SEARCH · PubMed Health

Results for “multiple 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 487 records · Page 27Linked to original sources

PubMed2004

Predicting thermal displacements in modular tool systems.

In the last decade, there has been an increasing interest in compensating thermally induced errors to improve the manufacturing accuracy of modular tool systems. These modular tool systems are interfaces between spindle and workpiece and consist of several complicatedly formed parts. Their thermal behavior is dominated by nonlinearities, delay and hysteresis effects even in tools with simpler geometry and it is difficult to describe it theoretically. Due to the dominant nonlinear nature of this behavior the so far used linear regression between the temperatures and the displacements is insufficient. Therefore, in this study we test the hypothesis whether we can reliably predict such thermal displacements via nonlinear temperature-displacement regression functions. These functions are estimated first from learning measurements using the alternating conditional expectation (ACE) algorithm and then tested on independent data sets. First, we analyze data that were generated by a finite element spindle model. We find that our approach is a powerful tool to describe the relation between temperatures and displacements for simulated data. Next, we analyze the temperature-displacement relationship in a silent real experimental setup, where the tool system is thermally forced. Again, the ACE algorithm is powerful to estimate the deformation with high precision. The corresponding errors obtained by using the nonlinear regression approach are 10-fold lower in comparison to multiple linear regression analysis. Finally, we investigate the thermal behavior of a modular tool system in a working milling machine and again get promising results. The thermally induced errors can be estimated with 1-2 microm accuracy using this nonlinear regression analysis. Therefore, this approach seems to be very useful for the development of new modular tool systems.

Computer Simulation↗

PubMed2004

Advanced statistics: linear regression, part I: simple linear regression.

Simple linear regression is a mathematical technique used to model the relationship between a single independent predictor variable and a single dependent outcome variable. In this, the first of a two-part series exploring concepts in linear regression analysis, the four fundamental assumptions and the mechanics of simple linear regression are reviewed. The most common technique used to derive the regression line, the method of least squares, is described. The reader will be acquainted with other important concepts in simple linear regression, including: variable transformations, dummy variables, relationship to inference testing, and leverage. Simplified clinical examples with small datasets and graphic models are used to illustrate the points. This will provide a foundation for the second article in this series: a discussion of multiple linear regression, in which there are multiple predictor variables.

Biometry↗

PubMed1995

Burns mortality and hospitalization time--a prospective statistical study of 352 patients in an Asian National Burn Centre.

A prospective study of 352 patients in an Asian National Burn Centre has been used to develop statistical predictive models for mortality and hospitalization time. The patients are largely of Asian origin. Total burn surface area (% TBSA) and presence of respiratory burns are significant independent predictors of mortality in the multiple logistic regression analysis with an accuracy of 98.3 per cent. Age is not a significant predictor of mortality in our patients. Age, % TBSA, full thickness % TBSA and respiratory burns are significant independent predictors of length of stay in hospital with a R2 value of 0.57 in the multiple linear regression analysis. There were 16 deaths, many of whom had developed multiple complications, common causes of which were sepsis, bronchopneumonia, DIVC and multiorgan failure. The final causes of death were septicaemic shock in 10 patients, extensive burns in four patients, ARDS in one patient and bleeding peptic ulcer in one patient. The development of these two mortality and morbidity predictive models is the first step in the evaluation of our results. These models have to be tested against a future set of patients. After confirmation they will aid in patient management, clinical audit, patient and family counselling. They will also serve as baseline standards for evaluation of new therapies, assist us in the allocation of resources and identifying the at-risk population for improvements in therapy.

Adolescent↗

PubMed1998

Factors associated with very early weaning among primiparas intending to breastfeed.

OBJECTIVES: The major objective of this study was to identify predictor variables that accurately differentiated breastfeeding women who weaned during the first 4 weeks, those who weaned between 5 and 26 weeks, and those who weaned after 26 weeks. Predictors were demographic variables, Theory of Planned Behavior (TPB) variables, breastfeeding knowledge, and difficulties experienced during the first month. METHODS: Primiparas who delivered healthy infants in an urban midwestern hospital provided initial data prior to discharge. Follow-up occurred at 1, 3, 6, 9, and 12 months. Following appropriate bivariate analyses, polychotomous logistic regression was used to determine predictors of weaning group. Linear multiple regression was used to predict intended duration. RESULTS: Most of the 84 women who weaned very early had intended to breastfeed considerably longer. According to the multivariate analysis, women who weaned earlier were younger, had completed fewer years of education, had a more positive bottle-feeding attitude and a less positive breastfeeding attitude, intended to breastfeed less time, had lower knowledge scores, had higher perceived insufficient milk scores, and planned to work outside the home. Variables postulated by the TPB to be direct predictors of intention explained 36% of the variance in intended duration. CONCLUSIONS: Women at risk for early weaning can be identified with reasonable accuracy using a TPB-based conceptual framework expanded to include breastfeeding specific variables. Casefinding using empirically derived screening methods and careful postpartum follow-up, along with professional intervention, should be used to avert unintended early weaning.

