SEARCH · PubMed Health
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.
Rapid accurate analysis of echocardiographic data is accomplished using a sonic digitizer and programmable calculator. This method allows the echocardiographer to select technically optimal areas of the recording for analysis. The resolution of the measuring device is 0.1 mm. A hardcopy printout of both measurement and calculation is provided. Instead of expensive on-line computer, an inexpensive programmable calculator is used.
MAD is a system that helps an expert data analyst in a specific application domain (like epidemiology or image analysis) to build reasoning models aimed at fulfilling specific tasks. These models may be subsequently used to guide doctors in the analysis of a set of data referring to a specific ground domain. Expert knowledge is represented at various levels: a general description of an application domain and various models that formalize the reasoning followed to perform specific tasks within a defined application domain. Reasoning models are represented as rules of propositional calculus, and a meta-knowledge permits to support knowledge acquisition. During the consultation, different external programs may be run when needed, without the doctor having to learn how to use them. MAD is written in Golden Common LISP and may be linked to any external software for data analysis, provided it runs under MS-DOS and does not require more than 192 Kb. Examples of application of the system to epidemiology and image analysis are given.
OBJECTIVE: To conduct a pooled data analysis in a group of patients defined by sex, menopausal status, and underlying disease in order to examine the effect of intermittent cyclical etidronate in the prevention and treatment of corticosteroid induced osteoporosis. METHODS: We selected 5 randomized, placebo controlled studies that examined the efficacy of intermittent cyclical etidronate therapy in which the raw data were available for analysis. Three were prevention studies and 2 treatment studies. The primary outcome was the difference between treatment groups in the percentage change from baseline in lumbar spine bone density. Secondary outcomes included the difference between treatment groups in the percentage change from baseline in femoral neck and trochanter bone density, and vertebral fracture rates. RESULTS: Results are separately pooled for the prevention and treatment studies. The prevention studies had significant mean differences (95% CI) between groups in mean percentage change from baseline in lumbar spine, femoral neck, and trochanter bone density of 3.7 (2.6 to 4.7), 1.7 (0.4 to 2.9), and 2.8% (1.3 to 4.2) after one year of treatment, in favor of the etidronate group. The treatment studies displayed a mean difference between groups in mean percentage change from baseline in lumbar spine bone density of 4.8 (2.7 to 6.9) and 5.4% (2.5 to 8.4) after one and 2 years of therapy. In the prevention studies, a reduced fracture incidence was observed in the etidronate group compared with the placebo group (relative risk 0.50; CI 0.21 to 1.19). CONCLUSION: Etidronate therapy was effective in preventing bone loss in the prevention studies and in preventing or slightly increasing bone mass in the treatment studies. A fracture benefit was observed in postmenopausal women treated with etidronate in the prevention studies.
To determine the relative incidence of recurrent carotid stenosis (RCS) and the effect of methodology on data analysis and interpretation, late results were obtained for 232 patients (270 procedures) from 1 to 51 months (mean 22 months) after carotid endarterectomy (group A). Patency of the carotid artery was confirmed by postoperative intravenous digital subtraction angiography (DSA) for most of the series, and a subset (subgroup A1) of 113 patients (129 procedures) also received DSA studies at later intervals of 4 to 49 months (mean 26 months). There were 23 late deaths and five late strokes. Only two of the strokes were ipsilateral to previous endarterectomy, and both of these patients had normal follow-up DSA studies. Late DSA imaging revealed either no RCS or only trivial defects (20% diameter or less) in 111 arteries, moderate (36% to 60%) RCS in nine, severe (70% to 90%) RCS requiring secondary procedures in eight, and internal carotid occlusion in one. Depending on the definition of RCS (secondary operation vs greater than or equal to 30% angiographic lesions), the cohort selected for analysis (group A vs subgroup A1), and the approach to calculations (crude vs cumulative), the incidence of recurrent stenosis after carotid reconstruction in this single study could be expressed within the extraordinary wide range of 3% to 32%. Although carotid endarterectomy was associated with uniformly low risk for late stroke, these results confirm that the reported recurrence rate may be substantially influenced by the method in which data are grouped and manipulated. Consistently presented data are essential to any comparisons concerning the surgical therapy for extracranial disease.
