SEARCH · PubMed Health

Results for “Logistic Models”

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 649 records · Page 36Linked to original sources

PubMed2004

Analyzing psychopathology items: a case for nonparametric item response theory modeling.

The authors discuss the applicability of nonparametric item response theory (IRT) models to the construction and psychometric analysis of personality and psychopathology scales, and they contrast these models with parametric IRT models. They describe the fit of nonparametric IRT to the Depression content scale of the Minnesota Multiphasic Personality Inventory--2 (J. N. Butcher, W. G. Dahlstrom, J. R. Graham, A. Tellegen, & B. Kaemmer, 1989). They also show how nonparametric IRT models can easily be applied and how misleading results from parametric IRT models can be avoided. They recommend the use of nonparametric IRT modeling prior to using parametric logistic models when investigating personality data.

Adult↗

PubMed2001

Application of statistical models for secondary data usage of the US Navy's Occupational Exposure Database (NOED).

Many organizations around the world have collected data related to individual worker exposures that are used to determine compliance with workplace standards. These data are often warehoused and thereafter rarely used as an information resource. Using appropriate groupings and analysis of OSHA data, Gómez showed that such stored data can provide additional insight on factors affecting occupational exposures. Using data from the Occupational Exposure Database of the United States Navy, the usefulness of statistical models for defining probabilities of exposure above permissible limits for observed work conditions is examined. Analyses have highlighted worker Similar Exposure Groups (SEGs) with potential for overexposure to asbestos and lead. In terms of grouping data, Rappaport et al. defined the Within-Between Lognormal Model, a scale-independent measure for quantifying between-worker variability within a selected worker group: (B)R.95 = exp[3.92s(sB)], representing the ratio of arithmetic mean exposures received by workers in the 97.5th and 2.5th percentiles. To help search for groups, the Proportional Odds Model, a generalization of the logistic model to ordinal data, can predict probabilities for group exposure above the Occupational Exposure Limit (OEL), or the Action Level (AL), which is one-half of the OEL. Worker SEGs have been identified for asbestos workers removing friable asbestos ((B)R.95 = 11.0) and nonfriable asbestos ((B)R.95 = 6.5); metal cleaning workers sandingspecialized equipment ((B)R.95 = 11.3), and workers at target shooting ranges cleaning up lead debris ((B)R.95 = 10). Estimated probabilities for the categories OEL support current understanding of work processes examined. Differences in probability noted between tasks and levels of ventilation validate this method for evaluating other available workplace exposure determinants, and for predicting probability of membership in categories that may help further define worker exposure groups, and determinants of excessive exposures. Thus, analyses of retrospective exposure data can help identify work site and work practice factors for efficient targeting of remediation resources.

Databases, Factual↗

PubMed2003

Penalized estimating equations.

Penalty models--such as the ridge estimator, the Stein estimator, the bridge estimator, and the Lasso-have been proposed to deal with collinearity in regressions. The Lasso, for instance, has been applied to linear models, logistic regressions, Cox proportional hazard models, and neural networks. This article considers the bridge penalty model with penalty sigma(j)/beta(j)/gamma for estimating equations in general and applies this penalty model to the generalized estimating equations (GEE) in longitudinal studies. The lack of joint likelihood in the GEE is overcome by the penalized estimating equations, in which no joint likelihood is required. The asymptotic results for the penalty estimator are provided. It is demonstrated, with a simulation and an application, that the penalized GEE potentially improves the performance of the GEE estimator, and enjoys the same properties as linear penalty models.

Air Pollution↗

PubMed2006

Predicting sustained virological response and anaemia in chronic hepatitis C patients treated with peginterferon alfa-2a (40KD) plus ribavirin.