Adolescent↗

PubMed1985

A free lignocaine index as a guide to unbound drug concentration.

A free lignocaine index was developed on the basis of measurements of plasma lignocaine and its principle binding protein, alpha 1-acid glycoprotein (AAG) in 80 samples from 16 patients admitted to the coronary care unit and given prophylactic lignocaine therapy. The free drug fraction, fu, of lignocaine was determined by equilibrium dialysis and its relationship to AAG and total lignocaine concentration (T) defined by multiple linear regression analysis as l/fu = 1.45 + 0.023 (AAG) -0.129 (T) (multiple r = 0.872, P less than 0.001). This relationship was used to calculate the 'free lignocaine index' as fu X T and compared with the observed value obtained by equilibrium dialysis of 178 samples from 41 separate subjects who received lignocaine after suspected myocardial infarction. There was a highly significant relationship (r = 0.933, n = 178, P less than 0.001) between the observed and predicted values. We conclude that the free drug index may be useful in rapidly assessing the unbound (free) concentration of lignocaine in plasma.

Humans↗

PubMed2000

Generalized additive distributed lag models: quantifying mortality displacement.

There are a number of applied settings where a response is measured repeatedly over time, and the impact of a stimulus at one time is distributed over several subsequent response measures. In the motivating application the stimulus is an air pollutant such as airborne particulate matter and the response is mortality. However, several other variables (e.g. daily temperature) impact the response in a possibly non-linear fashion. To quantify the effect of the stimulus in the presence of covariate data we combine two established regression techniques: generalized additive models and distributed lag models. Generalized additive models extend multiple linear regression by allowing for continuous covariates to be modeled as smooth, but otherwise unspecified, functions. Distributed lag models aim to relate the outcome variable to lagged values of a time-dependent predictor in a parsimonious fashion. The resultant, which we call generalized additive distributed lag models, are seen to effectively quantify the so-called 'mortality displacement effect' in environmental epidemiology, as illustrated through air pollution/mortality data from Milan, Italy.

Journal Article↗

PubMed1978

Skin colorimetry in Belize. I. Conversion formulae.

Comparisons between skin colorimetry reports have been hampered by the common use of two different types of portable reflectometers, which sample reflectances at different wavelengths. In an attempt to provide direct comparability between the two machines, multiple linear regression equations were derived from reflectance spectrophotometry readings on 308 Black Caribs and 175 Creoles in Belize, Central America, using both machines. Cross validation tests show the coefficients presented are applicable to independent data sets and generally applicable to other heavily pigmented populations. Comparisons with previously published conversion formulae, which were from very small samples using simple linear regression, show a definite improvement in predictive accuracy when using multiple regression equations based on a large sample.

Belize↗

PubMed2004

Prediction of gas chromatographic retention indices of a diverse set of toxicologically relevant compounds.

For a set of 846 organic compounds, relevant in forensic analytical chemistry, with highly diverse chemical structures, the gas chromatographic Kovats retention indices have been quantitatively modeled by using a large set of molecular descriptors generated by software Dragon. Best and very similar performances for prediction have been obtained by a partial least squares regression (PLS) model using all considered 529 descriptors, and a multiple linear regression (MLR) model using only 15 descriptors obtained by a stepwise feature selection. The standard deviations of the prediction errors (SEP), were estimated in four experiments with differently distributed training and prediction sets. For the best models SEP is about 80 retention index units, corresponding to 2.1-7.2% of the covered retention index interval of 1110-3870. The molecular properties known to be relevant for GC retention data, such as molecular size, branching and polar functional groups are well covered by the selected 15 descriptors. The developed models support the identification of substances in forensic analytical work by GC-MS in cases the retention data for candidate structures are not available.

Calibration↗

PubMed2005

Determination of the age of sherry wines by regression techniques using routine parameters and phenolic and volatile compounds.

This paper describes a study of the possibility of obtaining regression models by means of partial least squares (PLS) and multiple linear regression (MLR) that would enable us to correlate a series of parameters, such as the concentration of short-chain organic acids, higher alcohols, and phenolic compounds with the age of vintage Sherry wines or "anadas". The aim of this study is to characterize how these parameters evolve with aging. If this could be done, it would then be possible to guarantee the age of such wines using objective variables. A PLS regression model was established that allows the age to be predicted with a mean deviation of 1.6 years with respect to the age of the wines. In the case of the MLR, a model with 6 variables was obtained that gives a mean deviation of 3.3 years in the predictions.

