Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Statistical Regression”

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 1,585 records · Page 88Linked to original sources
PubMed2006

[Relationship between determining factors for depressive symptoms and for dietary habits in older adults in Mexico].

OBJECTIVE: To determine the factors that are commonly associated with both the dietary habits of older adults living in the community and depressive symptoms in this group. METHODS: Secondary analysis of data on Mexico City obtained by means of the multicenter study on Health, Well-being, and Aging (Salud, Bienestar y Envejecimiento, SABE) that was carried out in 1999 and 2000. The following variables were examined: presence or absence of depressive symptoms, according to Yesavage's Geriatric Depression Scale (GDS); mental status, as reflected by the score obtained on a modified, shortened version of Folstein's Mini-mental State Examination (MMSE); and functional capacity, as measured by Katz' Index of Independence in Activities of Daily Living. Those whose score on the shortened MMSE was 12 points or less were asked to respond to the Pfeffer Functional Activities Questionnaire, which was used to assess their ability to perform the basic activities of daily living. Information was obtained through self-report on the material, physical, psychological, and social aspects of purchasing, preparing, and consuming food products and of oral health status during the most recent 12-month period. Variables that were significantly associated with the results obtained on the GDS were included in a multivariate regression analysis; several statistical models were created, and variables that were shown to be statistically significant in the stepwise multivariate linear regression were used to determine the best-fitting explanatory model for the results obtained on the GDS. RESULTS: The average age of study participants was 64.4 +/- 8.6 years, and the prevalence of depressive symptoms was 66%. The score obtained on the GDS showed a significant association with the presence of arterial hypertension (P < 0.01), but not with the self-reported presence of diabetes, neoplasia, stroke, lung disease or heart disease. However, the use of dental prostheses (P < 0.01), urinary incontinence (P < 0.01), and falls (P < 0.01) were significantly associated with the results on the GDS. The intake of milk products, meat, fish, fowl, fruit, and vegetables was significantly lower in the group that had depressive symptoms. An inverse correlation was detected between the score obtained on the GDS on the one hand, and the number of complete meals consumed during the day (P < 0.01) and total fluid intake (P < 0.01) on the other. The determining factors that were most closely associated with these results were, in addition to the presence of arterial hypertension, the presence of cognitive impairment (P < 0.01), difficulty performing the basic activities of daily living (P = 0.03) and the instrumental activities of daily living (P < 0.01), poor mobility (P < 0.01), difficulty using the telephone (P < 0.01), and the self-perception of having poor memory (P < 0.01), of having insufficient resources with which to live (P < 0.01), and of having poor oral health (P < 0.01). These variables explained 31% of the variance seen in the R2 values linked to the SDG variables that were incorporated into the final explanatory model. CONCLUSION: A number of determining factors for depressive symptoms and the results obtained on the GDS resemble the factors that determine poor dietary habits among older adults. The potential existence of common causative mechanisms calls attention to the need for designing interventions aimed at preventing both types of problems and their negative consequences. These results confirm the need to take dietary habits and other parameters into account when studying depression in older adults.

Aged↗
PubMed1992

Covariance adjustment of survival curves based on Cox's proportional hazards regression model.

Cox's proportional hazards regression model is a useful statistical tool for the analysis of 'survival data' from longitudinal studies. This multivariate method compares the 'survival experience' between two or more exposure groups while allowing for simultaneous adjustment of confounding due to one or more covariates. In addition to the summary regression statistics, further insight on the exposure--response relationship can be gained by visually examining the covariates-adjusted survival curves in the respective comparison groups. Covariates-adjusted survival curves are usually computed by the 'average covariate method'. This method is, however, subject to potential drawbacks. A method that avoids these drawbacks is to estimate adjusted survival curves by the corrected group prognostic curves approach. We have written a computer program to construct survival curves by the latter method. The program is coded in the Interactive Matrix Language of SAS.

Humans↗
PubMed2006

Statistical analysis of test-day milk yields using random regression models for the comparison of feeding groups during the lactation period.