AIM: To assess the likelihood of a sustained virological response (SVR) vs. the likelihood of anaemia in patients with chronic hepatitis C. METHODS: Data from 1732 patients treated with peginterferon alfa-2a (40KD) plus ribavirin in two randomized, multinational studies were pooled. Probabilities of SVR and anaemia were modelled using the generalized additive logistic model, with numerous clinical variables considered for entry into the model. Baseline haemoglobin was only considered in the analysis for anaemia. RESULTS: The probability of anaemia increased from 6 to 16% as a function of the ribavirin dose kg(-1) (12-16 mg kg(-1)), whereas the relationship between SVR and ribavirin dose kg(-1) was influenced by hepatitis C virus (HCV) genotype. The probability of an SVR was not influenced by the ribavirin dose kg(-1) in patients with HCV genotype 2 or 3 infection, but increased as a function of ribavirin dose kg(-1) in patients with HCV genotype 1 infection (40-50% increase in probability of SVR for 12-16 mg kg(-1) dose ribavirin increase). The probability of an SVR in patients included with HCV genotype 1 decreased with increasing HCV RNA level to about 3 million copies ml(-1), but was relatively independent of increasing HCV RNA level thereafter. In addition, older age, a higher ribavirin apparent oral clearance and cirrhosis had a negative impact on achieving an SVR, but improved with increasing alanine aminotransferase (ALT) quotient. Sex and ribavirin dose kg(-1) were the most important prognostic factors for anaemia, followed by baseline haemoglobin, age, baseline ALT quotient and cirrhosis. CONCLUSION: This study supports individualizing ribavirin dosages by HCV genotype and body weight, and highlights several clinical variables that influence the likelihood of an SVR compared with anaemia in chronic hepatitis C patients treated with peginterferon alfa-2a (40KD) plus ribavirin.

Adult↗

PubMed1993

Pharmacodynamics of intravenous ciprofloxacin in seriously ill patients.

Seventy-four acutely ill patients were treated with intravenous ciprofloxacin at dosages ranging between 200 mg every 12 h and 400 mg every 8 h. A population pharmacokinetic-pharmacodynamic analysis relating drug exposure (and other factors) to infectious outcome was performed. Plasma samples were obtained and assayed for ciprofloxacin by high-performance liquid chromatography. Samples from patients were frequently cultured so that the day of bacterial eradication could be determined. The pharmacokinetic data were fitted by iterative two-stage analysis, assuming a linear two-compartment model. Logistic regression was used to model ciprofloxacin exposure (and other potential covariates) versus the probabilities of achieving clinical and microbiologic cures. The same variables were also modelled versus the time to bacterial eradication by proportional hazards regression. The independent variables considered were dose, site of infection, infecting organism and the MIC for it, percent time above the MIC, peak, peak/MIC ratio, trough, trough/MIC ratio, 24-h area under the concentration-time curve (AUC), AUC/MIC ratio (AUIC), presence of other active antibacterial agents, and patient characteristics. The most important predictor for all three measures of ciprofloxacin pharmacodynamics was the AUIC. A 24-h AUIC of 125 SIT-1.h (inverse serum inhibitory titer integrated over time) was found to be a significant breakpoint for probabilities of both clinical and microbiologic cures. At an AUIC below 125 (19 patients), the percent probabilities of clinical and microbiologic cures were 42 and 26%, respectively. At an AUIC above 125 (45 patients), the probabilities were 80% (P < 0.005) and 82% (P < 0.001), respectively. There were two significant breakpoints in the time-to-bacterial-eradication data. At an AUIC below 125 (21 patients), the median time to eradication exceeded 32 days; at an AUIC of 125 to 250 (15 patients), time to eradication was 6.6 days: and at AUIC above 250 (28 patients), the median time to eradication was 1.9 days (groups differed; P < 0.005). These findings, when combined with pharmacokinetic data reported in the companion article, provide the rationale and tools needed for targeting the dosage of intravenous ciprofloxacin to individual patients' pharmacokinetics and their bacterial pathogens' susceptibilities. An a priori dosing algorithm (based on MIC, patient creatine clearance and weight, and the clinician-specified AUIC target) was developed. This approach was shown, retrospectively, to be more precise than current guidelines, and it can be used to achieve more rapid bacteriologic and clinical responses to ciprofloxacin, as a consequence of targeting the AUIC.