Least-Squares Analysis↗

PubMed2001

High serum retinyl esters are not associated with reduced bone mineral density in the Third National Health And Nutrition Examination Survey, 1988-1994.

Hypervitaminosis A is sometimes associated with abnormalities of calcium metabolism and bone mineral status. A recent study found a negative association between reported dietary vitamin A intake and bone mineral density (BMD). Some segments of the U.S. population have high fasting serum retinyl ester concentrations, a physiological marker that may reflect high and possibly excessive vitamin A intake. We examined the association between fasting serum retinyl esters and BMD in the Third National Health and Nutrition Examination Survey, 1988-1994 (NHANES III), a large, nationally representative sample of the U.S. population. BMD was measured for the femoral neck, trochanter, intertrochanter, and total hip on all nonpregnant participants aged > or = 20 years; 5,790 participants also had complete data on fasting serum retinyl esters and covariates including age, body mass index (BMI), smoking, alcohol consumption, dietary supplement use, diabetes, physical activity, and, among women, parity, menopausal status, and the use of oral contraceptives or estrogen-replacement therapy. The sample included non-Hispanic white, non-Hispanic black, and Mexican American men and women. We examined the association between fasting serum retinyl esters and BMD at each site, controlling for covariates with multiple linear regression. We examined the association with osteopenia and osteoporosis with multiple logistic regression. Although the prevalences of high fasting serum retinyl esters concentration and low BMD were both substantial in this sample, there were no significant associations between fasting serum retinyl esters and any measure of bone mineral status.

Adult↗

PubMedDate not supplied

[Domestic violence during pregnancy and its relationship with birth weight].

OBJECTIVE: To determine the prevalence of domestic violence during pregnancy and its impact on birth weight and the immediate post-partum period. MATERIAL AND METHODS: We conducted a survey of 110 pregnant women who delivered at the Hospital Civil in Cuernavaca, Morelos. The questionnaire was applied by specialized personal. We used multiple linear regression to adjust for differences between birth weight means and multiple logistic regression for complications. RESULTS: In our study, women who suffered violence during pregnancy had three times more complications during delivery (Cl 95% 1.3-7.9). The difference in birth weight of newborns of these women was 560 g less (p < 0.01 adjusted by age and parity) in comparison with women who did not undergo violence during pregnancy. Women who suffered violence during pregnancy had a four times greater risk for having low birth weight babies (Cl 95% 1.3-12.3) than the non-battered women. CONCLUSIONS: We propose more research be done on this topic, including studies of other population groups. Also, health personnel should be educated that violence towards women could constitute a reproductive risk.

Adolescent↗

PubMed1997

Mammography screening and the increase in breast cancer incidence in Hawaii.

This ecological study investigated the association between mammography utilization and breast cancer incidence in Hawaii with the hypothesis that geographic areas with high mammography use have higher breast cancer incidence than geographic areas with low mammography use. Insurance claims for mammograms received during 1992 and 1993 were combined with breast cancer incidence data from the Hawaii Tumor Registry and data from the 1990 Census ZIP File. The claims data were obtained from four private and three public health plans and covered approximately 85% of women 40 years of age and older. Age-specific breast cancer incidence rates for the 79 ZIP code areas were regressed on mammography rates and selected aggregate demographic variables using multiple linear regression. An estimated 42% of women 40 years of age and older had received at least 1 mammogram during 1992 and 1993, with the highest rate (45%) in women ages 50-64 years old. Overall, 23% of the variation in age-specific breast cancer incidence could be predicted by mammography utilization, 23% by increasing age, and 4% by higher education. The relationship between mammography use and breast cancer incidence was strongest for women 50-64 years old and for localized disease. The magnitude of the association between breast cancer incidence and mammography utilization was comparable to the increase in breast cancer rates observed in Hawaii during the mid-1980s, supporting the hypothesis that the sharp increase in breast cancer incidence was attributable to screening and early detection. However, the long-term 1% increase in breast cancer incidence requires alternate explanations.

Adult↗

PubMed1996

Length of inpatient stay and recidivism among patients with schizophrenia.