Random regression models are widely used in the field of animal breeding for the genetic evaluation of daily milk yields from different test days. These models are capable of handling different environmental effects on the respective test day, and they describe the characteristics of the course of the lactation period by using suitable covariates with fixed and random regression coefficients. As the numerically expensive estimation of parameters is already part of advanced computer software, modifications of random regression models will considerably grow in importance for statistical evaluations of nutrition and behaviour experiments with animals. Random regression models belong to the large class of linear mixed models. Thus, when choosing a model, or more precisely, when selecting a suitable covariance structure of the random effects, the information criteria of Akaike and Schwarz can be used. In this study, the fitting of random regression models for a statistical analysis of a feeding experiment with dairy cows is illustrated under application of the program package SAS. For each of the feeding groups, lactation curves modelled by covariates with fixed regression coefficients are estimated simultaneously. With the help of the fixed regression coefficients, differences between the groups are estimated and then tested for significance. The covariance structure of the random and subject-specific effects and the serial correlation matrix are selected by using information criteria and by estimating correlations between repeated measurements. For the verification of the selected model and the alternative models, mean values and standard deviations estimated with ordinary least square residuals are used.

Animal Feed↗
PubMed1989

Fitting heteroscedastic regression models to individual pharmacokinetic data using standard statistical software.

In the analysis of individual pharmacokinetic data by nonlinear regression it is important to allow for possible heterogeneity of variance in the response. Two common methods of doing this are weighted least squares with appropriate weights or data transformation using a suitable transform. With either approach it is appealing to let the data determine the appropriate choice of weighting scheme or transformation. This article describes two methods of doing this which are easy to compute using standard statistical software. The first method is a generalized least squares scheme for the case where the variance is assumed proportional to an unknown power of the mean. The second involves applying a power transformation to both sides of the regression equation. It is shown that both techniques may be implemented using only nonlinear regression routines. Sample code is provided for their implementation using the SAS software package. However, the proposed methods are feasible using any software package that incorporates a nonlinear least squares routine, and are thus well suited to routine use.

Humans↗
PubMed2003

A cautionary tale of two statistics: partial correlation and standardized partial regression.

Partial correlation and standardized partial regression are widely used to control for other variables in cross-sectional and longitudinal studies. Two aspects of these statistics can radically alter the interpretation of data and so deserve attention. First, their sign will be opposite to that of the original zero-order correlation when this correlation is of the same sign as and smaller than the product of the other 2 correlations. Second, their size will be considerably larger than a weak zero-order correlation when either that correlation is positive and the product of the other 2 correlations is negative or that correlation is negative and the product of the other 2 correlations is positive. The first aspect is illustrated with data from J. M. Gottman and L. J. Krokoff (1989) on initial conflict behavior and initial and later marital satisfaction, and the 2nd aspect is illustrated with data from C. L. Heavey, C. Layne, and A. Christensen (1993) on the same issue. Their interpretation of their findings is questioned.

Conflict, Psychological↗
PubMed2000

[Categorizing variables in the statistical analysis of data: consequences for interpreting the results].

Frequently during the process of data analysis in epidemiological studies, the scale of one or more continuous variables is changed. The objective of this paper was to assess the consequences of categorizing variables during data analysis. We studied three situations with different scenarios for statistical analysis with regression models. The results show that dichotomizing continuous variables can substantially modify the relationships between dependent and independent variables. Thus, for example, in epidemiological studies trying to evaluate the effect of an exposure on a response, the magnitude and/or the direction of this effect can be biased by dichotomizing a variable. We therefore recommend avoiding, as much as possible, the categorization of variables when doing analyses.

Analysis of Variance↗
PubMed1998

[Epidemiologic and statistical interaction models and consequences for regression analysis].

Different ways to define interaction between exposition factors in epidemiological studies as well as the choice between additive and multiplicative no interaction leads frequently to confusion during data analysis. In their standard form methods of event data analysis such as Poisson or logistic regression assume a multiplicative parameterization of no interaction. However, evidence from empirical investigations as well as causal models of disease etiology, e.g. the simple independent action model of Finney or the sufficient-component-causes model of Rothman, suggest additive or other kinds of non-multiplicative concepts of no interaction. For additive structured data we illustrate the asymptotic bias ("interaction-bias") of main effect estimates which are based on inappropriate data analysis using multiplicative models and omitting significant or non-significant interaction terms. We show that both the epidemiological study design as well as the underlying causal model are determinants of the interaction structure of the data and should be considered in the model selection process. The definition of interaction should distinguish between risk, rate and odds if risks are not very small. Using generalized linear models with parametrical link functions we are able to analyze non-multiplicative interaction structures.