Adult↗

PubMed1986

Prognostic factors in malignant melanoma of the choroid: a retrospective survey of cases occurring in Northern Ireland between 1965 and 1980.

In Northern Ireland between 1965 and February 1980 one hundred eyes were enucleated following diagnosis of malignant melanoma of the choroid. The pathology of 87 cases was available for re-examination. Data on those patients still alive were examined together with the post-mortem reports of those deceased. A mathematical model was constructed to assess tumour volume more accurately. Survival was analysed using univariate and multivariate methods. Actuarial survival curves were constructed and two multivariate regression models were used--the proportional hazards model and the log-logistic model. In addition person years at risk analysis was used to compare survival with the Northern Ireland population. When the data were examined univariately, the location of the tumour, cell type, degree of invasion and the parameters of size (height, diameter and volume) were all significant prognostic indicators. When multivariate analysis was employed, the degree of invasion did not contribute significantly to the model, this being accommodated by tumour size. Cell type was also shown to be of less importance in the multivariate models. The various prognostic indicators together with the person years at risk analysis are discussed.

Adolescent↗

PubMedDate not supplied

Polychotomous multivariate models for coronary heart disease simulation. II. Comparisons of risk functions.

This is the second in a series of papers dealing with models of coronary heart disease. Three different types of statistical models are considered as risk functions: the multivariate logistic model, the Cox proportional hazard model and the Neyman exponential risk avoidance model. The types of models differ in the form hypothesized for the probability of occurrence of coronary heart disease outcomes: incident myocardial infarct, cardiac death, and death from other causes. Although the three risk functions are strikingly different, they can all be tested using the CRISPERS chronic disease simulation system. Simulations were performed using data from North Karelia, Finland. The polychotomous multivariate logistic risk function is convenient for studies involving increasing numbers of risk factors. The Cox proportional hazard regression model is shown to be unsuitable for the cohort dataset used as well as for some of the intended uses of the simulation models. The Neyman exponential risk avoidance model involves time in a quite different fashion. It has the inherent advantage of being easier to relate to underlying biological mechanisms because it is the integral of first order rate equations. It is concluded that more than one risk function should be evaluated for simulations of coronary heart disease.

Adult↗

PubMed1996

Race, socioeconomic status, and obesity in 9- to 10-year-old girls: the NHLBI Growth and Health Study.

The purpose of this investigation was to determine whether measures of socioeconomic status (SES) are inversely associated with obesity in 9- to 10-year-old black and white girls and their parents. Subjects were participants in the Growth and Health Study (NGHS) of the National Heart, Lung, and Blood Institute. Extensive SES, anthropometric, and dietary data were collected at baseline on 2379 NGHS participants. The prevalence of obesity was examined in the NGHS girls and parents in relation to SES and selected environmental factors. Less obesity was observed at higher levels of household income and parental education in white girls but not in black girls. Among the mothers of the NGHS participants who were seen, lower prevalence of obesity was observed with higher levels of income and education for white mothers, but no consistent patterns were seen in black mothers. Univariate logistic models indicated that the prevalence of obesity was significantly and inversely associated with parental income and education and number of parents in the household in white girls whereas caloric intake and TV viewing were significantly and positively associated with obesity. Among black girls, only TV viewing was significantly and positively associated with the prevalence of obesity. Multivariate logistic regression models revealed that lower parental educational attainment, one-parent household, and increased caloric intake were significantly associated with the prevalence of obesity in white girls; for black girls, only increased hours of TV viewing were significant in these models. It is concluded that socioeconomic status, as measured by education and income, was related to the prevalence of obesity in girls, with racial variation in these associations. A lower prevalence of obesity was seen at higher levels of socioeconomic status in white girls, whereas no clear relationship was detected in black girls. These findings raise new questions regarding the correlates of obesity in black girls.