OBJECTIVE: The study examined whether length of hospital stay is related to recidivism among psychiatric patients. A quasi-experimental approach was used to address limitations of controlled and epidemiological research. METHODS: Three matched groups, each consisting of 55 inpatients with schizophrenia, were selected from public psychiatric units with different mean lengths of stay. Regression models were used to compare the groups on three variables: time to first readmission (survival analysis), number of readmissions (ordinal logit regression), and total time in the community in the postdischarge year (multiple linear regression). RESULTS: An analysis based on the units with different lengths of stay, which was similar to that typically used in controlled studies, found no differences in the three outcome measures. However, a second analysis that examined data for all patients irrespective of their unit assignment found that inpatients treated for 30 days or less relapsed sooner than those with stays longer than 30 days. The disparity in results was largely due to overlapping quasi-experimental conditions: many patients on the short-stay units had a long lengths of stay, and vice versa. The first analysis supports an administrative policy of short stays. The second reinforces previous findings that a group of patients, primarily young males with onset of illness at an early age and multiple previous hospitalizations, is at greater risk of relapse with short-term treatment. CONCLUSIONS: The apparent contradiction between a unit- or patient-based analysis suggests that unit-based results should be interpreted with caution when used to make clinical or utilization review decisions.

Adult↗

PubMed1989

Determinants of maximum exercise capacity in patients with chronic airflow obstruction.

Patients with chronic airflow obstruction (CAO) often develop impairment of respiratory muscle function. We hypothesized that inspiratory muscle strength, as assessed by resting, peak inspiratory pressure (PIP) may be an important determinant of maximum exercise capacity in patients with CAO. Twenty ambulatory male patients (mean age, 56 +/- 3 years [+/- SE]) with CAO (FEV1, 1.72 +/- 0.21 L) comprised the studied population. Oxygen consumption at incremental cycle ergometry to tolerance (VO2max, 1.80 +/- 0.20 L/min) served as the dependent variable for regression vs measures of resting pulmonary function. Significant correlations with VO2max included power output in watts (r = 0.951), VEmax (r = 0.858), Dsb (r = 0.841), PIP (r = 0.816), age (r = -0.809), FEV1 (r = 0.763), and FVC (r = 0.663). The FEV1, Dsb, and PIP each entered into a multiple linear regression relationship describing VO2max. Also, when paired with VEmax as independent variables in multiple regression, PIP and Dsb each improved description of VO2max over VEmax alone (p less than 0.05), whereas FEV1 and FVC did not (p greater than 0.05). We conclude that factors other than ventilatory capacity also have a quantitative effect on VO2max and that PIP constitutes a determinant of maximum exercise capacity in patients with CAO.

Exercise Test↗

PubMed2003

A multiple regression model of combined duplex criteria for detecting threshold carotid stenosis and predicting the exact degree of carotid stenosis.

BACKGROUND AND PURPOSE: Combined duplex criteria are commonly used in vascular laboratories for evaluating carotid stenosis. However, most of these combinations are empirical, and systemic validation is lacking. This study was completed using a multiple regression method to evaluate the accuracy of different combined duplex criteria for detecting threshold carotid stenosis and predicting the exact degree of carotid stenosis on angiography. METHODS: Two hundred sixty-six sets of unilateral carotid duplex and angiographic data were randomly divided into 2 sets: a derivation set and a validation set. The derivation set was used to develop a multiple logistic regression model for detecting 70% threshold carotid stenosis. Age, sex, systolic blood pressure, diastolic blood pressure, Doppler peak systolic velocity (PSV), Doppler and diastolic velocity (EDV), the systolic carotid ratio (SCR), and ophthalmic artery flow direction were tested as independent variables. A multiple linear regression model was also developed for predicting the exact degree of carotid stenosis on angiogram. The validation set was then used to evaluate the accuracy of these models. RESULTS: According to the logistic regression strategy, the best multiple logistic regression model was as follows: probability of threshold carotid stenosis = exp(2.6 PSV - 6.2)/[1 + exp(2.6 PSV - 6.2)]. The best linear regression model was as follows: degree of carotid stenosis = 20.2 PSV - 7.4 EDV + 0.4 SCR + 8.5. Both models proved to be valid following an evaluation of the validation set. CONCLUSIONS: This study illustrates that Doppler parameters may be of use in predicting the exact degree of carotid stenosis and the probability of threshold carotid stenosis. This is important if duplex criteria are going to replace angiography as the only tool for selecting endarterectomy candidates.

Aged↗

PubMed2003

Source apportionment of polycyclic aromatic hydrocarbons in the urban atmosphere: a comparison of three methods.