In this paper we introduce a mixture model in which we combine logistic regression and piecewise proportional hazards models for analysis of duration data. The model allows simultaneous estimation of two sets of effects of covariates: one of the probability of an event and the other of the timing of the event. We illustrate the application of the model through an analysis of the effects of women's characteristics and of the acceptance of a one-child certificate on the birth of second children in China. Both factors affect the probability of having a second child, but only the acceptance of a one-child certificate has a significant and strong effect on the second-birth interval.
This paper describes methods of constrained and restrained refinement of EXAFS data which provide a means of substantially reducing the number of independent parameters compared to conventional least-squares methods commonly used. Constrained refinement allows a major reduction in the number of free parameters for a refinement of a structural model. In restrained refinement, additional structural information from well-characterized small molecules is used to provide additional observations in the data analysis. Even though these methods are of general application to the majority of complex systems, they are particularly valuable for biological molecules. The methods are of major advantage for ligands where significant multiple scattering is present, e.g., histidine, tyrosine, CO, CN, etc. The bases of these methods are described, and applications to some complex chemical and biological systems are given.
This research examined the applicability of using a neural network approach to analyze population pharmacokinetic data. Such data were collected retrospectively from pediatric patients who had received tobramycin for the treatment of bacterial infection. The information collected included patient-related demographic variables (age, weight, gender, and other underlying illness), the individual's dosing regimens (dose and dosing interval), time of blood drawn, and the resulting tobramycin concentration. Neural networks were trained with this information to capture the relationships between the plasma tobramycin levels and the following factors: patient-related demographic factors, dosing regimens, and time of blood drawn. The data were also analyzed using a standard population pharmacokinetic modeling program, NON-MEM. The observed vs predicted concentration relationships obtained from the neural network approach were similar to those from NONMEM. The residuals of the predictions from neural network analyses showed a positive correlation with that from NONMEM. Average absolute errors were 33.9 and 37.3% for neural networks and 39.9% for NONMEM. Average prediction errors were found to be 2.59 and -5.01% for neural networks and 17.7% for NONMEM. We concluded that neural networks were capable of capturing the relationships between plasma drug levels and patient-related prognostic factors from routinely collected sparse within-patient pharmacokinetic data. Neural networks can therefore be considered to have potential to become a useful analytical tool for population pharmacokinetic data analysis.
In this paper we present a data mining scenario that supports development of automated web-based documentation of data analysis for diagnosis and treatment. The documents can be seen as guidelines in a narrow sense, and are designed to include executable modules for the corresponding decision support systems. Our aim is to discuss the possibilities of identifying certain types of diagnoses and treatments for which guidelines can be generated and computerised more systematically.
An analog processing circuit is described which allow accurate measurement of the phase relationships between input angular acceleration and resulting eye velocity. Vestibular nystagmic data are processed via analog technics to yield slowphase eye velocity. The turntable velocity input is cross-correlated with the eye velocity output, using a Nicolet MED-80 minicomputer system. The resulting correlograms are further processed to obtain precise phase information. Test data analysis shows a system resolution within 1 degree. Data from human and animal subjects are portrayed.