Adult↗
PubMed2003

Estimation of left ventricular ejection fraction from Doppler derived myocardial performance index in patients with acute myocardial infarction: agreement with echocardiographic and radionuclide measurements.

OBJECTIVE: The aim of the study was to validate a previously published method to calculate left ventricular ejection fraction (EF) from the myocardial performance index (MPI or Tei-index) in patients with acute myocardial infarction (MI). METHODS: Sixty-one patients in sinus rhythm without overt heart failure were examined between 2 and 7 days after the acute MI. Doppler tracings from mitral inflow and left ventricular outflow were recorded together with two-dimensional echocardiographic (2DE) recordings. MPI was calculated from the Doppler tracings, and EF measured with the biplane Simpson's method. From MPI the EF was calculated by the formula EF = 0.60 - (0.34 x MPI). Radionuclide angiographic (RNA) measurements of EF were performed within 1 day of the Doppler echocardiography. RESULTS: Compared with radionuclide EF, MPI derived EF significantly underestimated EF by 0.03 (+/-0.013; P = 0.027), whereas there was no significant difference in mean EF between 2DE and RNA. There was no statistically significant difference in the agreement between MPI derived EF relative to RNA, or 2DE relative to RNA. The agreement between the three methods was only moderate with wide limits of agreement (+/-0.17). The relationship expressed by the proposed formula for calculating EF from MPI was not statistically significant in regression analysis in this patient population. CONCLUSIONS: No statistically significant relationship was found between MPI and EF by radionuclide angiography. However, MPI derived EF was as accurate as biplane echocardiographic measurements of EF when compared with radionuclide EF, but the agreement between methods was only moderate.

Echocardiography, Doppler↗
PubMed1985

Non-protein bound oestradiol, sex hormone binding globulin, breast cancer and breast cancer risk.

It has recently been found by various authors that despite a normal serum concentration of oestradiol (E2), the percentage of non-protein-bound or free E2 is abnormally high in breast cancer patients. Since it is the free E2 which is considered to be biologically active, confirmation of this finding would be most relevant to the pathogenesis of breast cancer. Using Hammond's centrifugal ultrafiltration dialysis method we have measured free E2 in heparinized plasma from 68 premenopausal women (a) at high familial risk of breast cancer (n = 18), (b) with benign breast disease (n = 17), (c) cured of T1N0M0 breast cancer at least 6 months previously (n = 17) and (d) normal controls matched for age, parity and Quetelet index (n = 16). Sex hormone binding globulin (SHBG) was measured as [3H]-dihydrotestosterone binding capacity. Free E2 and SHBG were also measured in the serum of (e) postmenopausal patients having breast cancer (n = 38) and (f) matched control cancer patients (n = 67). We confirmed a very good inverse correlation between log free E2 per cent and log SHBG (P less than 0.0001). The regression lines for groups (a)-(d) were not statistically different. The regression lines for groups (e) and (f) were identical and ran nearly parallel to those for groups (a)-(d) though somewhat lower. This small difference may be ascribed to menopausal status. Therefore, we found no difference in free E2 percentage, calculated free E2 concentration or SHBG between premenopausal women at risk, women with benign breast disease, patients cured for early breast cancer or having breast cancer and matched controls. However, postmenopausal breast cancer patients had a significantly higher total serum E2 concentration and, by consequence a higher calculated free E2 concentration compared to the carefully matched control group.

Adult↗
PubMed1987

A comparison of five anthropometric indices for identifying factors of malnutrition.

Five anthropometric indices--weight-for-age, height-for-age, weight-for-height, weight velocity, and height velocity--are compared for identifying factors of nutritional status of children. Data come from Matlab, the field station of the International Centre for Diarrhoeal Disease Research, Bangladesh. Weights and heights on approximately 1,400 children aged 12-60 months were taken 11 times at two-month intervals. Dwelling floor space, a proxy for socioeconomic status, is considered a long-term factor; the 1974-1975 Bangladesh famine, a medium-term factor; and season, a short-term factor of nutritional status. The figures of the indices by socioeconomic status at different times show their power to identify the factors visually; the regression analyses test statistical significance of the factors; and the standardized regression coefficients provide relative power of the indices to identify the factors. Weight-for-age and height-for-age appear best for identifying long- and medium-term factors; weight velocity, for identifying short-term factors. Weight-for-height is third in identifying long- and medium-term factors, second for identifying short-term factors. Height velocity identifies short-term factors only. The effect of short-term factors on height velocity appears approximately four months later than it appears on weight velocity. This study clarifies why certain indices do not predict mortality in this study population and others.