Polycyclic aromatic hydrocarbons (PAHs) are ubiquitous pollutants in urban atmospheres. Several PAHs are known carcinogens or are the precursors to carcinogenic daughter compounds. Understanding the contributions of the various emission sources is critical to appropriately managing PAH levels in the environment. The sources of PAHs to ambient air in Baltimore, MD, were determined by using three source apportionment methods, principal component analysis with multiple linear regression, UNMIX, and positive matrix factorization. Determining the source apportionment through multiple techniques mitigates weaknesses in individual methods and strengthens the overlapping conclusions. Overall source contributions compare well among methods. Vehicles, both diesel and gasoline, contribute on average 16-26%, coal 28-36%, oil 15-23%, and wood/other having the greatest disparity of 23-35% of the total (gas- plus particle-phase) PAHs. Seasonal trends were found for both coal and oil. Coal was the dominate PAH source during the summer while oil dominated during the winter. Positive matrix factorization was the only method to segregate diesel from gasoline sources. These methods indicate the number and relative strength of PAH sources to the ambient urban atmosphere. As with all source apportionment techniques, these methods require the user to objectively interpret the resulting source profiles.

Air Pollutants↗

PubMed2005

Elevated C-reactive protein augments increased arterial stiffness in subjects with the metabolic syndrome.

We examined whether the presence of an increasing number of metabolic syndrome "disorders" was associated with an increasing pulse wave velocity, which is recognized as a marker of cardiovascular risk, and evaluated whether an elevated plasma C-reactive protein level augments this increasing pulse wave velocity. Using a cross-sectional study design, C-reactive protein, metabolic syndrome-related anthropometric parameters, and pulse wave velocity were measured in 5752 middle-aged Japanese men (44+/-10 years old). In linear regression analyses, all of the metabolic "disorders" and the logarithm of the C-reactive protein significantly correlated with pulse wave velocity. Multiple linear regression analysis demonstrated that triglycerides, HDL cholesterol, mean blood pressure, fasting glucose, and the logarithm of the C-reactive protein were significant independent positive predictors of pulse wave velocity (R-square=0.38). The presence of an increasing number of metabolic "disorders" in the subjects was associated with an increasing pulse wave velocity (no disorders 1228+/-139 cm/s > or =3 disorders 1437+/-250 cm/s; P<0.01). Among subjects with the metabolic syndrome, pulse wave velocity was higher in cases with (1508+/-278 cm/s) than in those without an elevated C-reactive protein (1427+/-243 cm/s; P<0.01). In conclusion, an increase in arterial stiffness may constitute a pathophysiological basis for the increased risk of cardiovascular disease in patients with the metabolic syndrome and that an elevated C-reactive protein level may aggravate this cardiovascular risk.

Adult↗

PubMed2006

ALIFE@Work: a randomised controlled trial of a distance counselling lifestyle programme for weight control among an overweight working population [ISRCTN04265725].

BACKGROUND: The prevalence of overweight is increasing and its consequences will cause a major public health burden in the near future. Cost-effective interventions for weight control among the general population are therefore needed. The ALIFE@Work study is investigating a novel lifestyle intervention, aimed at the working population, with individual counselling through either phone or e-mail. This article describes the design of the study and the participant flow up to and including randomisation. METHODS/DESIGN: ALIFE@Work is a controlled trial, with randomisation to three arms: a control group, a phone based intervention group and an internet based intervention group. The intervention takes six months and is based on a cognitive behavioural approach, addressing physical activity and diet. It consists of 10 lessons with feedback from a personal counsellor, either by phone or e-mail, between each lesson. Lessons contain educational content combined with behaviour change strategies. Assignments in each lesson teach the participant to apply these strategies to every day life. The study population consists of employees from seven Dutch companies. The most important inclusion criteria are having a body mass index (BMI) > or = 25 kg/m2 and being an employed adult. Primary outcomes of the study are body weight and BMI, diet and physical activity. Other outcomes are: perceived health; empowerment; stage of change and self-efficacy concerning weight control, physical activity and eating habits; work performance/productivity; waist circumference, sum of skin folds, blood pressure, total blood cholesterol level and aerobic fitness. A cost-utility- and a cost-effectiveness analysis will be performed as well. Physiological outcomes are measured at baseline and after six and 24 months. Other outcomes are measured by questionnaire at baseline and after six, 12, 18 and 24 months. Statistical analyses for short term (six month) results are performed with multiple linear regression. Analyses for long term (two year) results are performed with multiple longitudinal regression. Analyses for cost-effectiveness and cost-utility are done at one and two years, using bootstrapping techniques. DISCUSSION: ALIFE@Work will make a substantial contribution to the development of cost-effective weight control- and lifestyle interventions that are applicable to and attractive for the large population at risk.

Adult↗