INTRODUCTION: Patient Data Management Systems (PDMS) for ICUs collect, present and store clinical data. Various intentions make analysis of those digitally stored data desirable, such as quality control or scientific purposes. The aim of the Intensive Care Data Evaluation project (ICDEV), was to provide a database tool for the analysis of data recorded at various ICUs at the University Clinics of Vienna. SETTINGS: General Hospital of Vienna, with two different PDMSs used: CareVue 9000 (Hewlett Packard, Andover, USA) at two ICUs (one medical ICU and one neonatal ICU) and PICIS Chart+ (PICIS, Paris, France) at one Cardiothoracic ICU. CONCEPT AND METHODS: Clinically oriented analysis of the data collected in a PDMS at an ICU was the beginning of the development. After defining the database structure we established a client-server based database system under Microsoft Windows NI and developed a user friendly data quering application using Microsoft Visual C++ and Visual Basic; RESULTS: ICDEV was successfully installed at three different ICUs, adjustment to the different PDMS configurations were done within a few days. The database structure developed by us enables a powerful query concept representing an 'EXPERT QUESTION COMPILER' which may help to answer almost any clinical questions. Several program modules facilitate queries at the patient, group and unit level. Results from ICDEV-queries are automatically transferred to Microsoft Excel for display (in form of configurable tables and graphs) and further processing. CONCLUSIONS: The ICDEV concept is configurable for adjustment to different intensive care information systems and can be used to support computerized quality control. However, as long as there exists no sufficient artifact recognition or data validation software for automatically recorded patient data, the reliability of these data and their usage for computer assisted quality control remain unclear and should be further studied.
For linear regression analysis of longitudinal data with Gaussian response, I propose a new model to generalize the traditional class of random effects models in which the random effects are deterministic polynomials with coefficients randomly distributed over subjects with mean zero. The generalization is accomplished by adding zero mean Gaussian 'disturbances' to the highest derivative of each random coefficient subject polynomial, independently at each observation time. The resulting random effects, which have mean zero at each observation time, are called disturbed highest derivative polynomials (DHDPs). The disturbances induce serial correlation and also allow the subject-specific DHDP time trends to be non-linear. I do not estimate the subject-specific DHDP time trends. Analysis is based on the marginal model, that is, the fixed effects or population model obtained by integrating the random polynomial coefficients and all disturbances out of the joint distribution of themselves and the response vector. This allows a 'population averaged' interpretation. One can select the DHDP order by an information criterion. When the population time trend is not correctly modelled, the optimal DHDP order will be larger than when it is correctly modelled. One can make the covariance matrix of the regression coefficients robust to errors in modelling the within-subject dependence. I describe the relationship of a DHDP to a smoothing polynomial spline, and show how to replace the DHDP model with a smoothing polynomial spline model for the within-subject dependence in the marginal model.
The methods of Bayesian statistics are applied to the analysis of fMRI data. Three specific models are examined. The first is the familiar linear model with white Gaussian noise. In this section, the Jeffreys' Rule for noninformative prior distributions is stated and it is shown how the posterior distribution may be used to infer activation in individual pixels. Next, linear time-invariant (LTI) systems are introduced as an example of statistical models with nonlinear parameters. It is shown that the Bayesian approach can lead to quite complex bimodal distributions of the parameters when the specific case of a delta function response with a spatially varying delay is analyzed. Finally, a linear model with auto-regressive noise is discussed as an alternative to that with uncorrelated white Gaussian noise. The analysis isolates those pixels that have significant temporal correlation under the model. It is shown that the number of pixels that have a significantly large auto-regression parameter is dependent on the terms used to account for confounding effects.
A simple program written in Microsoft BASIC for the analysis of radioimmunoassay data is presented. The program was designed for use by unexperienced operators on a TRS-80 Model III microcomputer. Standard curve data are fitted by linear regression after a logit response/loge dose transformation. Standard errors of individual observations are computed, and confidence intervals are obtained for a critical t of 0.1. Standard observations are eliminated when they lie outside the confidence interval from the calculated response on the fitted curve, and this is accomplished until no datum can be eliminated or until the coefficient of determination reaches an operator defined value. The program then prints the regression parameters and the retained standard values. Unknown sample data can then be entered, and the values are computed relative to the final regression parameters of the standard curve. The values can be modified by operator defined correction/dilution factors. The mean value of the slope for 28 assays is -0.99, while the predicted theoretical slope with a loge transformation is -1. This program finds its greatest use in research laboratories where a number of different RIA are performed and where it is of value to have some control over the computation process.