Anthropometry↗
PubMedDate not supplied

Impaired processing of famous faces in Alzheimer's disease is related to neurofibrillary tangle densities in the prefrontal and anterior cingulate cortex.

To examine the neuroanatomical correlates of impaired processing of famous faces in Alzheimer's disease (AD), we performed an anterograde clinicopathological study of 25 patients with clinically and neuropathologically confirmed AD. Famous face recognition, identification and naming was assessed using the Famous Face Test. The assessment of neurofibrillary tangle (NFT) and senile plaque (SP) densities was performed in ten cortical areas in both hemispheres, and statistical analysis was made using forward stepwise logistic regression models. A statistically significant relationship was found between NFT densities in Brodmann's areas 9 and 24 in both hemispheres and impaired famous face naming and identification. SP counts did not correlate with any of the neuropsychological parameters. These data suggest that NFT formation in prefrontal and anterior cingulate cortex, two areas involved in semantic memory processes, is a key event in famous face naming and identification deficits. In agreement with previous studies, they also indicate that SP densities are not a good pathological correlate of neuropsychological deficits in AD.

Aged↗
PubMed2005

On theoretical principles of constitution typology.

Constitution typology is thought of as a classification that is based on a few fundamental properties of the organism. These properties steadily distinguish individuals and permit the prediction of manifold secondary (external) features of types. At the transition from one constitutional type to another, concordant changes of trait values are determined by the respective changes of general causal factors. Theoretically, external constitution typology should be constructed on the basis of trait concordant "dynamics" under the change of fundamental factors at the individual level. Therefore, it does not depend on statistical relations characterizing population level. Causal factors and their influences on traits are impossible to be correctly and uniquely determined on the basis of population statistics relations. Causal factors do not coincide with the statistical ones of multivariate biometric analysis; equally, trait regression dependencies on statistical factors do not coincide with the causal influences.

Biometry↗
PubMed2005

Inverse correlation between soluble CD40 ligand and soluble CD40 is absent in patients with unstable angina.

The CD40/CD40 ligand (CD40L) system mediates inflammatory processes important in atherogenesis and plaque instability. The expression of CD40L on activated T cells was suppressed by soluble CD40 (sCD40) in vitro. However, the relationship between soluble CD40L (sCD40L) and sCD40 in unstable angina (UA) is still unknown. Thirty-seven consecutive patients with recent chest pain or discomfort were recruited. Patients with both Braunwald's class IB-IIIB and with coronary stenosis (or stenoses) of >75% were assigned to the UA group (n = 19, aged 67.2 +/- 8.2 years), and the rest to the control group (n = 18, aged 63.4 +/- 8.7 years). The serum levels of sCD40L and sCD40, and the plasma levels of matrix metalloproteinase (MMP)-9, were measured by enzyme-linked immunosorbent assays. A significantly inverse correlation between sCD40L and sCD40 was shown in the controls (r = -0.72, P = 0.0007), but was absent in the UA group (r = -0.16, P not significant), although there was no statistical significance between these groups in terms of serum levels of sCD40L or sCD40. The difference of the regression slopes of these regression lines was statistically significant (P < 0.01). Additionally, there was a significant correlation between sCD40 and plasma levels of MMP-9 in the patients with and without UA (r = 0.58, P = 0.0096), but no significant correlation between sCD40L and MMP-9 levels (r = 0.00, P not significant). The balance between CD40 and CD40L may be lost in patients with UA. Soluble CD40 expression may also be related to MMP-9 expression in atherosclerotic tissues.

Aged↗
PubMed2004

Gestational trophoblastic disease: is intensive follow up essential in all women?