Black People↗

PubMed1998

Prevalence proportion ratios: estimation and hypothesis testing.

BACKGROUND: Recent communications have argued that often it may not be appropriate to analyse cross-sectional studies of prevalent outcomes with logistic regression models. The purpose of this communication is to compare three methods that have been proposed for application to cross sectional studies: (1) a multiplicative generalized linear model, which we will call the log-binomial model, (2) a method based on logistic regression and robust estimation of standard errors, which we will call the GEE-logistic model, and (3) a Cox regression model. METHODS: Five sets of simulations representing fourteen separate simulation conditions were used to test the performance of the methods. RESULTS: All three models produced point estimates close to the true parameter, i.e. the estimators of the parameter associated with exposure had negligible bias. The Cox regression produced standard errors that were too large, especially when the prevalence of the disease was high, whereas the log-binomial model and the GEE-logistic model had the correct type I error probabilities. It was shown by example that the GEE-logistic model could produce prevalences greater than one, whereas it was proven that this could not happen with the log-binomial model. The log-binomial model should be preferred.

Cross-Sectional Studies↗

PubMed2005

Comparison of different mathematical models to analyze diminution kinetics of ultrasound contrast enhancement in a flow phantom.

Ultrasound (US) energy leads to intensity- and frequency-dependent destruction of US contrast agent (UCA) microbubbles. When applying repeated US pulses, this phenomenon can be detected as contrast diminution over time. Contrast diminution kinetics depend on the replenishment of UCA into the sample volume. Thus, it is related to organ perfusion. To analyze the contrast diminution kinetics following pulsed harmonic US application (SONOS 5500, 1.8-3.6 MHz, MI: 1.6, frame rates: 2, 4, and 6.67 Hz), we performed an in vitro study using SonoVue continuous infusion. Seven flow rates (4.5, 9, 13.5, 18, 22.5, 27 and 36 mL/min) were tested. Based on our results, three mathematical models (linear intensity decrease, exponential decay, and an exponential destruction/reperfusion model) describing diminution kinetics were compared. In 113 (89.7%) of 126 trials, a signal decrease was observed after US application. At higher flow rates (18 to 36 mL/min), curve fitting was not possible for the exponential models. For the linear model, intensity decrease depended significantly on the flow rate (p < or = 0.005, n = 7). A logistic model was fitted to the data, defining the slope in the dynamic range of quasilinear dependence for the different frame rates, as well as the inflection point: The higher the frame rate, the higher the flow rate at the point of inflection. For the exponential model, the contrast half-life was dependent on the flow rate (r = 0.95, p = 0.03, n = 6) only at the highest frame rate (6.67 Hz). The perfusion coefficient derived from the destruction/reperfusion model was not significantly related to the flow rate. In conclusion, the linear intensity decrease correlates well with the flow rate (i.e., flow velocity) and defines optimum frame rates for diminution imaging at different flow velocities. The exponential models, which required curve-fitting procedures, were determined to be inappropriate to describe flow in our phantom.

Contrast Media↗

PubMed1992

Determinants of child mortality in south-west Uganda.

Anthropometric and sociodemographic variables were taken from 4320 children in a baseline survey carried out in March-April 1988 in the district of Mbarara, south-west Uganda. After 12 months a follow-up survey assessed the mortality of the children during the preceding year. Lack of ownership of cattle, recent arrival in the village, using candles for lighting, being of birth order higher than 5 and having a father with less than 8 years of schooling were significantly associated with child mortality. The addition of mid-upper arm circumference significantly improved the logistic model of socioeconomic variables and mortality and did not diminish the predictive power of socioeconomic variables in relation to increased mortality. This suggests that nutritional status and specific socioeconomic factors are both, independently, important predictors of child mortality.

Anthropometry↗

PubMed2004

On the mechanistic underpinning of discrete-time population models with complex dynamics.