While PACU nurses are increasingly conducting research studies to validate nursing practice, it is important to consider logical rules for data analysis. Level of measurement is an important consideration when selecting statistical tests to analyze the data. Statistical tests need not remain a mystery since level of measurement is the deciding factor for selecting which tests are appropriate for answering the research questions or testing the hypotheses.
Protein expression profiling by proteomics has the potential to be an ideal tool for the description of changes in complex biological systems or in the characterization of a disease. This study analyzes in more quantitative terms how methodological drawbacks of the current technology can hamper the comprehensive analysis of large sets of spot patterns by statistical means. Irreproducibilities in spot intensities due to silver staining and geometric distortions of the spot patterns inherent to the electrophoresis procedure push even the semiautomatic alignment and matching to their limits. This leads to reduced matching efficiencies for identical spots if no additional exhaustive manual matching is performed for every single spot in all gels. As a consequence, only a limited number of spots can be matched in all gels of a large set. The statistical pattern analysis of such data sets will thus not allow the comprehensive description of relevant pattern changes.
Parkinson's disease (PD) is characterized by a degeneration of nigrostriatal dopaminergic neurons, which can be imaged with 123I-labeled 2 beta-carbomethoxy-3 beta-(4-iodophenyl) tropane ([123I]beta-CIT) and single-photon emission computed tomography (SPECT). However, the quality of the region of interest (ROI) technique used for quantitative analysis of SPECT data is compromised by limited anatomical information in the images. We investigated whether the diagnosis of PD can be improved by combining the use of SPECT images with morphological image data from magnetic resonance imaging (MRI)/computed tomography (CT). We examined 27 patients (8 men, 19 women; aged 55 +/- 13 years) with PD (Hoehn and Yahr stage 2.1 +/- 0.8) by high-resolution [123I]beta-CIT SPECT (185-200 MBq, Ceraspect camera). SPECT images were analyzed both by a unimodal technique (ROIs defined directly within the SPECT studies) and a multimodal technique (ROIs defined within individual MRI/CT studies and transferred to the corresponding interactively coregistered SPECT studies). [123I]beta-CIT binding ratios (cerebellum as reference), which were obtained for heads of caudate nuclei (CA), putamina (PU), and global striatal structures were compared with clinical parameters. Differences between contra- and ipsilateral (related to symptom dominance) striatal [123I]beta-CIT binding ratios proved to be larger in the multimodal ROI technique than in the unimodal approach (e.g., for PU: 1.2 vs. 0.7). Binding ratios obtained by the unimodal ROI technique were significantly correlated with those of the multimodal technique (e.g., for CA: y = 0.97x + 2.8; r = 0.70; P < 0.001). Concerning the correlations between SPECT data and clinical parameters, the significance levels in the multimodal ROI technique, for example, for the correlation between CA and the UPDRScom subscore (r = -0.49 vs. -0.32). These results show that the impact of [123I]beta-CIT SPECT for diagnosing PD is affected by the method used to analyze the SPECT images. The described multimodal approach, which is based on coregistration of SPECT and morphological imaging data, leads to improved determination of the degree of this dopaminergic disorder.
A 30-month prospective study of 27 Scandinavian boys with confirmed diagnosis of Duchenne muscular dystrophy was carried out to construct profiles of the natural history of the disease. Assessments which included measures of voluntary muscle strength and function were done at 3 monthly intervals except for the first and second which were separated by 1 month. Recently developed statistical methods for analysis of longitudinal data with repeated observations on the same individual were used avoiding the problem of induced serial correlations. This allowed for the construction of both reference and prediction profiles for the variables %MRC, motor ability, walking time for 10 m and the sum of myometry of seven muscle groups.
A BASIC computer program designed to facilitate analysis of twin data is presented. The program estimates genetic parameters and tests their statistical significance based on the genetic and environmental hypothesis.