OBJECTIVE: To determine the timescale of the registration process for gestational trophoblastic disease and its impact on hCG level at registration and subsequent need for chemotherapy. DESIGN: A prospective observational study using a standardised protocol for registration, assessment and treatment for molar pregnancy. SETTING: A supra-regional tertiary referral centre for gestational trophoblastic disease. PARTICIPANTS: A total of 2046 consecutive women registered between January 1994 and December 1998 with a diagnosis of molar pregnancy. METHODS: Data at and after registration, collected prospectively on a computerised database, were statistically analysed (by multiple logistic regression and ANOVA). MAIN OUTCOME MEASURES: Relationship between length of time to and hCG value at registration; also the subsequent need for chemotherapy. RESULTS: A total of 2046 women with a diagnosis of molar pregnancy were registered in the study period. The mean time interval between first evacuation and registration at the referral centre was 47 days (median 37, range 0-594). One hundred and five out of 2046 (5.1%) women needed chemotherapy. Sixty-three precent of the women (1296 out of 2046) had a normal level of urinary hCG (less than 40 IU/24 hours) at the time of registration and only one (0.08%) needed chemotherapy. Binary logistic regression analysis showed a statistically significant relationship between time to registration, hCG value, histology, pretreatment risk score and decision to administer chemotherapy. CONCLUSION: Women with gestational trophoblastic disease who were registered late were significantly more likely to have normal levels of hCG and were less likely to need chemotherapy. A less intensive follow up may be justified in women with gestational trophoblastic disease who are registered with a normal hCG level.

Adult↗
PubMed1990

Neo-osteogenesis of haversian trabeculae through a bone chamber implanted in a rabbit tibial cortex: a control model.

Neo-osteogenesis of cortical bone trabecular was observed as they regenerated into a bone chamber implant by appositional growth. Measures of change in bone area were obtained from 13 rabbits each week starting the third and ending the eighth week postimplantation. Observations were made using intravital microscopy and were analyzed using digital image processing. Images were computer-captured video frames equivalent to 78 (6 x 13) separate observations. They were measured by tracing the trabecular outlines with a digitizing crosshair each week and comparing changes in area as a percent of the circular field-of-view ("slit-gap") filled. Data supported the hypothesis that trabecular regeneration week 3 to week 8 followed the logistic growth curve regression: A% 100%/1 + 9.47e-0.7747(t-3) where A% is the percent of slit-gap area covered by bone and t is time, at a very high significance level. Nevertheless, a highly significant linear regression fit the data. Statistical analysis showed that the regression line could be fit to bone area measurements from weeks 3 to 8 (W3-W8) postimplantation, giving a constant neo-osteogenesis rate of 7.42 +/- 0.67 X 10(4) microns 2/day and a decreasing linear neo-osteogenesis rate from 73 microns/day at W4 to 21 microns/day at W8; the latter is based on a circle-segment approximation of trabeculum shape. This range approximated a bridge between ranges for cortical gap defect healing and porous ingrowth healing reported by other workers and supported the hypothesis that the BCI control model was a cross between gap healing and porous ingrowth.

Air Pressure↗
PubMed1997

Using regression models for prediction: shrinkage and regression to the mean.

The use of a fitted regression model in predicting future cases, either as a diagnostic tool or as an instrument for risk assessment is discussed. The regression to the mean effect implies that the future values of the response variable tend to be closer to the overall mean than might be expected from the predicted values. The extent of this shrinkage is studied for multiple and logistic regression models, and is found to be related to simple goodness-of-fit statistics of the original regression. Shrinkage is a particularly serious problem if the sample size is small and/or the number of covariates is large. Shrinkage of predictors is illustrated by two examples. A more general formulation is suggested.

Biometry↗
PubMed1994

MIAMOD: a computer package to estimate chronic disease morbidity using mortality and survival data.

The MIAMOD package is for use in epidemiological analysis of chronic degenerative diseases. The package has been designed to use survival and mortality data supplied by the user, to estimate incidence and prevalence rates, to forecast future mortality and morbidity levels and to provide life tables. Estimates are derived by modelling incidence as an age-period-cohort function and by the equations relating specific cause mortality and prevalence to incidence and survival. The model parameters are evaluated by fitting mortality data by means of a weighted non-linear regression model. Numerical and statistical methods used to solve the regression and to compute standard errors, projections, and life tables are discussed. The program is written in standard FORTRAN 77 and runs on a mainframe computer. The different program units are illustrated together with the principal implementation options. An example of application is provided to show a typical output pattern.

Breast Neoplasms↗