We present a mechanistic underpinning for various discrete-time population models that can produce limit cycles and chaotic dynamics. Specific examples include the discrete-time logistic model and the Hassell model, which for a long time eluded convincing mechanistic interpretations, and also the Ricker- and Beverton-Holt models. We first formulate a continuous-time resource consumption model for the dynamics within a year, and from that we derive a discrete-time model for the between-year dynamics. Without influx of resources from the outside into the system, the resulting between-year dynamics is always overcompensating and hence may produce complex dynamics as well as extinction in finite time. We recover a connection between various standard types of continuous-time models for the resource dynamics within a year on the one hand and various standard types of discrete-time models for the population dynamics between years on the other. The model readily generalizes to several resource and consumer species as well as to more than two trophic levels for the within-year dynamics.

Animals↗

PubMed2003

Reproductive factors and familial predisposition for breast cancer by age 50 years. A case-control-family study for assessing main effects and possible gene-environment interaction.

BACKGROUND: The effect of environmental/lifestyle factors on breast cancer risk may be modified by genetic predisposition. METHODS: In a population-based case-control-family study performed in Germany including 706 cases by age 50 years, 1381 population, and 252 sister controls, we investigated main effects for environmental/lifestyle factors and genetic susceptibility and gene-environment interaction (G x E). Different surrogate measures for genetic predisposition using pedigree information were used: first-degree family history of breast or ovarian cancer; and gene carrier probability using a genetic model based on rare dominant genes. Possible G x E interaction was studied by (1) logistic regression using cases and population controls including an interaction term; (2) comparing results using sister controls and population controls; (3) case-only analysis with logistic regression and (4) a mixture logistic model. RESULTS: Familial predisposition showed the strongest main effect and the estimated gene carrier probability gave the best fit. High parity and longer duration of breastfeeding reduced breast cancer risk significantly, a history of abortions increased risk and age at menarche showed no significant effect. We found significant G x E interaction between parity and genetic susceptibility using different surrogate measures. In women most likely to have a high genetic susceptibility, high parity was less protective. Later age at menarche was protective in women with a positive family history. No evidence for G x E interaction was found for breastfeeding and abortion. CONCLUSIONS: These findings corroborate results from other studies and provide further evidence that the magnitude of protection from parity is reduced in women most likely to have a genetic risk in spite of the limitations of using surrogate genetic measures.

Abortion, Therapeutic↗

PubMed2002

Exploratory causal modeling in epidemiology: are all factors created equal?

The purpose of this study was to demonstrate the consequences of analyzing sequentially caused relationships with models assuming equally proximate causation. Monte Carlo simulations of data with well defined causations were performed. The logistic modeling approach was strongly misleading if a distant causal factor was treated as a factor being equally distant to the outcome as a proximal causal factor. In contrast, simple pathway analysis was able to correctly identify the true causation. In causal pathways, the relative risk of an intermediate cause with respect to the outcome needs to have a certain magnitude for the effect of the distant variable to be passed on. The results further show that the true relative risk of the distant variable is not dependent on its baseline prevalence. In contrast, the prevalence of the intermediate variable must be small enough to carry the influence of the distant variable through the causal chain. Practical epidemiologic exploration of etiological factors is presently dominated by stepwise multiple regression. This type of exploration is not model free but is often intuitively based on the structural assumption of equal proximity of all potential factors to the outcome. Equal proximity, however, is not likely in many etiologies, especially not if the causal factors under consideration are of different quality, like psychological and biological factors. In cases of causal pathways with some factors more distant and others more proximal to the outcome, the former tend to be dismissed by equal proximity modeling. Upstream exploration of more distant etiological factors is hindered by endemic stepwise multiple regression modeling, treating all variables as being equal in proximity to the outcome.

Confounding Factors, Epidemiologic↗

PubMed2005

Modeling anti-KLH ELISA data using two-stage and mixed effects models in support of immunotoxicological studies.

