in the central portion of the vertebrae. Trabecular bone mineral density (mg/ml) was determined in 538 healthy women ranging in age from 20 to 80 using GE CT/T scanners at 80 kVp. Various statistical regressions were performed for the entire population to describe the general pattern of bone loss from the spine; a cubic model (r = -0.69, SEE = 26.0 mg/ml) was found to be statistically superior (p less than 0.01) to linear, quadratic, or exponential models. An average bone loss was approximated from these regression analyses with a yearly absolute rate of -2.02 +/- 0.097 mg/ml per year (p less than 0.0001). The average rate of change for premenopausal women was -0.45 mg/ml per year (p less than 0.05), for perimenopausal women was -4.39 mg/ml per year (p less than 0.0001) and for postmenopausal women was -1.99 mg/ml per year (p less than 0.0001). QCT values were also stratified into 5 and 10 year age groups and analyzed separately for pre- and postmenopausal women. The 5 and 10 year interval stratification revealed no identifiable bone density decrements prior to midlife using analysis of variance statistical methods; significant losses of bone mineral density were noted to correspond with the usual time of menopause and to continue into old age. Various two-phase regressions were employed using age and menstrual status to improve the description of age- and menopause-related bone loss.(ABSTRACT TRUNCATED AT 250 WORDS)

Adult↗
PubMed2006

Analysis of traffic injury severity: an application of non-parametric classification tree techniques.

Statistical regression models, such as logit or ordered probit/logit models, have been widely employed to analyze injury severity of traffic accidents. However, most regression models have their own model assumptions and pre-defined underlying relationships between dependent and independent variables. If these assumptions are violated, the model could lead to erroneous estimations of injury likelihood. The classification and regression tree (CART), one of the most widely applied data mining techniques, has been commonly employed in business administration, industry, and engineering. CART does not require any pre-defined underlying relationship between target (dependent) variable and predictors (independent variables) and has been shown to be a powerful tool, particularly for dealing with prediction and classification problems. This study uses the 2001 accident data for Taipei, Taiwan. A CART model was developed to establish the relationship between injury severity and driver/vehicle characteristics, highway/environmental variables and accident variables. The results indicate that the most important variable associated with crash severity is the vehicle type. Pedestrians, motorcycle and bicycle riders are identified to have higher risks of being injured than other types of vehicle drivers in traffic accidents.

Accidents, Traffic↗
PubMed1984

A simple computer program for Scatchard plot analysis of hormone receptors including statistical analysis on a low cost desk top calculator.

A simple procedure for the estimations of standard errors in hormone binding parameters derived from Scatchard analysis is proposed by assuming an univariate linear regression statistic model. The complete set of necessary formulae is presented and a practical way for calculation is solved using the binding of human growth hormone to rat adrenal membranes, as a practical example. The comparison between the results obtained by the method proposed and by using a well proved and accepted computer program, shows unequivocally that the differences are practically undistinguishable. The complete program to calculate the Scatchard parameters and their statistical analysis is available on request.

Binding Sites↗