During preclinical drug development, the immune system is specifically evaluated after prolonged treatment with drug candidates, because the immune system may be an important target system. The response of antibodies against a T-cell-dependent antigen is recommenced by the FDA and EMEA for the evaluation of immunosuppression/enhancement. For that reason, we developed a semiquantitative enzyme-linked immunosorbent assay to measure antibodies against keyhole limpet hemocyanin. To our knowledge, the analysis of this kind of data is at this moment not yet fully explored. In this article, we describe two approaches for modeling immunotoxic data using nonlinear models. The first is a two-stage model in which we fit an individual nonlinear model for each animal in the first stage, and the second stage consists of testing possible treatment effects using the individual maximum likelihood estimates obtained in the first stage. In the second approach, the inference about treatment effects is based on a nonlinear mixed model, which accounts for heterogeneity between animals. In both approaches, we use a three-parameter logistic model for the mean structure.

Analysis of Variance↗

PubMed2005

Appropriate assessment of neighborhood effects on individual health: integrating random and fixed effects in multilevel logistic regression.

The logistic regression model is frequently used in epidemiologic studies, yielding odds ratio or relative risk interpretations. Inspired by the theory of linear normal models, the logistic regression model has been extended to allow for correlated responses by introducing random effects. However, the model does not inherit the interpretational features of the normal model. In this paper, the authors argue that the existing measures are unsatisfactory (and some of them are even improper) when quantifying results from multilevel logistic regression analyses. The authors suggest a measure of heterogeneity, the median odds ratio, that quantifies cluster heterogeneity and facilitates a direct comparison between covariate effects and the magnitude of heterogeneity in terms of well-known odds ratios. Quantifying cluster-level covariates in a meaningful way is a challenge in multilevel logistic regression. For this purpose, the authors propose an odds ratio measure, the interval odds ratio, that takes these difficulties into account. The authors demonstrate the two measures by investigating heterogeneity between neighborhoods and effects of neighborhood-level covariates in two examples--public physician visits and ischemic heart disease hospitalizations--using 1999 data on 11,312 men aged 45-85 years in Malmo, Sweden.

Aged↗

PubMed2004

Adjust quality scores from alignment and improve sequencing accuracy.

In shotgun sequencing, statistical reconstruction of a consensus from alignment requires a model of measurement error. Churchill and Waterman proposed one such model and an expectation-maximization (EM) algorithm to estimate sequencing error rates for each assembly matrix. Ewing and Green defined Phred quality scores for base-calling from sequencing traces by training a model on a large amount of data. However, sample preparations and sequencing machines may work under different conditions in practice and therefore quality scores need to be adjusted. Moreover, the information given by quality scores is incomplete in the sense that they do not describe error patterns. We observe that each nucleotide base has its specific error pattern that varies across the range of quality values. We develop models of measurement error for shotgun sequencing by combining the two perspectives above. We propose a logistic model taking quality scores as covariates. The model is trained by a procedure combining an EM algorithm and model selection techniques. The training results in calibration of quality values and leads to a more accurate construction of consensus. Besides Phred scores obtained from ABI sequencers, we apply the same technique to calibrate quality values that come along with Beckman sequencers.

Algorithms↗

PubMed1987

The effect of agricultural price-support policies on interregional and rural-to-urban migration in Korea: 1976-1980.

The impact of rice price-support policies, designed to increase farmers' income and reduce rural-urban migration in the Republic of Korea, is examined for the period 1976-1980 using a polytomous logistic model. "Our findings revealed that the elasticity of migration with respect to rice yield per origin farm household is positive and is significantly different from zero. The elasticities of migration with respect to rate of urbanization, particularly urban concentration-agglomeration, and population size of the destination are also positive and are significantly different from zero.... Our findings questioned the wisdom of employing rice price price-support programs as a viable policy for reducing interregional and rural-to-urban migration in Korea."

Agriculture↗